Retail Big Data. How IoT and Cashier Data can improve Retails finance performance with examples.
CATWING
Data Science
Retail
IoT
@SergiSergiev sergiev.sergi@gmail.com in/sergiev
Overview
Technologies
Retail Pyramid
Demo
Technologies
Current situation
Cash register
Product list/catalogue
Warehouse/Inventory
Open workbooks
Limited by available memory and system resources
Total number of rows and columns on a worksheet
1,048,576 rows by 16,384 columns
Total number of characters that a cell can contain
32,767 characters
Other sources IoT
WiFi
Analytics
External data
Bluetooth Analytics
Door
Counters
Sensors
WiFi analytics Positive • Passive interaction •
Customer behavior
•
Venue metrics
•
Marketing measurement
•
Analytics to use
Drawback •
Coverage around 50 %
•
High installation costs
•
Precision between 1-5 m
•
New data not clear concept