Customer Insight Magazine - Spring 2016

www.customer-insight.co.uk | spring 2016

THE UKCSI IS AT ITS HIGHEST POINT SINCE JANUARY 2014

EDITORIAL

Foresight With six speakers and some lively Q&A sessions, a

The latest UKCSI results are on page 31 and of

Editor

transcript of TLF’s 16th annual Customer Experience

seized by the top two companies, Amazon and Utility

Conference (see page 11) would have generated a lot

Warehouse, to communicate a positive message to

of text. Keynote speaker Matthew Syed enthralled the

customers and re-inforce their loyalty. It also fits in

audience of 200 customer experience professionals with

Rachel Allen’s article on warming up customers on page

some thought provoking anecdotes and conclusions from

21. Many organisations still don’t realise the extent to

his new book “Black Box Thinking”, which we have

which communicating with customers effectively will help

reviewed on page 36. Black Box Thinking is all about the

to improve satisfaction and loyalty.

benefits of learning from failure and you can glimpse the

One of the topics that’s currently of great interest

essence of his thesis by contrasting the safety records

to the customer research community is text analytics.

of the airline and hospital industries. The difference,

Customers are posting comments about you on Twitter,

according to Matthew is the thorough, open-minded

Facebook and many less frequented corners of the web.

and positive way in which airlines approach learning

You get customer communications directly through

from failure compared with the defensive, superficial

a range of channels and you may have thousands of

and negative stance traditionally adopted by the medical

comments from customer surveys. And this is without

profession.

the growing number of software applications that seek

The airlines approach is clearly the best model for

to generate customer feedback or to stimulate two way

customer experience professionals who can learn from

communications by text or email with your customers.

lapsed customers, problems and complaints and any

In his Latest Thinking article on text analytics (page 7),

expression of customer dissatisfaction. Tom Newey

Stephen Hampshire says “I’m extremely excited about the

described at the same conference how RSA had used

potential of these techniques.” He goes on to explain how

customer feedback to very successfully map and improve

text analytics works, what it can do and what it can’t.

the complaints experience. The key was the fact that the

He also looks at the tipping point, in terms of comment

survey data had quantified exactly how much customer

volume, beyond which using a computer to analyse them

satisfaction increased or decreased depending on how

may become cheaper than using a human.

RSA staff behaved at key moments of truth along the complaints journey. Turn to page 12 if you would like a perfect half page summary of how to improve your own complaints journey.

EDITORIAL Editor Nigel Hill ADVERTISING Marketing Manager Charlotte Ratcliffe DESIGN & PRODUCTION Production Editor Chris Newbold

CONTACTS

Nigel Hill

particular interest this time is how the opportunity was

Creative Director Rob Egan Graphic Designers Becka Crozier Jordan Gillespie PRINTER AB Print Group Ltd

Customer Insight is the magazine for people who want to deliver results to employees, customers and any other stakeholders as part of a coherent strategy to create value for shareholders. We publish serious articles designed to inform, stimulate debate and sometimes to provoke.

We aim to be thought leaders in the field of managing relationships with all stakeholder groups. www.customer-insight.co.uk info@customer-insight.co.uk Customer Insight C/O The Leadership Factor Taylor Hill Mill Huddersfield HD4 6JA

NB: Customer Insight does not accept responsibility for omissions or errors. The points of view expressed in the articles by contributing writors and/or in advertisements included in this magazine do not necessarily represent those of the publisher. Whilst every effort is made to ensure the accuracy of the information contained within this magazine, no legal responsibility will be accepted by the publishers for loss arising from use of information published.

All rights reserved. No part of this publication may be reproduced or stored in a retrievable system or transmitted in any form

or by any means without prior written consent of the publisher. © CUSTOMER INSIGHT 2016

ISSN 1749-088X

www.customer-insight.co.uk | Spring 2016 Customer Insight 3

C O N T E N T S

-

Text Analytics Latest Thinking on this complex subject

11

TLF Customer Experience Conference Learnings from some of the UK’s leading customer service professionals

CONTRIBUTORS

06

S P R I N G

2 0 1 6

20

Warming Up Customers How to optimise response rates and engagement in customer surveys

Stephen Hampshire

Darren Wake

Rachel Allen

Sarah Stainthorpe

Conference speaker, Renaissance polymath and occasional climber

Customer relations expert, socialite and frequent runner

Customer satisfaction evangelist, author and lover of the outdoors

Human data miner, insight sorcerer and prone to mad challenges

4 Customer Insight Spring 2016 | www.customer-insight.co.uk

CONTENTS

LATEST THINKING Text Analytics

06

18

Find out how those funny little postage stamp things are improving your fresh fruit

25

How do you know which customers are truly loyal and which are vulnerable to competitors?

17

How do you launch a new range of supplements in a crowded market?

CONFERENCE TLF Conference

11

CASE STUDIES Winterbotham Darby Health circle It’s Fresh

16 17 18

RESEARCH Warming Up

20

RESEARCH Competitor Analysis

25

CONFERENCE Complaints Conference

38

Customer Effort is the new NPS, and a much better measure for many companies. But what is it?

30

CUSTOMER UKCSI

31

BOOK REVIEW Black Box Thinking, Matthew Syed

36

QUICK GUIDE Customer Effort

38

Greg Roche

Nigel Coxon

Customer satisfaction standard bearer, communicator and misguided football fan

Industrial marketing expert, divergent thinker and astrophotographer

DESIGNERS

Published by

Becka Crozier

Jordan Gillespie

Right brain mastermind, gastronomic ascetic and deadpan comedienne

Creative magus, genuine tyke and 20ft wave rider

www.customer-insight.co.uk | Spring 2016 Customer Insight 5

L AT E S T T H I N K I N G If you’re anything like me, you’ll have come to be suspicious of technologies that are long on marketing and short on case studies. Text analytics, unfortunately, is a typical example. I could list you a whole raft of suppliers who offer it, either as a sole software offering or as an adjunct to their research or consultancy work. It’s much harder to give you a list of organisations who are getting lots of value from using text analytics with customers, though many are trying. That doesn’t mean that text analytics has nothing to offer. Far from it. I’m extremely excited about the potential of these techniques. But it’s important to use them in the right place and in the right way, rather than simply throwing them into the mix because they sound cutting-edge. Getting value from text analytics requires you to be clear about what you’re trying to achieve, and prepared for the investment of time and resources needed for the best results. In this article I’m going to map out the landscape of text analytics, explain what it can and can’t do, and discuss some of the particular challenges of using it with customer survey data. I’ll also show a couple of examples that may whet your appetite. What is text analytics? Text analytics, or text mining, is about using words as raw data for the types of analysis that have traditionally been used with numbers. To do that we need to teach our computers a limited understanding of “natural language” text. In the past researchers often approached this from a relatively technical, linguistic, standpoint. Does grammar matter? For example, computers can be taught grammar. They can figure out, for example, the subject and object of a sentence. You can see this in action online using this tool - try a sentence of your own and see how well the computer does. In practice, this approach might be used to set up a programme to scan financial news feeds for acquisition data. From a chunk of text such as this: Our programme could figure out that the prospective buyer is Whirlpool and the prospective acquisition is Aga Rangemaster. It could also pick out the likely price. Scraping articles from across the web, we could construct a database based on articles like this. “Bag of words” Most applications of text analytics use a simpler “bag of words” approach. This means that we’re making no attempt to understand the grammar of a sentence or the way words in a document relate to each other. It turns out that, for many practical applications, you can do very well without going any further than looking at the words that appear in a document. Let’s say we were trying to classify newspaper articles by subject.

One includes the words “football”, “Chelsea”, “midfielder”, “referee” and another includes words like “EBITDA”, “FTSE”, “shareholders”. Pretty easy to categorise them, isn’t it? With enough examples to learn from, we can successfully categorise much more difficult cases. Understanding grammar is not always necessary, we can often do a good job just by looking at which words appear. Let’s look at an example of how text analytics can help us. Spam, spam, spam, spam The text analytics tool that most people use most often, without realising it, is their spam filter. These use a mixture of indicators (or “features”) to figure out which messages are spam and which are real. Those features might include things about the provenance of the message like whether the sender matches the from address, but also specific words that are highly correlated with spam (like “viagra”). This is a really good example of what we should try to emulate: A very clearly-defined task A high success rate (tuneable to balance false-positives and false-negatives) A combination of types of features Learning from mistakes to improve. How might we go about doing something similar with our data? Understanding what data we have The first job, in most text mining tasks, is to process our raw data to make it better suited for later steps. Even if you never pursue any other text analytic paths, and there are good reasons you might not, as we’ll see in a moment, this sort of analysis is well worth doing. This work gives us easy access to quantitative information about our text data, something we’re often likely to forget to consider. Size and shape, tidying up Start with the basics. How many “documents” do we have and how many words? In the case of survey text, we’d normally treat each person’s response to a single question as a document. This means that our documents are much shorter than some used in other tasks, but there are good reasons to treat each open-ended question as a separate text analysis task. Processing the text in those documents starts with some tasks to tidy them up. More or less universally: Remove punctuation Convert everything to lowercase Spellcheck Remove “stopwords”, such as “the”, “and”, which don’t add meaning. Lists of stop words in various languages are readily available. And many people would also: Expand contractions (e.g. “don’t” to “do not”) Invert negatives (e.g. “not good” to “bad”, although this is not always as simple as it first seems) Remove custom stopwords, known to be common and uninformative for that task (e.g. “account” for bank customers) Stem words (so that “customer” and “customers” become “custom”) After that we should have a set of data with the same number of documents, but far fewer words. The words that remain are more interesting, and much more likely to differ between types of customer. The next step is to think about what words we have left, and the best way is usually to look at the ones that appear most often.

Frequency You may be thinking that this is just a long-winded way of getting to a word cloud, and you’d be right. Word cloud software often does some of the processing we’ve just talked about behind the scenes. The difference is that we’ve made all these decisions ourselves, and this is just the start. The frequency of occurrence of words is not just the basis for a pretty picture. It’s a quantitative measure that we can use to understand our data. We can compare subgroups based on how customers feel (e.g. “promoters” vs “passives” and “detractors”) or customer type (e.g. gender). And we can monitor trends and changes over time. As well as looking at which words appear, it can be instructive to look at which words tend to appear together. n-grams and co-occurrence Like much of the jargon to do with text analytics, “n-grams” sound frighteningly complex, but are actually a very simple idea. There are many examples of concepts which are made up of more than one word, but which actually refer to a single idea. Some examples: “account manager”, “newcastle upon tyne”, “customer service”. You may well have come across this phenomenon When making word clouds. When two or more words consistently appear together like this, you have an n-gram. Two words together is a bigram, three is a trigram. These can be interesting to look at in the same way as the most frequent words, and often form a useful feature for any models we may go on to build. For example, bigram features such as “cheap watches” are useful for spam filters. More generally, we can look to see if there are any patterns in the way words occur. Do some words tend to occur in the same document (e.g. “slow” and “service”) across many comments with similar meaning? “Co-occurrence” can be used, in a large enough data set, in very much the same way as you would look at the correlation between numerical variables. It gives us a quantitative basis for understanding how strongly two concepts are linked together in customers minds. Looking at the patterns of correlation between words is interesting in its own right, but it also underpins the work we may go on to do looking for ways to categorise the topic of each comment. Classifying documents What we’ve done so far is interesting, sometimes even insightful, and it’s useful to be able to think quantitatively about text data. But it isn’t giving you what I know you’re reading this article to get to. You want a computer that can do all the hard work of reading loads of comments and coding them up into themes. So…is that possible? Sort of. Let’s start by considering what is it that happens when you read through comments and code them? You use your knowledge of English (or whatever language) to figure out what a comment is “about”, and then assign it a code or theme. To a large extent this is based on the words in the comment. Usually there’s a bit of back-and-forth between comments and coding frame while you figure out to what level of detail each theme needs to be broken down, striking the right balance between usability and precision. Often you combine the text of the comment with your knowledge about the business to make “intelligent” inferences about meaning. In those terms, a summary of current text analytics is: It’s relatively easy to train a computer to categorise based on words. The person refining the model will refine categories to reflect the same balance. Step 3 is, as it stands, very difficult. This requires understanding of what causes issues, not just grouping words, and that requires more complexity. It’ll happen one day, but the tools that will help here (e.g. “Deep Learning” networks) are cutting-edge and still being developed. In basic terms, the best you can expect is for a computer to do roughly as well as a human who knows absolutely nothing about your business. The question now becomes one of cost-benefit. If text analytics maxes out at “as good as a a cheap human”, maybe it’s cheaper to pay a human than it is to train and refine software. Why do you think Amazon employs people in its warehouses? Because they bring a quintessential humanity to the picking and packing; or because it’s cheaper (for now) than automation? The decision criterion here is not quality, but quantity. If you have 200 comments to analyse then text analytics is a pointless waste of time and money. If you have 200 million it’s a no-brainer. What we can say for sure is that if you have fewer than 1,000 documents (i.e. 1,000 comments for an individual question probed in a survey, not for the whole survey) it will be quicker to read and code the comments than it will to teach a machine how to analyse them. In most practical applications, setting up classification models still requires a surprising amount of human effort. Most models use “supervised learning”, which basically means that the computer needs a sample of human-coded documents from which to learn. In principle you can cheat the learning process by using off-the-shelf models that may suit your situation (e.g. one trained on comments from a customer satisfaction survey in your industry), but most organisations prefer to develop their own. How far are we away from getting computers to do the whole process? Perhaps not all that far. It is possible to get the computer to look at the words that appear in documents and, based on patterns of which words tend to occur together, group documents together. “Topic modelling” is the broad term for this, and there are a number of technical approaches that can be quite successful. Unfortunately there are, as yet, few good examples when it comes to using topic modelling with survey open-ends. Part of the problem is that our “documents” are much shorter than newspaper articles, which means there are relatively few terms for us to base our classification on. We’ll solve this problem, though. Either with topic modelling, or with even more sophisticated tools such as “deep learning” neural networks. It’s only a matter of time. I think a lot of progress will have been made by 2020. Predicting behaviour Figuring out what customers are talking about, and grouping comments into themes, is useful. No question about that, particularly if we can start to look at how these themes change over time or look at differences between types of customer. But what difference does it make? Perhaps the most exciting thing made possible by thinking analytically about words is that we can start to incorporate them in predictive models. This means that we can combine the words people use with the scores they give us, and potentially other information such as behaviour, to build more complete predictive models. What can we predict? The starting point is often overall satisfaction or NPS, but there’s no reason to stop there. Why not look at actual defections? Or sales? The more insight we have into each individual’s emotional landscape, the more likely we are to be able to predict their behaviour. Scores are useful for this, and (unlike some) I doubt we’ll be getting rid of them any time soon, but why not take advantage of the verbatim comments as well? Think about the difference between these two customers: “6 out of 10 - Yeah, no problems. Did what it said on the tin.” “6 out of 10 - It was just ok. I’ll probably try someone else next time.“ Neither of these customers is particularly happy, but if we can make use of their words we’re much more likely to be able to see that the second is most likely to stray. Summary This has been a brief summary of a very deep field. We’re flirting with Artificial Intelligence, and embracing its more respectable cousin Machine Learning. There’s no question that as machine learning tools develop they will take over more and more tasks that were previously the preserve of humans. Understanding text is just one example, although it’s certainly a big one. So what should you do about it? If I could choose three take-aways, they’d be these: Start thinking quantitatively about survey open-ends. There’s a lot of potential value in doing so, and some of it is relatively easy to get to. Don’t get excited by “text analytics”, the sexy buzzword. Think hard about why you want to do it, and honestly about whether it’s the right tool for the job. Then, when you have a cogent answer, get excited about being able to solve a real problem that was impossible before. If your problem is interesting enough, then it’s worth taking the time to solve it properly. Be prepared for a considerable investment of time to refine your model. If you’re anything like me, you’ll have come to be suspicious of technologies that are long on marketing and short on case studies. Text analytics, unfortunately, is a typical example. I could list you a whole raft of suppliers who offer it, either as a sole software offering or as an adjunct to their research or consultancy work. It’s much harder to give you a list of organisations who are getting lots of value from using text analytics with customers, though many are trying. That doesn’t mean that text analytics has nothing to offer. Far from it. I’m extremely excited about the potential of these techniques. But it’s important to use them in the right place and in the right way, rather than simply throwing them into the mix because they sound cutting-edge. Getting value from text analytics requires you to be clear about what you’re trying to achieve, and prepared for the investment of time and resources needed for the best results. In this article I’m going to map out the landscape of text analytics, explain what it can and can’t do, and discuss some of the particular challenges of using it with customer survey data. I’ll also show a couple of examples that may whet your appetite. What is text analytics? Text analytics, or text mining, is about using words as raw data for the types of analysis that have traditionally been used with numbers. To do that we need to teach our computers a limited understanding of “natural language” text. In the past researchers often approached this from a relatively technical, linguistic, standpoint. Does grammar matter? For example, computers can be taught grammar. They can figure out, for example, the subject and object of a sentence. You can see this in action online using this tool - try a sentence of your own and see how well the computer does. In practice, this approach might be used to set up a programme to scan financial news feeds for acquisition data. From a chunk of text such as this: Our programme could figure out that the prospective buyer is Whirlpool and the prospective acquisition is Aga Rangemaster. It could also pick out the likely price. Scraping articles from across the web, we could construct a database based on articles like this.

6 Customer Insight Spring 2016 | www.customer-insight.co.uk

L AT E S T T H I N K I N G

I

f you’re anything

TEXT ANALYTICS like me, you’ll have come to be suspicious of technologies that are long on marketing and short on case studies. Text

analytics, unfortunately, is a typical example. I could list you a whole raft of suppliers who offer it, either as a sole software offering or as an adjunct to their research or consultancy work. It’s much harder to give you a list of organisations who are getting lots of value from using text analytics with customers, though many are trying.

That doesn’t mean that text analytics has nothing to offer. Far from

In practice, this approach might be used to set up a programme to

it. I’m extremely excited about the potential of these techniques. But

scan financial news feeds for acquisition data. From a chunk of text

it’s important to use them in the right place and in the right way,

such as this:

rather than simply throwing them into the mix because they sound cutting-edge. Getting value from text analytics requires you to be clear about what you’re trying to achieve, and prepared for the investment of time and resources needed for the best results.

In this article I’m going to map out the landscape of text analytics, explain what it can and can’t do, and discuss some of the particular challenges of using it with customer survey data. I’ll also show a couple of examples that may whet your appetite.

What is text analytics?

Text analytics, or text mining, is about using words as raw data for the types of analysis that have traditionally been used with numbers.

To do that we need to teach our computers a limited understanding of “natural language” text. In the past researchers often approached this from a relatively technical, linguistic standpoint.

Does grammar matter?

For example, computers can be taught grammar. They can figure out, for example, the subject and object of a sentence. You can see this in action online using the Reed-Kellogg Diagrammer tool - try a sentence of your own and see how well the computer does.

Our programme could figure out that the prospective buyer is Whirlpool and the prospective acquisition is Aga Rangemaster. It could also pick out the likely price. Scraping articles from across the web, we could construct a database based on articles like this.

“Bag of words” Most applications of text analytics use a simpler “bag of words” approach. This means that we’re making no attempt to understand the grammar of a sentence or the way words in a document relate to each other. It turns out that, for many practical applications, you can do very well without going any further than looking at the words that appear in a document.

Let’s say we were trying to classify newspaper articles by subject.

One includes the words “football”, “Chelsea”, “midfielder”, “referee” and another includes words like “EBITDA”, “FTSE”, “shareholders”. Pretty easy to categorise them, isn’t it? With enough examples to learn from, we can successfully categorise much more difficult cases. Understanding grammar is not always necessary, we can often do a good job just by looking at which words appear. Let’s look at an example of how text analytics can help us.

www.customer-insight.co.uk | Spring 2016 Customer Insight 7

L AT E S T T H I N K I N G

Spam, spam, spam, spam

Size and shape, tidying up

After that we should have a set of data

Start with the basics. How many “docu-

with the same number of documents, but far

use most often, without realising it, is their

ments” do we have and how many words? In

fewer words. The words that remain are more

spam filter. These use a mixture of indicators

the case of survey text, we’d normally treat

interesting, and much more likely to differ

(or “features”) to figure out which messages

each person’s response to a single question as

between types of customer.

are spam and which are real. Those features

a document. This means that our documents

might include things about the provenance of

are much shorter than some used in other

we have left, and the best way is usually to

the message like whether the sender matches

tasks, but there are good reasons to treat

look at the ones that appear most often.

the from address, but also specific words

each open-ended question as a separate text

that are highly correlated with spam (like

analysis task.

The text analytics tool that most people

“viagra”).

The next step is to think about what words

Frequency

Processing the text in those documents

You may be thinking that this is just a

starts with some tasks to tidy them up. More

long-winded way of getting to a word cloud,

should try to emulate:

or less universally:

and you’d be right.

• A very clearly-defined task

• Remove punctuation

the processing we’ve just talked about behind

This is a really good example of what we

Word cloud software often does some of • A high success rate (tuneable to balance false-positives and false-negatives) • A combination of types of features • Learning from mistakes to improve.

How might we go about doing something

•C onvert everything to lowercase

the scenes. The difference is that we’ve made

• Spellcheck

all these decisions ourselves, and this is just

emove “stopwords”, such as “the”, •R

the start.

“and”, which don’t add meaning. Lists of

The frequency of occurrence of words

stop words in various languages are readily

is not just the basis for a pretty picture.

available.

It’s a quantitative measure that we can use

similar with our data?

to understand our data. We can compare

Understanding what data we have The first job, in most text mining tasks, is to process our raw data to make it better suited for later steps. Even if you never pursue any other text analytic paths, and there are good reasons you might not, as we’ll see in a moment, this sort of analysis is well

And many people would also:

“promoters” vs “passives” and “detractors”) xpand contractions (e.g. “don’t” to “do •E not”)

As well as looking at which words appear,

although this is not always as simple as it

it can be instructive to look at which words

first seems)

tend to appear together.

emove custom stopwords, known to be •R common and uninformative for that task

to quantitative information about our text

(e.g. “account” for bank customers) tem words (so that “customer” and “cus•S tomers” become “custom”).

to consider.

or customer type (e.g. gender). And we can monitor trends and changes over time.

• I nvert negatives (e.g. “not good” to “bad”,

worth doing. This work gives us easy access data, something we’re often likely to forget

subgroups based on how customers feel (e.g.

n-grams and co-occurrence Like much of the jargon to do with text analytics, “n-grams” sound frighteningly complex, but are actually a very simple idea.

There are many examples of concepts which

Raw text

Processed text

are made up of more than one word, but which

Comments

5981

5981

actually refer to a single idea. Some examples:

Total words

38297

23342

Unique words

4079

1975

Words used once

2182

867

clouds. When two or more words consistently

Words used 50+ times

131

95

appear together like this, you have an n-gram.

Top 10 words

the

2416 (6%)

price

615 (3%)

and

1456 (4%)

noth

533 (2%)

they

1045 (3%)

staff

445 (2%)

same way as the most frequent words, and

more

669 (2%)

store

422 (2%)

often form a useful feature for any models we

have

570 (1%)

get

376 (2%)

not

491 (1%)

shop

370 (2%)

are

450 (1%)

rang

357 (2%)

see if there are any patterns in the way words

for

424 (1%)

good

349 (1%)

occur. Do some words tend to occur in the

there

379 (1%)

just

344 (1%)

nothing

373 (1%)

till

319 (1%)

8 Customer Insight Spring 2016 | www.customer-insight.co.uk

“account manager”, “newcastle upon tyne”, “customer service”. You may well have come across this phenomenon when making word

Two words together is a bigram, three is a trigram.

These can be interesting to look at in the

may go on to build. For example, bigram features such as “cheap watches” are useful for spam filters. More generally, we can look to

same document (e.g. “slow” and “service”) across many comments with similar meaning?

L AT E S T T H I N K I N G

Percentage of customers using word

space discount last left point bargain habit cashier slight layout show great petrol bag kind poor meal reason everyday depend didnt brought said right town cut pack competit particular morrison bought keep free year higher week friendli pleasant trolley that wouldnʼt person deliveri area seem manag sainsburi low fault littl general card wider present chicken

check supermarket drinkquick sometim tesco checkout stuff apologis butcher usual plenti well valu back bring often everyth friend much time meat queue rather

money larger

basic

wouldnt

bit need

refurbish

isnt

fruit

one

got

get till better

none

main

self

foodextra far

arent

best

counter

will

take

didnʼt

bell

now

item

next

express room

two

took

conveni

select

certain

full

problem

end

sort

milk

live

went

tea

busi

long

date

feel

sure

late

put appear whole

veget tell theyr

might

pop

away dear isnʼt

thing donʼt

wasn’t gone

look

new

pay

walk

car

half

regular spend

canʼt

cook old

near sell

shout wide

custom someth shelvtincome never say came slow love shelf ask limit brand way even bad place locat mani wasn’t train see three though high day cloth apart give section wrong smaller

wait potato

fine

easi

voucher number speed

“Co-occurrence” can be used, in a large

Classifying documents

5%

good

5%

till

5% 4% 4%

bigger

4%

offer

4%

realli

4%

1. You use your knowledge of English (or whatever language) to figure out what a

enough data set, in very much the same way as you would look at the correlation between

5%

just

know

there

line

difficult dearer

shop

product

move

open nicediffer buy lower local expens alway stock sainsbury’ member also around improv work

short corner

els handi callround queu

hour

tend

size

fact

wine access

home fish

6%

get

reduc bake

small

6%

rang

market

make

good staff price realli bigger product

just

prefer

6%

store

build

varieti choic

havent

nothrang

onto

ring

closer

help

anyth quit bread

shop

store

park

couldnt

deal

7%

staff

qualiti servic happi think run offer want enough fresh anoth fair

like know can

someon special use cost find big

lot

clean

9%

noth

level close cant

peopl cheaper

10%

price

What we’ve done so far is interesting,

comment is “about”, and then assign it a code or theme. To a large extent this is

numerical variables. It gives us a quantita-

sometimes even insightful, and it’s useful to be

tive basis for understanding how strongly two

able to think quantitatively about text data. But

concepts are linked together in customers’

it isn’t giving you what I know you’re reading

2. Usually there’s a bit of back-and-forth

minds.

this article to get to. You want a computer that

between comments and coding frame

can do all the hard work of reading loads of

while you figure out to what level of detail

between words is interesting in its own right,

comments and coding them up into themes.

each theme needs to be broken down,

but it also underpins the work we may go on

So…is that possible? Sort of. Let’s start by

striking the right balance between usabil-

to do looking for ways to categorise the topic

considering what is it that happens when you

of each comment.

read through comments and code them?

Looking at the patterns of correlation

based on the words in the comment.

ity and precision.

3. Often you combine the text of the comment with your knowledge about the business to make “intelligent” inferences

think

staff

about meaning.

store bigger

In those terms, a summary of current text analytics is:

till shop

1. It’s relatively easy to train a computer to one

rang

categorise based on words.

offer

2. The person refining the model will refine categories to reflect the same balance.

product

3. Step 3 is, as it stands, very difficult. This don’t

just

get

requires understanding of what causes issues, not just grouping words, and that requires more complexity. It’ll happen one

better noth

day, but the tools that will help here (e.g.

know

“Deep Learning” networks) are cuttingedge and still being developed.

realli

good

In basic terms, the best you can expect

like

is for a computer to do roughly as well as a

lower

human who knows absolutely nothing about

thing price

your business.

The question now becomes one of costbenefit. If text analytics maxes out at “as

stock

good as a cheap human”, maybe it’s cheaper to pay a human than it is to train and refine

www.customer-insight.co.uk | Spring 2016 Customer Insight 9

L AT E S T T H I N K I N G

Why did you leave?

Combining text and scores to predict behaviour

Topic 1

Topic 2

Topic 3

price

rude

moved

cost

unhelpful

location

expensive

better

far

cheaper

never

closer

money

tried

house

Using LDA to discover topics from distribution of words

“seamless”

“hassle”

sat

“cheaper”

nps

“speed”

vfm

“advice”

software. Why do you think Amazon employs

means there are relatively few terms for us

we’re much more likely to be able to see that

people in its warehouses? Because they bring

to base our classification on. We’ll solve this

the second is most likely to stray.

a quintessential humanity to the picking and

problem, though. Either with topic modelling,

packing; or because it’s cheaper (for now)

or with even more sophisticated tools such as

than automation? The decision criterion here

“deep learning” neural networks. It’s only a

This has been a brief summary of a very

is not quality, but quantity. If you have 200

matter of time. I think a lot of progress will

deep field. We’re flirting with Artificial Intel-

comments to analyse then text analytics is

have been made by 2020.

ligence, and embracing its more respectable

a pointless waste of time and money. If you have 200 million it’s a no-brainer. What we can say for sure is that if you have fewer than

Predicting behaviour Figuring out what customers are talking

Summary

cousin Machine Learning. There’s no question that as machine learning tools develop they will take over more and more tasks that were

1,000 documents (i.e. 1,000 comments for an

about, and grouping comments into themes,

previously the preserve of humans. Under-

individual question probed in a survey, not

is useful. No question about that, particularly

standing text is just one example, although

for the whole survey) it will be quicker to

if we can start to look at how these themes

it’s certainly a big one.

read and code the comments than it will to

change over time or look at differences

teach a machine how to analyse them.

between types of customer. But what differ-

In most practical applications, setting up classification models still requires a surpris-

So what should you do about it? If I could choose three take-aways, they’d be these:

ence does it make?

Perhaps the most exciting thing made

• Start thinking quantitatively about survey

ing amount of human effort. Most models use

possible by thinking analytically about words

open-ends. There’s a lot of potential value

“supervised learning”, which basically means

is that we can start to incorporate them in

in doing so, and some of it is relatively

that the computer needs a sample of human-

predictive models. This means that we can

coded documents from which to learn. In

combine the words people use with the scores

principle you can cheat the learning process

they give us, and potentially other informa-

sexy buzzword. Think hard about why you

by using off-the-shelf models that may suit

tion such as behaviour, to build more com-

want to do it, and honestly about whether

your situation (e.g. one trained on comments

plete predictive models. What can we predict?

it’s the right tool for the job. Then, when

from a customer satisfaction survey in your

The starting point is often overall satisfaction

you have a cogent answer, get excited

industry), but most organisations prefer to

or NPS, but there’s no reason to stop there.

about being able to solve a real problem

develop their own.

Why not look at actual defections? Or sales?

How far are we away from getting com-

The more insight we have into each individu-

easy to get to.

• Don’t get excited by “text analytics”, the

that was impossible before.

• If your problem is interesting enough,

puters to do the whole process? Perhaps not

al’s emotional landscape, the more likely we

then it’s worth taking the time to solve it

all that far. It is possible to get the computer

are to be able to predict their behaviour.

properly. Be prepared for a considerable

to look at the words that appear in docu-

Scores are useful for this, and (unlike

ments and, based on patterns of which words

some) I doubt we’ll be getting rid of them any

tend to occur together, group documents

time soon, but why not take advantage of the

together. “Topic modelling” is the broad term

verbatim comments as well? Think about the

for this, and there are a number of technical

difference between these two customers:

approaches that can be quite successful.

Unfortunately there are, as yet, few good examples when it comes to using topic modelling with survey open-ends. Part of the problem is that our “documents” are much shorter than newspaper articles, which

investment of time to refine your model.

“6 out of 10 - Yeah, no problems. Did what it said on the tin.”

“6 out of 10 - It was just okay. I’ll probably try someone else next time.“ Neither of these customers is particularly happy, but if we can make use of their words

10 Customer Insight Spring 2016 | www.customer-insight.co.uk

Stephen Hampshire Client Manager TLF Research stephenhampshire@leadershipfactor.com

CONFERENCE

16th Annual Customer Experience Conference

Almost 200 people gathered together at One Great George Street between the Houses of Parliament and Horse Guards Parade for TLF’s 16th Annual Customer Experience Conference. It was the 4th consecutive year for the venue, home to the Institution of Civil Engineers. Anyone who’s seen the amazing Georgian interiors will know why delegates rate this higher than any other venue on the post-conference satisfaction survey.

A

t the start

of the Conference, TLF

shared the 2015 results of the UK

profit maximisation above all else), cutting cost on customer service is easy. It

Customer Satisfaction Index (UKCSI),

doesn’t have an immediate perceptible

showing a 4th consecutive fall in satis-

impact unlike sales boosting alternatives

faction across UK consumers. With GDP

such as a big advertising campaign or a

growing, unemployment down, infla-

price cutting promotion. But boy does it

tion at an all time low and consumer

have a long term impact!

confidence improving why was the UKCSI

There is growing interest in the

falling? A clue might lie in Tesco’s story

Walkaway £. It’s increasing. Why?

– fewer staff delivering customer service,

Customers are becoming more confident

stingier Clubcard rewards and stores

all the time. They have more informa-

starved of investment. It might be the

tion and channels to find alternatives.

most notorious example of putting profit

But companies often don’t make it easy

before customer experience but it’s by

for customers to stay. For example, they

no means the only one. Bigger standard

make it much easier for potential new

deviations on the UKCSI results revealed

customers to get in touch as well as

a growing gap between the best organ-

offering them better deals. Ask yourself

isations and those that were not putting

– how easy is it for existing customers

customers first.

seeking customer service to get in touch

The trouble is that for companies like Tesco that are in financial trouble (or

with your company compared with new ones searching for information on deals?

those who are not but put short term

www.customer-insight.co.uk | Spring 2016 Customer Insight 11 Greg Roche

CONFERENCE

AI and the Customer Stephen Hampshire, Client Manager at TLF, used the opening

anything about, who they mix with and are influenced by and how they feel. To get this fully rounded view that will maximise your

conference presentation to consider how customer experience might

ability to anticipate future customer behaviour you will need to

be affected by the development of Artificial Intelligence. Stephen

combine talking to customers as well as collecting volumes of data

Hawking and some other eminent scientists have singled out AI as

about their past behaviour.

the biggest future threat to the human race. At present the days of

As Stephen pointed out, AI will get ever better at physical tasks and

computers being able to improve their own intelligence and operate

rational thinking but Emotional Intelligence will take much longer. He

independently of any human generated programme seem well in the

cited Google Translate as a good example. It’s improved by leaps and

future but some very basic forms of AI are already intruding on the

bounds over the last 2 or 3 years but still struggles with the difference

customer experience. Speech recognition is an example.

between tu and vous!

There is enormous interest in the customer management world in

So for the future of customer insight, you’ll have to invest in

using big data to understand customer behaviour. The attempts remain

AI and big data, but don’t forget

primitive but with location tracking, escalating volumes of customers’

about EI. For that you’ll still have

purchasing history and, further ahead, advances in face recognition,

to get insights from the customers

organisations will become increasingly knowledgeable about

themselves.

customers’ past behaviours. They’ll certainly remember what you did last time, and all the times before that, better than you do. But it won’t help them to get better at understanding why you did it.

Basing predictions about future customer behaviour purely on information about how they behaved in the past is very dangerous.

You have to understand the drivers of that behaviour which include other things happening in their lives that organisations don’t know

Stephen Hampshire Tom Newey

RSA – improving the complaints experience The first of 4 TLF customer research clients to present was Tom

But having the right research outputs is only the starting point. It

Newey who delivered a very useful explanation of how RSA had used

will all be wasted if the company doesn’t use it properly. Here’s how

customer feedback to map and improve the complaints experience.

RSA used it:

Their survey data had quantified exactly how much customer

1. Internal feedback

satisfaction increased or decreased depending on how RSA staff had

Extensive internal feedback to all using story telling visuals produced

behaved at key moments of truth along the complaints journey. With

by TLF ensured that all 65 Customer Relations staff understood the

this data collected and mapped it is a short step to mobilise staff to

Customer Journey and the key Moments of Truth.

focus on the moments of truth and the behaviours that make the

2. Agree actions

biggest difference. And here are the things that RSA did focus on to

Don’t tell staff what they have to do, explore the research findings and

successfully improve the complaints experience.

customer insights with staff and agree actions with them. It won’t be

1. Listen and acknowledge

difficult. With the right research outputs the customers will already

Listen carefully to the customer’s problem, address them by name and

have told staff exactly what they have to do.

acknowledge that you have understood the complaint.

3. Audit

2. Inform

Introduce systems to monitor staff behaviours and include in their

Tell the customer exactly what will happen next and how long it will

121s and performance reviews. In February 2014 staff maintained that

take.

they were carrying out the required actions but according to customers

3. Clarify

were doing so only 77% of the time. A target of 98% was set and

Make sure you have exactly the right understanding of the outcome

compliance was up to 96% by 2015.

the customer expects.

4. Review

4. Follow-up

Introduce a clear visual dashboard to give staff and management

Call the customer at the end to make sure the customer is happy with

regular updates on whether the actions are being implemented and

the process and, especially if the outcome was not their favoured one,

how much it’s improving customer satisfaction.

fully understands the reasons for the decision.

Once these actions have been agreed and communicated to staff,

And it worked. The greatly improved staff compliance with the agreed actions did result in a correspondingly good increase in

you can use customer experience modelling (CEM) to quantify exactly

customer satisfaction scores. Their Complaints Satisfaction Index was

how much these staff behaviours are happening. Or not. If they

up by 7% a year later and most pleasingly for Tom they generated an

happen consistently customer satisfaction will go up. As Tom said,

even bigger satisfaction gain amongst customers whose complaint was

Customer Journey infographics are a great way to communicate this.

not upheld (up 13%) demonstrating that whilst the outcome is always important, the way staff handle the complaint makes a big difference.

12 Customer Insight Spring 2016 | www.customer-insight.co.uk

CONFERENCE

Service with GUSTO With a new CEO and Chairman appointed in 2013 it has been a time

For a company that is changing

of considerable change at N Brown, a traditional catalogue mail order

so rapidly, understanding customers

firm (think JD Williams, High and Mighty) transitioning to a younger

is everything, Helen pointed out.

target customer, an online business model and now its first ever

That includes understanding their

high street stores. The stores are literally a shop window to showcase

customers on many different levels

its more fashionable power brands such as Simply Be and Jacamo.

– understanding them physically,

But Helen Jack from Customer Insight and Marketing emphasised

understanding them attitudinally and

that none of these innovations will succeed without committed staff

understanding them behaviourally.

delivering world class service. Hence GUSTO:

Segmenting customers by behaviour

Helen Jack

is much more advanced than the

GLOW with pride

traditional customer marketing segmentations by demographics. What

UNDERSTANDING is everything

successful companies are interested in is what customers do, and why

SAVING makes sense

they do it, and not how old they are, and what gender they are.

TOGETHERNESS is crucial

In a rapidly altering market that is very influenced by changes in

OPPORTUNITY exists everywhere

technology and the growth of online browsing, and online shopping,

2014 was the Year of the Employee at N Brown, with a Value

Helen demonstrated that only the most adaptive,

Pack distributed to every one of the 4,000 employees detailing the

flexible and quick changing companies,

company’s objectives, strategy and values. To ensure that staff fully

can understand and respond to the

understood it all and their own roles in making it happen, briefings

ever changing demands of

were organised with 80 Briefing Packs distributed to management to

the fashion world

make sure the same messages were communicated throughout the

customers.

company.

Helen explained that N Brown are committed at the most senior levels to “putting the customer at the heart of the business.” This is driven by many internal initiatives, one of which is the cross functional and cross level ACE forum (Achieving Customer Excellence.)

This brings together different parts of the organisation to ensure there is a broad and companywide view of the customer, and what will make the customer experience better. This also aids the communication of the customer strategy across the business.

Storytelling Sitting in an armchair Jackanory style, TLF Designer, Rob Ward,

clouds based on customer comments and infographics illustrating

was next on stage to show the audience how design can be used to

survey findings. But the impact can be greatly magnified by video

greatly enhance the impact of messages and insights from customer

and sound and showed how this can be effectively achieved in very

surveys. Tom Newey from RSA had already evidenced the benefit of

simple ways. A good example was the N Brown clothes rail – a

clarifying the Customer Journey and its key Moments of Truth. Rob

simple but relevant video of the key customer feedback messages

showed how good design will combine eye catching graphics with

sliding along the rail on hangers with musical accompaniment.

convincing statistics and insights from research.

What TLF has learned over and over again from working with clients is that customer survey results make a great impact at

customer comments appearing one by one on the background

first but as time goes

graphic with a real customer voicing the comment each time.

on organisations need

Employees’ eyes are obviously attracted by the messages appearing

to make more and

on the screen but it’s the real customer voices that make the biggest

more effort to bring

impact – you can feel the emotion as well as hear the comment.

survey results to life to

Rob Ward

The impact can be increased using the voices of real customers.

This can be very simple like the post-it notes summarising

The next step up in investment and impact is seeing the

fully engage staff. And

customer. A simple way is to accompany the audio with a customer

nothing does this better

name and photo but the best way is to video the customer. A

than the Voice of the

very effective example that Rob showed began with an audio of a

Customer. Rob started

customer ringing the company’s call centre, white text crawling

off by showing some very

across the dark screen adding impact to the frustrating customer

simple graphic techniques

experience. But the messages really hit home when the customer

for this such as word

then appeared in person explaining, in a very measured and

www.customer-insight.co.uk | Spring 2016 Customer Insight 13

CONFERENCE

articulate way why the poor experience had driven this very reasonable

benefits of customer satisfaction surveys was certainly a message that

man to become extremely dissatisfied and a vocal complainer.

had been taken on board by the next presenters.

Using high impact communications techniques to magnify the

Striding up the League Table The next presentation was from a B2B environment. Helen

There was extensive

Gutteridge and Scott Jackson from the British Gypsum customer

internal feedback of

satisfaction team explained how the company had improved its

all survey results and

customer satisfaction from 79% in 2007 (not quite in the top half

customer comments.

compared to other organisations) to 87.5% at the end of 2014 – right

This didn’t just cover

up there in the top 5% compared to other organisations across all

employees but also

sectors.

external companies such as hauliers who are important partners in

Helen Gutteridge & Scott Jackson

Although it was average on TLF’s huge Satisfaction Benchmark

delivering the customer experience. The staff were further engaged

League Table, British Gypsum were devastated by what they saw as a

by introducing customer satisfaction-related pay, which motivated

very poor result in 2007 and the staff were shocked and upset by the

them enormously. Almost without exception in TLF’s experience, the

critical customer comments. This led to the development of a 5 year

companies that improve customer satisfaction the most have bonuses

vision – “To be the UK’s best building material supplier as defined

for all staff when satisfaction targets are hit.

by our customers.” At the core of the 5 year strategy was a plan “to

Throughout the satisfaction improvement journey, there was

be the best at the things that matter most to customers”. But a big

also extensive communication of the survey results to customers as

factor in British Gypsum’s amazing progression up the Satisfaction

well as updates on the implementation of satisfaction improvement

League Table was that the ‘doing best what matters most’ plan

action plans. In his Conference opening address, Greg Roche had

wasn’t just one plan for the whole business, it was 6 plans, one for

highlighted the importance of feeding back customer satisfaction

each segment. British Gypsum’s sample size was large enough to get

results to customers and, above all, telling them what actions you are

importance scores, satisfaction scores and satisfaction gaps for 25

taking as a result of listening to customers. He showed an example

key customer requirements in each of the 6 segments – distributors,

of an excellent 12 page customer feedback report produced by Tyco

main contractors, sub-contractors, merchants, architects and house

in America, but even they hadn’t gone to the trouble taken by British

builders.

Gypsum, who have recently produced a full length DVD which was

To help develop the 6 satisfaction improvement plans, staff visited

sent to all customers. It showed what British Gypsum had learned

customers and observed them at key points along the customer

from their customer satisfaction survey across the entire business

journey such as ordering the product, taking deliveries, using the

and what actions they were taking and it was reinforced by interviews

product, needing technical help etc. Based on the survey data and

with staff from all departments who explained exactly what they were

the observation, detailed Customer Journey Maps were developed

doing to deliver a world class customer experience. Without doubt

for each segment. This, in turn, led to event-based surveys around

this will have made maximum impact and will further enhance British

key Moments of Truth. Bigger annual customer satisfaction surveys

Gypsum’s reputation for customer focus.

monitored progress as the Customer Satisfaction Index steadily worked its way up the League Table.

Brand Matters Rob Warm, Head of Member

Rob then went on to say that one of the difficulties facing a

Relations at the National Housing

membership organisation is the varying and sometimes contradictory

Federation began by showing

demands of different members, e.g. larger compared with smaller

the audience a couple of posters

housing associations. On top of this is the changing nature of the

that they would have walked past

relationship most membership bodies now have with their members,

if they had arrived via nearby

driven by the need to become more commercial. These were the kind

Westminster tube station. As well

of factors that motivated NHF to do their first membership satisfaction

as very effectively illustrating the

survey. With members’ renewal decision having a growing cost-

difficulty of getting on the London

benefit element, NHF needs to make sure that it remains relevant and

housing ladder they illustrate one

delivers value for money, basically by doing best what matters most

of the key roles the Federation

to its members. The gaps between importance and satisfaction were

plays in the eyes of its members –

therefore very useful to NHF in developing action plans following the

campaigning on housing issues.

survey.

Rob Warm 14 Customer Insight Spring 2016 | www.customer-insight.co.uk

However, Rob’s favourite question was “Is

image. His parting shot? “Don’t under-estimate

NHF an organisation that you are proud to be

the importance of brand”. It’s not just for B2C. It

a member of?” His reasoning was that whilst

makes a difference to all organisations.

members’ renewal decision will be increasingly a

The keynote speaker was Matthew Syed who

commercial one going forward, if your customers

made Conference history by being the first speaker

are emotionally attached they are much more

ever invited back by TLF because delegates thought

likely to make a ‘rational’ decision that your

he was so good the first time round. To read about

product or service provides value for money.

Matthew’s ideas and his new book, turn to page 36.

Hence the posters in the tube. As well as delivering commercial services efficiently, NHF needs to promote the profile of the Federation and the sector to constantly reinforce its status and

Greg Roche Director TLF Research gregroche@leadershipfactor.com

CUSTOMER EXPERIENCE WORLD LONDON 2016

RENAISSANCE, HEATHROW

CONFERENCE | 18 - 19 MAY

CUSTOMER EXPERIENCE WORLD IS DIFFERENT - BECAUSE WE NOW NEED TO BE DIFFERENT. THE CUSTOMER EXPERIENCE SHOULD OFFER THE CORPORATE MORE INSIGHT INTO THE CUSTOMER’S BEHAVIOUR RESULTING IN A FAR MORE TARGETED SALES AND SERVICES APPROACH. ULTIMATELY 4 OUT OF 5 CUSTOMER EXPERIENCE MANAGEMENT INITIATIVES ARE STILL FAILING. WE ARE NOW IN THE AGE OF THE CUSTOMER LED APPROACH. THE CUSTOMER IS FIRMLY IN CONTROL. CUSTOMER EXPERIENCE EXPERTS WILL PROVIDE YOU WITH THEIR CUSTOMER EXPERIENCE INSIGHTS AND EXPERIENCES. PARTICIPATING IN ROUND TABLE DISCUSSIONS WILL GIVE YOU THE OPPORTUNITY TO DISCUSS YOUR OWN EXPERIENCE CHALLENGES GAINING FURTHER PRACTICAL KNOW HOW FOR YOUR UNIQUE EXPERIENCE REQUIREMENTS. Reserve your place using the code ‘MEDPAR16’ to receive a 20% discount www.focusgroupevents.com/CEW-London-2016/

F E AT U R E

“This survey made me hungry and I am trying to lose weight”

S

o said one

Basically a Usage and Attitude (U&A) sur-

pack of bite sized brioche would offer some

with objectives including:

exciting product development opportunities – as highlighted by these respondents.

• Occasions when eaten • How are they used • Consumer attitudes regarding - Healthiness

savoury and sweet?”

“I wish they did more different fillings.”

“I used to live in France and there we had brioche

- Packaging

in different shapes coated with large sugar crystals

• Product development opportunities

totally different to what’s available here.”

“I would like to see various varieties of brioche available in stores.”

ing the survey. Winterbotham Darby is

an award-winning supplier of high quality

“Maybe brioche could be sold as a multi-pack of

- Quality levels - Flavours respondent after complet-

its own rather than adding a filling, a mixed

vey, the research focused mainly on brioche

It was agreed the best way to obtain these Popular ideas for new flavours included

chilled, frozen and ambient food products.

insights was to survey a nationally repre-

Although the company supplies product

sentative sample of UK consumers who had

fruit, raisin/sultana and cinnamon. Baked-

across a wide range of categories includ-

purchased Continental Morning Goods in the

in-store brioche is also a big opportunity with

ing antipasti, bakery, chilled desserts, fresh

previous 3 months. TLF Panel was identified as

an overwhelming 91% of consumers giving it

pasta, frozen desserts, seafood and sausages,

the best data collection method for the speed

the thumbs up.

perhaps it was the photos of freshly baked

of data collection, the representativeness and

brioche rolls that accompanied this survey

granularity of the sample and the ability to use

“I work in a bakery and we also sell the pre-

that prompted the respondent’s comment!

product images. In addition to demographic

packaged version but there is nothing like the

Winterbotham Darby supplies retail and food

information, questions were asked about

service with clients such as Waitrose, M&S

respondents’ living arrangements, family unit

and Tesco. With full marketing, technical,

type, grocery spend and shopping habits as

logistics and sales support they work in part-

well as product usage and preferences.

freshly baked smell!”

The results split the responses according to the store where respondents bought brioche

Within a month a questionnaire had been

so were extremely well received by both Win-

agreed, the fieldwork had been conducted, the

terbotham Darby and the supermarkets, and at

responses had been analysed and charts pro-

the time of print the results are being used to

with TLF Panel since 2011 on a number of

duced and Winterbotham Darby had presented

further develop the products and their market-

consumer insight projects and in July 2015

the findings to their supermarket clients.

ing, which we’ll see in the shops before long!

nership with their customers to keep them at the forefront of their markets.

Winterbotham Darby has been working

commissioned a survey for one of their key

New product opportunities were identi-

categories – pre-packed Continental Morn-

fied by comparing consumers’ current buying

ing Goods. Consisting of Brioche, Croissant,

habits with their interest in product items that

Darren Wake

Pain au Chocolat, Waffle, Crepes and Danish,

they don’t currently buy. This identified bite

Business Development

the Continental Morning Goods market is

sized brioche as the most promising product

Manager

growing over 12% per annum, so it’s a big

development area. Since most consumers

opportunity.

regard brioche as a special treat and eat it on

16 Customer Insight Spring 2016 | www.customer-insight.co.uk

TLF Research darrenwake@leadershipfactor.com

F E AT U R E

H

F t for Purp se

ealthcircle are a strategic and

creative

advertising and marketing agency

The survey began by showing competitors’ current multivitamin print ads with

specialising in the healthcare and Pharma

questions designed to understand how much

industries with offices in London, New York

each ad grabbed the respondent’s attention,

and Singapore. Their client, Nature’s Way,

how engaging the ads were, how memorable

produce multivitamin and healthcare products

the headlines were and what main message

like pro-biotics and herbal supplements. The

each ad had conveyed. The questions then

market for dietary supplements has been very

moved on to the panellist’s own opinions

strong over the last decade due to the ageing

about matters such as ad layout, visual appeal

population, greater interest in health and

and effectiveness. Finally respondents did

wellbeing, more awareness of preventative

a word association exercise for each ad and

healthcare, more self-directed research on the

rated their overall likelihood to purchase the

internet and a considerable amount of adver-

product after seeing the ad.

tising by the leading brands. But there are

After showing competitor ads and obtain-

challenges as sales of branded vitamins and

ing opinions, the survey showed 4 concept

mineral supplements in the UK have flatlined

adverts that the agency was considering as

recently since a shift towards healthier diets

press ads for their multivitamin product.

has impacted many consumers’ perceived

After they had seen the 4 new concept ads

need for supplements plus own label sales

for Nature’s Way, respondents were probed

are growing. However, sales of demographi-

for memory recall then asked to rate key

cally targeted supplements have continued to

attributes for each concept. Word association

fare well reflecting the consumer’s desire for

questions were also asked in relation to the 4

personalisation. These trends made it more

ad concepts as well as likelihood to purchase

important for Nature’s Way to make the best

the product based on seeing each of the ads

use of its advertising budget as the American

in future. Finally respondents were asked to

company seeks to establish itself in the UK as

rank the 4 concepts in order of preference.

well as highlighting the importance of demographically targeted products.

A winning concept

Ad testing

identified opinions and preferences overall

In September 2015 Healthcircle and Nature’s

Following analysis by TLF, the results and by segment and clearly established a

Way asked TLF Panel to test some press ads,

favourite concept, which was taken forward

competitor ads and some new advertising con-

and developed by Healthcircle for use in a

cepts. The survey featured visuals of the ads

print advertising campaign destined initially

and concepts and was focused on their target

for the Boots Health and Beauty Magazine,

market of 30-49 year old females. The results

which is free to Advantage card holders and

were segmented on full / empty nest, age of

Dare, which is Superdrug’s beauty and well-

children, multivitamin purchase and usage,

being magazine.

Darren Wake Business Development Manager TLF Research darrenwake@leadershipfactor.com

brand preference, purchase frequency, where purchased and likelihood to purchase more multivitamins in future.

www.customer-insight.co.uk | Spring 2016 Customer Insight 17

CASE STUDY

“This is a major breakthrough in the fight to combat food waste”

F

the retailer, and from the retailer into the

ripening process of a wide range of produce –

home. As well as addressing much bigger

including peaches, strawberries and tomatoes

delivering solutions for food freshness. From

global issues, its products are also designed

– allowing them to stay fresh for longer. The

a global population of approx. 6.5 billion

to increase consumer satisfaction through

filter works by absorbing ethylene, a natural

people, 1 billion are still undernourished and

delivering better taste and quality due to the

gas emitted by fruit as it ripens. It locks

to-date, there is a lack of effective, safe and

increased freshness of the products. As well

away the ethylene, which is a type of plant

ethical technology to help resolve this matter.

as reducing wastage, this is a benefit that

hormone, and so delays the point at which

The UN World Food Program, the world’s

appeals strongly to FFT’s retailer customers.

the fruit becomes over-ripe. The seven mil-

ood

Freshness Technology (FFT) is an

innovations company that focuses on

largest humanitarian agency, warned the

lion tons of food thrown away by households

world will only be able to produce enough

in the UK alone is mainly because the use-by

food for everyone in 2050 if food security is

date has passed or because it has gone off.

made a top priority. Food charity, Love Food

A useful tip is that as well as extending the

Hate Waste says that seven million metric

shelf-life of fresh food by up to four days, it

tons (MT) of food and drink is wasted in UK

can, for a limited period, be re-used at home,

households every year, with fresh produce

e.g. by placing it in a bowl of fruit.

among the top items trashed. FFT has there-

TLF Panel survey

fore made it their mission to make products that sustain the world’s fresh food supply

The Freshness Strip

and has concentrated on developing innovative technologies that address significant

FFT had won a contract with Morrisons to insert the Freshness Strip into punnets of

Shown in the photo, the product that was

strawberries and raspberries and were keen

unmet needs throughout the entire supply

the subject of a recent TLF Panel survey was

to understand the impact the freshness strip

chain; from the grower and manufacturer to

the Freshness Strip. It’s a little bigger than a

makes, how the fruit compares to fruit from

postage stamp and readers may have spot-

other supermarkets and how it impacts food

ted it and, like me, perhaps been intrigued

wastage levels. FFT then wanted to present

by it, when opening a pack of fresh fruit

the results to Morrisons to say the strip has

bought from the local supermarket. Well this

had a positive impact on customers and they

tiny piece of material is the answer to a big

should continue to use the strip. Utilising TLF

problem: that of rotting fruit. The product

Panel we surveyed a representative sample of

is actually a kind of filter that slows the

UK adults in July 2015 via an online survey.

Darren Wake Business Development Manager TLF Research darrenwake@leadershipfactor.com

18 Customer Insight Spring 2016 | www.customer-insight.co.uk

F E AT U R E

Qualification questions were applied and only

Outcomes

food waste and could save the fresh produce

FFT were very pleased with the survey

industry tens of millions of pounds each year.

and / or raspberries from Morrisons in the

results, which they presented to Morrisons,

It will also mean that shoppers will be able to

previous 5 weeks went on to complete the

and with the subsequent outcomes. Morrisons

keep fruit and vegetables for longer without

survey.

decided to extend the use of the Freshness

feeling pressured to eat them within days of

Strip and other supermarkets have adopted

buying them.”

consumers who had purchased strawberries

Consumers who had purchased Morrisons’ strawberries only were asked questions

it too. In November, M&S, who were already

relating to strawberries only, those who had

using It’s Fresh! for strawberries, raspber-

purchased Morrisons’ raspberries only were

ries and blackberries, became the first retailer

asked questions relating to raspberries only

in the world to use it in pears. With adverse

media coverage, including The Mail on Sun-

and those who had purchased both strawber-

media comment about the 200,000 tonnes

day and The Times in December. The Mail on

ries and raspberries in the last 5 weeks were

of food wasted by supermarkets each year,

Sunday headline says it all - “The tiny gas-

asked both sets of questions. Within three

there is a growing incentive for companies to

guzzling strip that can keep

days the panel had generated a sample of

adopt the technology. So as well as Morrisons

fruit fresh for up to four

almost a thousand who had met the eligibility

and M&S, Tesco, Waitrose and Co-op are now

days’ longer (and cuts

criteria.

using It’s Fresh! and the strip can now be

the food mountain)”.

Customers like the Freshness Strip 54% of respondents had noticed the It’s

and cherries.

And there’s salad too. It is already used for tomatoes and avocados and at the

punnets and 62% for raspberries. Of those

request of retailers FFT is now working

who had noticed a difference to the fruit

on its application to flowers. Simon Lee,

quality in their recent purchases contain-

the founding Director of FFT expects the

ing the Freshness Strip, the overwhelming

strip to work very well with some flowers

majority (93% for strawberries and 94% for

where ethylene is a problem, such as roses

raspberries) thought the quality and freshness

and orchids.

them to buy more of those products in future.

FFT has also benefitted from very positive

found in packs of peaches, nectarines, plums

Fresh! filter in the Morrisons’ strawberry

were better, and thought that this would lead

Media coverage

Tesco ambient salad and avocado technical manager, Steve Deeble, said: “This is a major breakthrough in the fight to combat

www.customer-insight.co.uk | Spring 2016 Customer Insight 19

RESEARCH

I

mprove response rates

and quality of response through effective pre-survey communications.

A well executed warm up campaign will make your survey more cost effective by shorten-

ing the field work period (achieving interviews or responses more quickly) and increasing the response rate. It could also stimulate response from a wider range of customers, generating a more representative result and improve the quality of feedback by making participants more engaged with the survey.

20 Customer Insight Spring 2016 | www.customer-insight.co.uk

RESEARCH

“... selling the benefits of taking part in the research in the warm-up should act as an encouragement to take part” Warming up staff – why it matters

- how the data will be collected (i.e. tele-

organisation. Less satisfied customers are

phone, web etc.), who is conducting the

often those who most welcome the oppor-

hear about the survey is from customers.

research, what time of day contact might

tunity to say what they think. Interest-

This reflects badly on the organisation and is

occur. Establish the credibility of the

ingly, research shows that being invited to

demoralising for staff who wonder why they

agency with staff.

take part in a survey can improve customer

It can be embarrassing if the first staff

have not been informed. In markets with

- an understanding of what custom-

satisfaction. It conveys the message that the

strong customer-supplier relationships staff

ers should expect. Staff do not need to

organisation cares about customers, is not

can also play an important part in encourag-

understand the intricacies of the meth-

afraid to seek feedback and values custom-

ing customers to participate.

odology but an idea of what customers

What to tell staff

will be asked is useful. For example,

ers’ opinions.

• Staff can be afraid of the repercussions from

knowing customers will be asked to give

survey findings. Management should allay

scores and encouraged to make com-

staff fears around how the results will be

tance of the survey and its role in the

ments about products and services. For

shared and handle feedback responsibly,

organisation’s plans. Staff need to know

some industries reassurance that cus-

particularly verbatim comments which may

the survey will measure performance and

tomers will not be asked for personal or

refer to staff by name. Customer feedback

the findings will be used to make improve-

financial information may be important.

can support staff as it often reflects their

• Staff should be in no doubt of the impor-

ments. The survey gives the organisation

•C ustomer facing staff need to understand

an opportunity to highlight to staff the link

how they can encourage participation,

between satisfaction and loyalty and its

e.g. being positive about the benefits for

impact on the organisation.

customers. It is important staff do not try to

• Staff may need to answer questions from customers to give them enough information to deal with queries. This may include:

own thoughts or ideas.

influence customers’ responses by suggesting negative or positive feedback.

•S taff may assume customers will not like being approached to take part. In our expe-

Rachel Allen

survey is starting and finishing, when

rience, customers are happy to take part

Client Manager

the results will be available, how results

if the survey is well designed. Customers

TLF Research

will be shared and what will happen

often tell us they enjoyed the experience,

rachelallen@leadershipfactor.com

afterwards.

even when they are dissatisfied with the

- survey schedules including when the

www.customer-insight.co.uk | Spring 2016 Customer Insight 21

RESEARCH

Warming up customers – why it

What encourages customers to

matters

take part in the survey?

More customers will take part in a more

• Customers want to believe the organisation

committed manner if they understand the

is prepared to make changes from which

purpose of the survey and the potential ben-

they will benefit and that performance will

efits for themselves from taking part. One of

improve their experience based on feedback.

the main reasons warming up matters is that

Assure them of this.

it can help customers overcome the negative

• Customers need to be reassured that giving

attitudes they may have about taking part in

their opinion will require no effort and the

a survey. The reasons can be addressed in the

process will be straightforward. For a tele-

warm up communication.

phone survey this may involve call-backs

What stops customers taking part in the survey?

and appointment setting. For postal surveys this may involve provision of a business reply envelope. This may extend to surveying in foreign languages or providing large

• Suspicion – who are we talking to and where did you get my details from?

•P ersonal details – customers need to be

print versions of questionnaires.

• Customers need to feel that they can say what they think without challenge or reper-

confident that their contact information and

cussions. The interviewer needs to allow

responses, which are confidential per-

the customer to share their views. For other

sonal information, are going to be treated

collection methods, such as post or email

‘properly’. This means assurances about

the customer should be given the opportu-

data protection compliance and the offer

nity to make comments and explain their

of respondent anonymity (which can be assured through the Market Research Society Code of Conduct) •T he introduction – in our experience a wordy introduction puts customers off taking part. Customers do not want to hear a lengthy explanation of why the survey

scores with an open question.

• Customers will be engaged in the survey if the questions are interesting to them. This means asking relevant and unambiguous questions designed around what matters to customers.

• Customers need to know what is going to

matters. They decide to take part quickly

happen with their responses and how their

and want to get on with the questions so if

information will be used. The MRS Code

they’ve already been warmed up a telephone

of Conduct stipulates that data can only be

or online introduction can be very short

used for the purpose for which it was gath-

•T he questions – if the content is irrelevant

ered. This means that it cannot be ‘abused’

or uninteresting customers will opt out. If a

e.g. used for selling or marketing purposes

postal survey is difficult to read or fill in it

(without their explicit agreement).

will not get completed •T he interviewer – a pushy or intimidating interviewer can scare off customers

Opt outs Is it worth giving customers a number to

•P ast experience – ‘Nothing happened last

call so they can ‘opt out’ of taking part in the

time I gave feedback’ or ‘Last time I did a

survey? In our experience, this is unnecessary

survey the company was trying to sell to

and creates additional work because resource

me’

is required to handle incoming communica-

•F ear or lack of confidence – worry that the

tion and update databases. This approach can

survey will take a long time or they won’t

also adversely affect response rates and add

be able to answer the questions.

weeks to the survey schedule. Furthermore, customers may rule themselves out of taking part when a call from a skilled interviewer would have elicited feedback. Organisations sometimes offer ‘opt out’ to prevent upsetting their most valuable customers when these are actually the customers whose feedback they should be encouraging most.

22 Customer Insight Spring 2016 | www.customer-insight.co.uk

RESEARCH

Incentives We do not recommend offering incentive payments apart from for focus group attendance. It can be difficult to decide on an appropriate incentive suitable for all customers. This is further complicated by rules and regulations around what can be offered (e.g. organisations cannot offer vouchers for their own products). B2B customers may be forbidden from accepting incentives. Incentives may also influence customers’ responses although this will be impossible to determine.

Instead, selling the benefits of taking part in the research in the warm-up should act as an encouragement to take part.

What to tell customers?

Use the warm up communication to tell customers: • Why you are conducting a survey and the benefit to them of taking part – e.g. to help you understand what you need to do differently to meet their needs and satisfaction • An outline of the approach – how contact will be made • What you would like them to do and when – e.g. take part there and then, make an

e er Nam Custom Address Address Address Address ber: ce num

Referen

er 21 Octob

211085

2015

ers custom rvice to rd of se order to larly. In le standa ib here gu w ss re s po ed ea t ss those ar e highes lly asse identify ovide th to be impartia ents. to help BC to pr s A ds em er ee at ov n e om pr e im objectiv ormance to enable cust ke to se It is our e that our perf u may li survey ev here yo cting a d so I beli e condu ds and those w ar e sional an w nee a profes do this eet your sultancy d out in m n ie y co rr ll fu ca ch ar be we ill h, a rese eeds to earch w ocess n Researc TLF Res terview l this pr ted TLF behalf. impartia erefore appoin on our phone in d re le ey an ai te rv n a te su on nge ta cura ve th esti carry ou x to arra To be ac ive a qu er. We ha to help e mann in this work, to een xxx and xx by email, rece of your time onses objectiv ed le tw s sp tt ct se re be li ta eciali you in eat your aw to l be con ld appreciate a which sp st to contact you wil h will tr be es. (Or ally wou . TLF Researc would like to dr ut re in e m W do their ) es 10 em if opriate. d servic points that you ciated with th around ucts an lasting , as appr the post ck on our prod any particular your name asso h ug ro e d ba th ar noted an if there ving feed us by gi nce. However, em to be k for th de ove u can as in confi yo , to impr on nti we need e a first our atte s where le us to provid ea h. ar is t ab ligh you w to high ges to en r aim of ke analysed ecessary chan ant step in ou ould li will be t any n s and w import results er en ry e ve th om em , a st pl rvey r cu l im is as il su ou th w l e e al rd th w ga to and After e re ide ormance r customers. W service we prov our perf ou e level of rvice to class se y improving th r your help. ll fo continua u in advance k yo to than er

stom Dear Cu

ly

ncere Yours si

t Brown fficer Emmet O ecutive Chief Ex

appointment for a convenient time or return a questionnaire • Whether they will get feedback about the survey findings. This can be a powerful incentive to take part especially in B2B and special interest markets • What you will do with the survey results – i.e. use the findings when making plans or changes • Who you are and how they can contact you – introduce the agency and provide names and numbers to avoid suspicion • Consider the signatory – the person signing the letter needs to carry weight, conveying the importance of the survey to customers.

www.customer-insight.co.uk | Spring 2016 Customer Insight 23

RESEARCH

How to warm up customers?

There are a number of ways to get the message to customers. Employ a range of

Maximising response rates As well as an effective warm-up campaign, the following factors will further impact response rates:

approaches to get the best reach across your customer base:

•T he quality of the data - poor quality

• The length of the questionnaire – a long

contact data (i.e. data that is incorrect or

questionnaire may affect response. Custom-

incomplete) cannot generate response. This

ers may not respond or may not complete

explaining to customers why, when and how

means if names and address details are

the survey. This can particularly apply to

the survey is going to take place

inaccurate or telephone numbers are wrong

postal surveys. Where a longer survey is

there may be little or no response. This

unavoidable, make customers aware of this

message on a blog can assure customers

includes recognising any time zone differ-

upfront and explain why. Under the MRS

that if they are contacted there is nothing to

ences if customers are overseas. At best,

Code of Conduct any given timescales for

worry about

response will be hard to come by. At worst,

• Letters/email – send out a letter/email

• Social media – a tweet, Facebook post or

• Newsletters – an article or simply a line or

the exercise will be an embarrassing or

two, in a prominent position, can provide

expensive waste of time. The quality of your

relevant information

data may quickly become apparent as you

• Website - a paragraph on the home page or relevant pages as a reminder a survey

plan your warm up campaign.

•T he ‘appropriateness’ of contacts – if the

completion have to be accurate.

• The layout of the questionnaire – wordy questionnaires in a small font, or questions that do not flow will deter customers.

• The interviewer – interviewing is a skill that requires training. Trained interviewers

is taking place. Including a link to the

‘wrong’ person, or an unsuitable person,

can encourage customers to take part in a

agency conducting the research adds further

is contacted they are unlikely to take part.

survey and gather unbiased feedback quickly

credibility

If they do take part, their response will not

and effectively. Poor response rates can be

• Staff – contact centre staff can make customers aware when they call; account managers can personally inform customers • Posters – where customers visit counters, branches or offices, posters can be displayed • Regular mail (such as invoices, receipts, acknowledgements) - include a ‘footnote’ to inform customers

be useful. This particularly applies to B2B surveys where the job role of the respondent may be vital to gathering relevant feedback.

•T he purpose of the survey – if the reason

due to inexperienced interviewers.

• Previous history - if a customer has completed a survey previously (either for your organisation or another) and been misled

for conducting the survey is of little interest

or found their information has been used

or relevance to the customer they will be

inappropriately they may be unlikely to take

unlikely to take part.

part.

•T he ‘characteristics’ of the contacts – some

• Action resulting from previous feedback – if

customers are more difficult to get hold of

a customer has given up time to complete

to an email signature can be a cheap and

than others and this can affect response. In

the survey in the past but had no response,

effective way of building awareness

these cases, it makes sense to allow more

no feedback or seen no change, they will be

• Events – meetings or conferences can

time for the survey. For example, securing

• Email signatures – adding a short message

include a spot of information sharing

Key points:

interviews with customers who are frequently travelling may take longer.

• Your relationship with contacts - where the

less likely to take part again.

• Some response rates will always be on the ‘low’ end of the scale. Busy professionals who are time poor may be less likely to

relationship with customers is perceived to be

take part in a survey than customers with

weak customers will be less likely to respond.

time on their hands. To avoid disappoint-

If an organisation goes for lengthy periods

ment be realistic about response rates,

the field work taking place for customers to

of time without having any contact with

allow adequate time for the survey process

recall being warmed up but in good time for

customers, they may find response is lack-

and consider the data collection method

them to take the message in, usually 7-14

ing. However, it may be important to include

carefully.

days ahead of the field work starting. Com-

these customers in the survey. How will an

munication can continue whilst the survey

organisation understand these customers,

is live in the form of a reminder (along the

gather their feedback or assess their potential

ing very important decisions on the basis of

if it makes no attempt to talk to them?

survey results it makes sense to invest some

• Use a combination of communication methods for a wide reach • Deliver the communication close enough to

lines of ‘Thank you if you have taken part – but if not, we would still appreciate your

• The time (month, day, hour) of the call or

Since organisations will often be mak-

time and budget to ensure that you get the

email – Bank Holidays, national events and

best possible response rate and quality of

popular sporting events can affect response.

response. An effective warm up campaign is

database. There is little sense in warming

E.g. Europeans holiday during August and

an often overlooked but crucial part of this

up some customers and not others. You may

response can be poor during this time. It can

plan.

vary the approach or message slightly to

also be difficult to capture feedback in the run

suit customer types

up to Christmas when customers are busy.

feedback’) • Warm up every customer on the contact

24 Customer Insight Spring 2016 | www.customer-insight.co.uk

RESEARCH

Customers tend to benchmark organisations very widely, comparing them with their experiences across many different sectors.

O

ften customers’ recent

experience is

limited to one organisation per sector.

supply markets, for example, dual or multiple sourcing is widespread. In the leisure sec-

They simply don’t currently deal with more

tor customers often frequent more than one

than one local council, one mortgage lender,

restaurant, tourist destination or theatre. It

one mobile phone provider or one housing

is therefore very useful for some companies

association. At other times, however, custom-

to understand how customers make com-

ers are much more active in making com-

parisons and choices between competing

parisons; when they bought their house and

suppliers.

needed a new mortgage for example or when their annual phone contract or insurance policy was due for renewal. In some markets

6.5

7

7.5

8

8.5

9

9.5

10

Fruit & vegetables

Stock availability

Bakery

Cleanliness

Market perception

Queue times

For companies in competitive markets

customers may habitually use one supermar-

where churn is common, a survey must cover

ket, for example, or drive one make of car for

all the factors that influence customers’

three or four years before replacing it, but are

choice and evaluation of suppliers enabling

nevertheless frequently making comparisons

the company to see how it compares against

between competing suppliers even though

its competitors on all the most important

they are not switching.

supplier selection criteria used by customers.

In other markets customer promiscuity is

(See Figure 1)

much more prevalent. In most industrial

6

Price

Fresh meat

Café

XYZ Ltd

Competitor 1

Competitor 2

Fig 1: Competitor comparisons

www.customer-insight.co.uk | Spring 2016 Customer Insight 25

RESEARCH

Provided the factors have also been scored for importance, a weighted index can be calculated for each supplier, the outcome providing an accurate reflection of how the market perceives the relative standing of the competing brands. Since customers’ attitudes precede their behaviours, Figure 2 will provide a reliable guide to future customer behaviour in the market and its consequent impact on market share so provides a sound basis for decisions about how to improve competitive positioning.

XYZ Ltd

85.8%

Competitor 1

84.9%

Competitor 2

77.5% Fig 2: Market Standing

-2.0

Competitor gaps

-1.5

-1.0

-0.5

0

0.5

1.0

1.5

2.0

1.5

2.0

Stock availability

There are significant differences between

To make the most impact on improving

Price

the satisfaction of its own customers, XYZ

figures 3 and 4. Stock availability is the

should focus on addressing a small number

obvious PFI for XYZ if based on the satisfac-

Queue times

of PFIs (priorities for improvement) based on

tion gaps but queue times are a much bigger

Fruit & vegetables

its biggest satisfaction gaps (i.e. where it is

area of under-performance against Com-

least meeting its customers’ needs. So based

petitor 1. Since in the real world there would

on Figure 3, XYZ should be prioritising stock

probably be at least 20 important customer

Fresh meat

availability.

requirements covered on the survey, and all

Cleanliness

However, in a highly competitive market

Bakery

companies have finite resources, XYZ may

Café

there is also another dimension to consider;

have to make choices between increasing the

XYZ’s relative performance compared with

satisfaction of its own customers or clos-

its main competitors, especially its biggest

ing the gaps with Competitor 1. Putting the

threat, Competitor 1. Figure 4 shows the

two sets of data together into a competitor

competitor gaps between XYZ and Competi-

matrix would be a useful start for making this

Queue times

tor 1.

decision.

Price

Fig 3: Satisfaction Gaps for XYZ

-2.0

-1.5

-1.0

-0.5

0

0.5

1.0

Cleanliness Stock availability

2 Meat

1.5

Bakery

Stock Availability

Fruit & vegetables

1 Café

Price 0.5

Fig 4: Competitor gaps: XYZ versus Competitor 1

Queue times

Satisfaction gaps

Requirements closest to the top left hand

Fruit and vegetables

0

corner of Figure 5 represent XYZ’s main areas Fresh meat

-0.5

of weakness, in terms of failing to satisfy

Bakery

Cleanliness

its own customers and under-performing its main competitor. Whilst ‘stock availabil-

-1

ity’ would emerge as XYZ’s main PFI based on measuring the satisfaction of its own

-1.5

customers, the data from across the market suggests that improving ‘queue times’ would make a bigger difference to XYZ’s market

Cafe -2

-2

-1.5

-1

-0.5

0

0.5

Competitor gaps Fig 5: Competitor matrix

26 Customer Insight Spring 2016 | www.customer-insight.co.uk

1

1.5

2

position against Competitor 1.

RESEARCH

Switching The main characteristic of very competi-

e.g. visiting a new restaurant ‘for a change’.

A competitor analysis must identify the

tive markets is the prevalence of switching.

Hofmeyr1 calls this ‘ambivalence’ and points

customers most and least likely to switch2.

Customers see changing from one supplier

out that in some markets customers are

This should include the company’s own cus-

to another as relatively easy so often feel it

loyal to more than one supplier. They will

tomers and competitors’ customers since the

is worth switching for even a small increase

sometimes visit a different restaurant even

company must understand how to defend its

in satisfaction. They may even switch just to

though they are completely satisfied with

own vulnerable customers as well as how to

find out if an alternative supplier is better,

their favourite restaurant. In some markets

target and attract its competitors’ most vul-

since it is easy to switch back if it isn’t. In

therefore, companies need a much deeper

nerable customers. This is done by dividing

highly competitive markets this promiscu-

understanding of customers’ loyalty attitudes

one’s own and the competitors’ customers

ity reaches its height when customers switch

and behaviour.

into loyalty segments as shown in Figure 6.

simply for a different customer experience,

Faithful

Vulnerable

Flirtatious

Available

Our customers

Competitors’ customers

Strongly loyal, rate our performance highly,

Strongly loyal, rate competitor highly, little

little interest in competitors

interest in us

Apparently loyal customers but high level of

Repeat buyers with competitors but little

inertia or some interest in competitors

positive loyalty and some interest in us

Little positive loyalty, actively interested in

Little loyalty to competitors, may be

alternatives

receptive to our advances

Customers showing a strong preference for

Competitors’ customers who already rate us

alternative suppliers

as superior to their existing supplier

Fig 6: Loyalty segments

Few companies have the resources to successfully implement acquisition and retention strategies across all segments. Figure 7 illustrates the situation for a supplier with one competitor but in a very promiscuous market there will be several competitors, each with their own strengths, weaknesses and customer profiles. The starting point for strategic decisions on retention and acquisition strategies is therefore to understand the distribution of the customer base across the four loyalty segments.

Our customers

Faithful

Vulnerable

Flirtatious

Available

Competitors’ customers

Reward loyalty, stimulate referrals, strong focus on service recovery factors

Don’t target

Strong focus on PFIs, communications

May be worth targeting if competitors are

campaigns and loyalty schemes to build

failing to meet their need in areas where you

positive loyalty

perform strongly

Objective assessments of costs and benefits of

Go for the jugular, especially where

retaining this group. Strong focus on closing

you believe your strengths match their

any perception gaps

priorities

Cut losses. Chances of retention very low

Should be easy prey but make sure they’re not habitual switchers

Fig 7: Loyalty segments

www.customer-insight.co.uk | Spring 2016 Customer Insight 27

RESEARCH

0%

10%

20%

30%

40%

50%

60%

70%

0%

10%

20%

30%

40%

50%

Figure 8 depicts a company with a very secure customer base, which should take

Faithful

Faithful

steps to reward and protect the loyalty of its many faithful customers, whilst implementing strong measures to attract any of the competitors’ available and flirtatious customers, provided they have a suitable needs

Vulnerable

Vulnerable

profile.

By contrast, the supplier shown in Figure 9 has a customer base that is typical of a company devoting too much resource to winning

Flirtatious

Flirtatious

new customers at the expense of satisfying and retaining its existing ones. This company needs to seriously re-think its strategic priorities. An example is the MBNA case study from Harvard, where the company was not

Available

Available

Fig 8: Secure customer base

Fig 9: Disloyal customer base

keeping its customers long enough for them to become sufficiently profitable. MBNA’s ‘zero defections’ strategy based on delivering exceptionally high levels of service to targeted customers, moved the company from the 38th to the largest bank card provider in the USA over two decades3,4.

Segmentation To optimise strategic decisions of the type

heavily influenced by demographic factors.

outlined in Figure 7, a company must seg-

In others, such as groceries and cars, a more

ment customers and build detailed profiles of

complex level of attitudinal and psycho-

the predominant types of customer in its own

graphic profiling is often necessary to fully

and its key competitors’ loyalty segments.

understand the differences between loyalty

One of the earliest academic authorities

segments. These may include core values

on customer segmentation was Yoram Wind5,

such as the importance placed on individual

who suggested some less commonly used

liberty, health and fitness and family values

segmentation variables, which, in his view,

or deeply held beliefs such as commitment to

often provided more insight than standard

the environment, fair trade food or specific

classification data such as demograph-

political or charitable causes. Sometimes, the

ics. Wind’s preferred segmentation criteria

best way to profile customers is to start with

included:

their tangible behaviour such as when they buy, how they buy (channel), how often they

• Needs Segmentation (called benefits segmentation by Wind) • Product preference • Product use patterns • Switching behaviour • Risk aversion

“For many companies, satisfying and retaining their existing customers is only half the battle”

buy and how much they buy, then search for demographic, psychographic or geographic

a small range of TV channels. Rather than

differences within the behavioural segments.

asking its customers direct questions about

Loyalty personality

their behaviour in its own market (e.g. likelihood of renewing their policy), an insurance

For even more intelligence, a company

company might ask about their media usage

• Deal-proneness

can often draw insightful conclusions about

and shopping behaviour. Customers that use

• Media use

customers’ loyalty by asking them ques-

a very small range of media and are highly

• Loyalty behaviour

tions about their behaviour in other walks

loyal to one supermarket for their grocery

of life. Media usage is an obvious example.

shopping are displaying a more favourable

Some people are promiscuous users of media,

loyalty personality than those who often

include demographic, geographic, behav-

hopping across many TV, radio and inter-

shop at three or four different supermarkets

ioural, lifestyle and psychographic details.

net channels, whilst others may get their

and have very diverse media habits. What-

In some markets, such as pensions or health

information and entertainment from one

ever they say about their intentions to renew

care, customers’ attitudes and behaviours are

newspaper, one or two radio stations and

their policy, customers demonstrating strong

In B2C sectors classification data can

28 Customer Insight Spring 2016 | www.customer-insight.co.uk

RESEARCH

Fig 10: Decision free analysis 100%

loyalty behaviours in other markets are more

1

likely to be loyal insurance customers.

81.3

In B2B, readily available demographic

Over 55

criteria include company size, usage volume

Under 55

and industry segment but B2B companies 46%

can often gain an insight advantage over

54%

2

competitors by adopting the kind of loyalty

3

92.4

personality techniques described in the previous paragraph. A good example would be risk

Still Working

aversion. De Bruicker and Summe classified buyers into segments such as ‘inexperienced

74.8

Retired

12%

Without Children

34%

generalists’ and ‘experienced specialists’6.

4

The former are less knowledgeable and less

With Children 10%

5

88.6

44%

6

95.1

79.5

7 69.3

confident and therefore less price sensitive but much more demanding in terms of

ABC1

C2DE

support. By contrast, experienced specialists are less interested in customer service and relationships but will unbundle and haggle over all aspects of a purchase.

5%

Income up to £20000 p.a

7%

25%

10 84.9

Decision tree analysis

Income over £20000 p.a

11 90.4

9

96.4

92.7

Outside London / South East

A good way of using survey data to identify

9%

8

London / South East 19%

6%

12

loyalty segments, decision tree analysis iden-

13

97.2

tifies the biggest differences between seg-

94.0

ments by sequentially dividing a sample into a series of sub-groups with each split chosen because it accounts for the largest part of the

This makes it possible to profile the most

References

remaining unexplained variation. The easiest

secure customers, the most flirtatious or

1. Hofmeyr, Jan (2001) “Linking loyalty

way to understand this process is to work

any other loyalty segment. The company

measures to profits”, The American

through the decision tree shown in Figure 10.

concerned would be well advised to target

Customer Satisfaction and Loyalty Con-

(See Figure 10)

retired over 55s on modest incomes outside

ference, American Society for Quality,

The process starts with the entire sample,

London. As well as having very high levels of

Chicago

indicated by the 100% above the first box,

satisfaction with the benefits delivered by the

ice and Hofmeyr (2001) “Commit2. R

which is numbered 1 in its top right hand

company they also account for a sizeable 19%

ment-Led Marketing”, John Wiley &

corner. The 81.3 refers to the customer sat-

of customers in the target market.

isfaction index for the sample in question.

This could be the entire sample, or, more usefully a sub-set of it, such as the ‘flirta-

Sons, New York eskett et al (2003) “The Value-Profit 3. H

Conclusion

Chain”, Free Press, New York

For many companies, satisfying and

4. H eskett et al (1997) “The Service-Profit Chain”, Free Press, New York

tious’ segment, or a competitor’s ‘available’

retaining their existing customers is only half

segment. The data examined does not have

the battle. In a highly competitive market a

to be overall satisfaction. It could be a loyalty

company could be achieving high levels of

index, a Net Promoter Score or an individual

customer satisfaction but still losing market

factor such as ‘quality of advice’. The process

share if a competitor is acting on better brand

then looks for the single dichotomous vari-

perception and competitor benchmarking

sure your customers keep com-

able that accounts for the biggest difference

insight. A competitor who understands the

ing back”, Harvard Business Review,

in satisfaction variation across the sample

priorities and profiles of a rival’s flirtatious

January-February

and, in this example, finds that it is age. It

and available loyalty segments and invests in

can split any variable into only two groups

delivering and promoting appropriate ben-

at each stage, and in this example the two

efits to them will always have the edge. To

Nigel Coxon

age segments that account for the biggest

succeed, companies in competitive markets

Business Development

variation across age groups are over- and

must therefore use surveys to understand

Manager

under-55s, which now become boxes 2 and 3.

market perception and should develop loyalty

And so it goes on.

segments for their own and competitors’

5. W ind, Yoram (1978) “Issues and Advances in Segmentation Research”, Journal of Marketing Research, August 6. D e Bruicker and Summe (1985) “Make

TLF Research nigelcoxon@leadershipfactor.com

customer bases.

www.customer-insight.co.uk | Spring 2016 Customer Insight 29

CONFERENCE

Financial Services Complaints Management Forum

A

t the start

of the 14th annual Financial Services Complaints

Management Forum the FCA provided an update on their

complaints reporting requirements, which will be implemented in June 2016.

These are the main points:

Santander) mentioned they had found it

complaints from those who do complain and

• An extension of the ‘next business day

valuable to carry out scenario testing inter-

making changes to alleviate this. It is also

rule’, where firms are permitted to handle

nally – i.e. to test that complaint handling

critical that it is made easy for customers to

complaints less formally, without sending

personnel would approach the same case in

complain.

a final response letter, to the close of three

the same way, so that they know they are

business days after the date of receipt.

delivering outcomes consistently.

• A requirement to report all complaints,

•M ore guidance on the definition of a com-

He also talked about the importance of culture and empowering staff to deliver good customer service, with the ultimate focus

including those handled by the close of

plaint would be welcomed. It seemed the

being on the customer rather than compli-

three business days. (Previously there was

case that most companies were recording all

ance, as summarised in the Best Practice

no requirement to report those handled

expressions of dissatisfaction as complaints

Complaints Model.

informally under the ‘next business day

(rather than just those that involve a mate-

rule’).

rial loss) due to the ease of delivering this

tive analytics and how RSA had used their

message to frontline staff.

existing data to predict future behaviour and

• A requirement for financial services companies to raise consumer awareness of

•T hose on the panel mentioned they had

Roger Binks from RSA talked about predic-

trends. He gave some examples they had

the Ombudsman service, by sending the

greatly reduced and simplified the com-

found using their home claims data where

‘summary resolution communication’ to

plaint categories recorded – and this was

customers who called more than twice had

complaints handled less formally under the

beneficial in producing more insightful root

a 70% chance of later making a complaint.

‘three business days rule’ as well as to the

cause data.

Also, for storm and flood claims specifically, where the claim had been open for 6 weeks

more serious complaints whose resolution

and there had been no proactive contact

took longer. This explains to consumers

Stephen Hampshire from TLF talked about

how to appeal to the Ombudsman if they

the fact that the focus on compliance and the

from RSA, there was an 80% chance of that

are not satisfied with the outcome of their

avoidance of risk has meant that companies

customer making a complaint. RSA then used

complaint.

are unlikely to deliver good customer service.

this data to change their process and intro-

This was echoed in the FCA paper and the

duced outbound calls within this period which

Swiss Re case study later presented.

reduced the number of complaints logged.

• New rules limiting the cost of calls consumers make to firms to a maximum ‘basic rate’, including all post-contractual calls and all complaints calls.

He reminded us that customer complaints are only the tip of the iceberg and that not all customers who experienced a problem will

Iain Law

The panel discussion around putting the

make a complaint. Typically, between 1 in 4

Client Manager

FCA’s thematic review into practice shared

and 1 in 5 will not make a complaint. This

TLF Research

some practical insights for businesses:

highlights the importance of pre-empting

iainlaw@leadershipfactor.com

• Those on the panel (Barclays, Lloyds and

complaints by understanding the causes of

30 Customer Insight Spring 2016 | www.customer-insight.co.uk

CUSTOMER

The UKCSI is at its highest point since January 2014 The UK Customer Satisfaction Index, the national measure of customer satisfaction based on over 39,000 customer responses, 75.2 stands at 77.0 (out of 100) in January 2016, up 0.8 points 74.1 compared to July 2015 (76.2) and up one point 72.0 compared to January 2015 (76.0).

Jan-09

Jul-09

Jan-10

78.0 77.3

78.2

77.9

77.4

77.1

77.0

76.7 76.3

76.0

76.2

75.6

Jul-10

Jan-11

Jul-11

Jan-12

Jul-12

Jan-13

Jul-13

Jan-14

Jul-14

Jan-15

Jul-15

Jan-16

Covering 13 sectors of the economy, the UKCSI is conducted by TLF Research for the Institute of Customer Service.

The biggest year on year sector increases are for Utilities (1.9 points), Public Services (National and Local) (1.7) and Insurance (1.6). The scores for Retail (Food), Transport & Telecommunications & Media have increased by at least one point. The other sectors have remained relatively flat with only Banks & Building Societies seeing a slight drop (down 0.4 points).

That’s how to promote the fact that you’re the UK’s most highly regarded companies for customer satisfaction.

First, they’re feeding it back to customers to make sure they notice and second they’re making customers feel great with Amazon offering a discount worth up to 20% and Utility Warehouse offering £25 retail vouchers for existing customers taking a second product. Very clearly they’ve used the news as an opportunity to make customers even more satisfied than they were before.

www.customer-insight.co.uk | Spring 2016 Customer Insight 31

CUSTOMER

The Top Scoring Organisations in UKCSI Jan 2016 rank

Organisation

January 2016 score

January 2015 score

January 2015 rank

Change in score Jan 2015 - Jan 2016

1

Amazon.co.uk

86.6

86.7

2

-0.1

2

Utility Warehouse

86.4

NO DATA

3

first direct

85.7

86.7

2

-1.0

44

4

Specsavers

85.4

81.9

19

3.5

5

Waitrose

85.0

83.5

7

1.5

organisations registered a fall of one point or more

6

John Lewis

84.9

87.2

1

-2.3

7

New Look

84.3

79.1

63

5.2

8

SAGA Insurance

84.1

79.9

47

4.2

9

Nationwide

83.7

83.8

6

-0.1

10

M & S (food)

83.5

83.9

5

-0.4

11

Aldi

83.4

83.0

10

0.4

12

Mini

83.3

NO DATA

12

Trailfinders

83.3

NO DATA

14

Virgin Atlantic

83.2

80.3

34

2.9

14

Iceland

83.2

82.3

17

0.9

16

Hyundai

83.1

80.8

27

2.3

16

Ocado

83.1

85.4

4

-2.3

18

Wilkinson

82.9

80.0

42

2.9

19

Superdrug

82.8

81.8

21

1.0

20

Boots UK

82.7

80.8

27

1.9

organisations improved by one point or more

Exceeded 80 in every UKCSI since 2013 Amazon.co.uk Waitrose John Lewis Nationwide M&S (food) Aldi Iceland Hyundai Boots UK Greggs/Baker’s Oven LV= M & S (Non-food) Premier Inn ASDA Kia Argos Škoda

Top 50 Organisations in UKCSI (scores out of 10)

The areas of greatest difference between the top 50 organisations and the rest are focused largely around

Staff doing what they say they will do

96

Difference

Remaining Organisations

1.8

7.0

Speed of resolving your complaint

6.6

The handling of the complaint

6.7

5.2

1.7

4.9

1.8

4.9

people measures, complaint handling and speed of service.

Sarah Stainthorpe Client Manager

Staff understanding the issue

1.6

6.9

The attitude of staff

6.9

The outcome of the complaint

6.7

5.3

1.5

5.4

1.4

5.3

TLF Research sarahstainthorpe@leadershipfactor.com

The ease of getting through (phone)

32 Customer Insight Spring 2016 | www.customer-insight.co.uk

8.2

1.1

7.1

CUSTOMER

Top 20 Improved Organisations 20 most improved organisations - one year

Sector

UKCSI score January 2016

UKCSI score January 2015

Change Jan 2015 -Jan 2016

T-Mobile

Telecommunications & Media

75.6

66.6

9.0

Odeon

Leisure

80.4

71.6

8.8

Cineworld

Leisure

80.9

72.9

8.0

The Co-operative Food

Retail (Food)

79.9

72.3

7.6

Jaguar

Automotive

81.6

74.4

7.2

Nationwide

Insurance

80.0

73.5

6.5

Ryanair

Transport

70.5

64.1

6.4

Zurich

Insurance

76.4

70.2

6.2

Abellio Greater Anglia Trains

Transport

69.7

63.5

6.2

GiffGaff

Telecommunications & Media

82.1

76.1

6.0

Southeastern Trains

Transport

69.5

63.6

5.9

Haven Holidays

Tourism

81.2

75.8

5.4

Peugeot

Automotive

79.4

74.1

5.3

New Look

Retail (Non-food)

84.3

79.1

5.2

TNT

Services

75.5

70.3

5.2

AA Insurance

Insurance

81.4

76.3

5.1

Volvo

Automotive

80.1

75.6

4.5

Homebase

Retail (Non-food)

78.5

74.1

4.4

More Th>n

Insurance

80.3

75.9

4.4

M&S

Insurance

76.8

72.6

4.2

SAGA Insurance

Insurance

84.1

79.9

4.2

T-Mobile has seen the biggest increase – nine points - since January 2015. With the exception of Nationwide (Insurance) and Zurich, the most improved organisations are not drawn from the sectors that have seen the biggest improvements. Ryanair has continued its improving trend, though it remains below the Transport sector average (73.5).

The most significant improvements in these organisations’ scores are in customers’ online experience, speed of service generally and especially in writing and measures associated with complaint handling.

Customer Satisfaction & Sales Growth Organisations with higher customer

with a UKCSI less than the sector average.

satisfaction in the Retail (Food) sector on

Lidl and Aldi have continued to post the

average achieve better sales growth and

strongest sales results, each exceeding 17%

market share. Organisations with a UKCSI at

year on year growth for the 12 weeks prior

least one point higher than the sector average

to the 11 October 2015. Iceland, Waitrose,

achieved average sales growth of 7.6%

Sainsbury’s and The Co-operative Food also

compared to a drop in sales of 0.4% for those

achieved positive sales results.

-1%

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

7.6%

3.8%

-0.4%

Food retailers with UKCSI at least 1 point above average Food retailers close to sector average (+ / - one point) Food retailers with UKCSI at least 1 point below sector average

www.customer-insight.co.uk | Spring 2016 Customer Insight 33

CUSTOMER

Navigating the multi-channel environment This section examines how customers use different channels, how this

Customers tend to be more satisfied when using websites or dealing

varies by sector and type of experience and the implication for channel

“in person” with organisations, but there are a number of striking

strategy and customer satisfaction.

differences by sector.

Whilst customers use a diversity of ways to interact with

Where customers are checking account information, getting a quote,

organisations, “in person” (46.9%), “website” (22.6%) and “over the

making an enquiry or purchase multi-channel usage tends to attract

phone” (20.2%) are the most popular channels. Social media, text and

relatively high levels of satisfaction. However, where customers have

webchat usage is most marked by 18 – 24 year olds, comprising 2.1%

experienced a problem or complaint, multi-channel use may signify

of primary interactions with organisations for this age group.

dissatisfaction and unresolved issues.

Percentage of customers using each channel as their primary means of interaction (by age group) 0

20

40 4.0%

46.9%

UK all-sector average

60

80

22.6%

100 4.9%

20.2%

18 to 24 25 to 34 35 to 44 45 to 54 55 to 64 65 and above In person

Over the phone

Webchat

In writing

Email

App

On their website

Text

Social media

On average, customers are more satisfied when using websites or “in person” as their primary way of interacting with organisations.

But there are a number of striking differences by sector: Most used channel (% of customers using this as their primary means of interaction)

Channel with highest satisfaction (0 - 100)

Largest gap in channel satisfaction compared to UK average

UK all-sector average

In person

46.9%

Website

78.8

N/A

N/A

Automotive

In person

60.3%

In person

80.2

Email

+5.0

Banks & Building Societies

In person

45.4%

App

80.6

In writing

+4.6

Insurance

Over the phone

43.7%

Website

79.7

Over the phone

+5.4

Leisure

In person

81.8%

In person

79.5

In writing

+5.8

Public Services (Local)

In person

65.8%

In person

77.2

Website

-11.3

Public Services (National)

In person

59.5%

Website

76.6

Over the phone

-9.8

Retail (Food)

In person

80.4%

Website

84.8

Website

+6.0

Retail (Non-food)

In person

51.3%

Website

85.9

Website

+7.1

Services

In person

55.2%

In person

80.0

Website

-4.5

Telecommunications & Media

Over the phone

48.4%

In person

78.0

Webchat

-5.6

Tourism

Website

45.0%

Website

81.0

Over the phone

+6.0

Transport

In person

45.8%

Website

77.3

In person

-6.9

Utilities

Over the phone

41.1%

Website

75.0

In person

-4.3

34 Customer Insight Spring 2016 | www.customer-insight.co.uk

CUSTOMER

Customers’ channel usage and satisfaction by sector: key take-outs

Sectors

• Automotive • Leisure

• Retail (Food) • Retail (Non-food)

Customers’ channel usage and satisfaction

Key implications

“In person” is the most frequently used channel and that which attracts the highest customer satisfaction.

A strong need for direct, personal engagement with customers.

Whilst “in person” is the most frequently used channel, website interactions achieve, on average, higher customer satisfaction.

• Develop employees’ customer service and emotional intelligence skills and measure the impact on customer satisfaction • Measure the quality of customer experience across channels.

• “In person” is the most widely used channel • Transport • Banks & Building Societies

• There is evidence that use of Apps, particularly in Banking, can boost customer satisfaction, albeit with, at present, relatively small numbers of customers • A major concern for Transport organisations is that “in person” - the most used channel by customers – shows the highest negative gap with the UKCSI average

Develop employees’ customer service and emotional intelligence skills and measure the impact on customer satisfaction.

• “In person” is the most used channel • Public Services (National) • Public Services (Local)

• Public Services (National) scores highest for website interactions, lowest for “over the phone”

The requirement to improve customer experiences in the core “in person”, “over the phone” and “in writing” channels.

• Public Services (Local) scores highest for “in person” interactions, lowest for “in writing”

• “Over the phone” is the most common method for interacting with organisations, but customer satisfaction is 3.3 points lower than the UK average for that channel Telecommunications & Media • Webchat was used in 2% of customer interactions, but average satisfaction was 64.8, the lowest of any channel in the sector and 5.6 points below the UK average for this channel

• Improving telephone-based customer service is essential in this sector • Webchat needs to complement rather than replace the core channels through which customers interact with organisations.

• “Over the phone” is both the most used channel and amongst those where the sector’s performance is closest to UK average customer satisfaction (1.4 points below) Utilities

• Website interactions earn the highest levels of customer satisfaction, though this is 3.8 points below the UK average for this channel

A need to develop a consistent and integrated experience across all channels.

• “In person” interactions (74.3) are second only to website for customer satisfaction but are 4.3 points below the UK average for this means of customer interaction

www.customer-insight.co.uk | Spring 2016 Customer Insight 35

BOOK

BOOK REVIEW

What could that be? The clue is that it’s something to do with ‘Growth Mentality’.........

It’s also the title of Matthew Syed’s new book, just published in September 2015. Matthew is the only keynote speaker who has ever been invited back to TLF’s Customer Experience Conference because the first time round he was given the best ever audience satisfaction scores and reviews. Until this time. Scores, comments and tweets from the delegates couldn’t praise him enough. So what’s so good about Matthew Syed. The one word answer is hope. He gives everyone hope that whatever their current circumstances, whatever their future objectives they have the opportunity and ability to shape their destiny and hit their goals. And it’s not because he comes out with motivational platitudes based on whatever the latest positive thinking flavour of the month is. His assertions are grounded in fact, science and published academic research. For those who didn’t see Matthew the first time round, let’s start with a bit of background.

Practice Currently a columnist and feature writer for The Times, Matthew was the

“Creative breakthroughs always begin with multiple failures.”

England table tennis number one for almost a decade, three times Commonwealth Champion, and twice competed for Great Britain in the Olympic Games. The origin of his theories and research into human performance was the seemingly remarkable coincidence that when Matthew was playing, over half of the top players in the UK were from his home town of Reading. But not from anywhere in Reading, they all lived on the same street as Matthew. Now it transpires that on that same street was a table tennis club, with its own building where the tables were permanently erected. The club’s coach, one of the best in the country, also happened to be a teacher at the local school and he gave the children access to the club to practice before school, after school, weekends, holidays, and, of course, gave them a lot of coaching. This meant that Matthew and his young friends all met the main criteria for world class performance: 1. A huge quantity of practice (as a very broad generalisation a minimum of 10,000 hours is needed) 2. More importantly, very high quality practice, informed and shaped by high

Nigel Hill

quality feedback.

TLF Research

The talent myth For many years it was conventional wisdom that people who excelled, whether table tennis players, footballers, pianists or chess grand masters did so because they had talent – attributes such as fast reaction times, hand eye co-ordination, memory etc that gave them a big advantage in their chosen field. However, it has now been widely demonstrated that the attributes are not bestowed at birth but developed through a huge amount of high quality practice.

36 Customer Insight Spring 2016 | www.customer-insight.co.uk

BOOK REVIEW

Indeed, it has been shown that talent, especially the belief that you’ve got it, tends to be detrimental. In one test, young people were given some tasks to perform and regardless of outcome were told they had done very well.

However, half were told that as they’d done so well they must be very talented whilst the rest were told they must have worked really hard. Over a few weeks they were given more tasks but were allowed to choose which tasks they undertook. Two very interesting outcomes followed.

The answer is that they are all Black Box Thinkers.

Firstly, the talented half tended to choose easier tasks. In

Whether they’re developing a new product, honing a core

their desire to protect their status of being talented they

skill or just trying to get a critical decision right, what

were very risk averse and didn’t want to fail. The hard

all these Black Box Thinkers have in common is that

workers chose more difficult tasks as they assumed that

they aren’t afraid to face up to mistakes. In fact, they

even if they failed sometimes, hard work would enable

see failure as the best way to learn. Rather than deny-

them to improve and gradually become good enough to

ing their mistakes, blaming others or attempting to spin

succeed at more difficult tasks. And this they duly did. By

their way out of trouble, these institutions and individuals

the end of the experiment, the hard workers had signifi-

invest considerable time and effort into scrutinising their

cantly improved their performance but, on average, the

errors and failures and exploring every angle for how they

attainment of the talented bunch had declined. It is now

could do things differently next time. They see this as an

widely accepted that for children, and especially boys, it

important part of their future strategy for success.

is detrimental to make them think they’re talented, but to always associate success with hard work.

Digital memory span tests (repeating a string of numbers in sequence) have shown that where people start on the bell-shaped curve has no predictive capacity for the extent to which they can improve. It’s all down to the

Black Box Thinking

quantity and quality of practice, the latter being deter-

Highlighting some of the examples featured in his new

mined mainly by quality of feedback and a person’s belief

book, Matthew poses the following questions:

that they can use it to improve. In other words being shown what you can do to be better next time. Matthew

• What links the Mercedes Formula One team with • What is the connection between Dave Brailsford’s Team

Google, Team Sky and the Mercedes Formula One team all share a ‘growth mentality’ culture with a positive

Sky and the aviation industry?

• What links the inventor James Dyson and the basketball player Michael Jordan?

calls this moving from a ‘fixed’ to a ‘growth’ mentality. In the book, Matthew points out that the likes of

Google?

attitude towards failure. He describes how small changes in culture can have a huge benefit across many different types of institutions and concludes that these lessons can be applied across every field from sport to education, from business to health.

Someone who really should know about learning from mistakes and turning failure into success is James Dyson.

This is his verdict on Black Box Thinking. “Creative breakthroughs always begin with multiple failures. This brilliant book shows how true invention lies in the understanding and overcoming of these failures, which we must learn to embrace.”

www.customer-insight.co.uk | Spring 2016 Customer Insight 37

QUICK GUIDE

· The July-August 2010 issue of the Harvard Business Review featured an article, by Dixon et al, entitled Stop Trying to Delight Your Customers. · Over a three year period, the authors had conducted research, with 75,000 B2B and B2C customers about their recent interactions with staff and self-service contact centres. Based on the

“making it easy for customers to resolve questions and problems is a key way to enhance customer loyalty”

research, the authors made some interesting assertions about the link between being easy to deal with and loyalty. · The authors went on to identify five loyalty building tactics that they · The authors discovered, through their research, that delighting customers doesn’t build loyalty. However, reducing the effort that customers need to put into getting their requests handled and problems solved does increase loyalty. This is understandable.

believed every organisation should adopt: - Reduce the need for repeat calls by anticipating and dealing with issues.

- Provide staff with the tools to deal with the emotional side of customer interactions.

· There is little point in organisations attempting to delight a customer

- Reduce the need for customers to switch service channels.

by making an extravagant gesture if they fail to get the basics right.

- Gather feedback from disgruntled customers and use it.

And, it’s fair to say, most organisations do not consistently deliver

- Focus on problem solving and not speed.

on the basics. It’s even fairer to say that, when things go wrong, most organisations do not handle problems and complaints well.

· There were some other interesting observations made in the article: - Consumers are drawn to companies that offer quality products at a

· Think of it like this, free deliveries would not work well for Amazon if the books that you ordered were continually out of stock or delivered late. In a nutshell, what the research shows is that there is nothing to be gained from making over the top gestures if you are not satisfying your customers when it comes to the basics.

good value. Most often, customers defect when the company fails to deliver on customer service.

- When customers reported that a “very low effort” was required to have their problem resolved: 94% intended to repurchase and 88% said they would increase their spending.

- When the customer considered the effort to resolve the problem as

· Based on their findings, the authors developed a new metric: the Customer Effort Score. The authors claim that the Customer Effort

“very high,” a whopping 81% intended to spread negative word of mouth.

Score is a better predictor of loyalty than the Net Promoter Score. · The findings indicate that “making it easy” for customers to interact · There is some debate over the best wording for the Customer Effort

with customer service departments to resolve questions and problems

question but at TLF we believe that you should adopt a flexible

is a key way to enhance customer loyalty. When staff are properly

approach depending on factors such as industry or type of transac-

trained and possess the information, tools and authority this takes

tion if it is a transactional survey. However, as a generalisation a

away the need for customers to make multiple calls and makes it

good question for a simple transaction would be:

faster for customers to deal with the organisation.

On a scale of 1 -10 how much effort did you have to put into getting your query resolved with Company X, where 1 means a lot of effort and 10 means no effort at all? ·T he study uncovered three primary reasons customers have to

Rachel Allen

expend extra effort/work hard to get their problem or question

Client Manager

resolved. These are having to:

TLF Research

-P lace multiple calls to address the same issue.

- Repeat the same information to multiple contact points.

-S witch from one service channel to another.

38 Customer Insight Spring 2016 | www.customer-insight.co.uk

rachelallen@leadershipfactor.com

FREE QUESTION

IN OUR MONTHLY CONSUMER SURVEY TLF Panel are offering you a FREE no strings attached question in our monthly consumer survey (worth £245) - giving you access to the views and opinions of our panel of over 55,000 UK individuals.

SOUNDS GREAT - HOW DO I GET MY QUESTION? Just email UK@leadershipfactor.com to reserve your question WHAT QUESTION SHOULD I ASK? You don’t need to know the question you want to ask - we can offer some suggestions if you’re unsure.

THE OMNIBUS SURVEY IS USED FOR PR - HEADLINE AND ‘NEWS’ CREATION

CONSUMER ATTITUDE TESTING

CONCEPT TESTING

BRAND RESEARCH

ADVERTISING TRACKING

PRICING SURVEYS

WHAT RESULTS WILL I GET? The survey will be completed by at least 2,000 members of our panel and the results from your question will be split by age, gender and region.

When we close the survey we will email you the results and they’re yours to keep.

Email UK@leadershipfactor.com to reserve your question TLF PANEL RATE CARD

THE PRICE INCLUDES:

NUMBER OF QUESTIONS

NUMBER OF RESPONSES

PRICE (EXCLUDING VAT)

10 15 20 10 15 20

1000 1000 1000 2000 2000 2000

£1,100 £1,300 £1,600 £1,750 £1,900 £2,100

MONTHLY OMNIBUS SURVEY NUMBER OF QUESTIONS 1 3

QUESTIONNAIRE SCRIPTING & DESIGN ADVICE RESULTS ANALYSIS DATA TABS FOR 5 SPLITS: OVERALL, GENDER, AGE, CITY, REGION ANY NECESSARY SCREENING ALL INCENTIVES

NUMBER OF RESPONSES

PRICE (EXCLUDING VAT)

2000 2000

£245 £445

EARNING, KEEPING & RESTORING

TRUST HALF DAY BRIEFING

THE BRIEFING INCLUDES

Trust, within organisations and between organisations and their customers, is almost universally held to be crucial. But what is it?

We know it when we feel it, but trust can be surprisingly hard to define with precision. Even harder is to be clear about the behaviours that build or erode trust.

The psychology of trust in relationships How to measure trust Qualitative insight to explore trust

In this briefing we look at what trust means, give an overview of some different models of trust that have been advanced, and identify the behaviours that are crucial to creating trust inside and outside a business.

The trust equation The ABCDs of trust How trust pays

LONDON ETC Venues One Drummond Gate Victoria London SW1V 2QQ May 11th 2016 09:15 - 12:30

LONDON ÂŁ155 (ex VAT)

ETC Venues One Drummond Gate Victoria London SW1V 2QQ July 12th 2016 09:15 - 12:30

Call Charlotte on 01484 467004 or book online at www.tlfresearch.co.uk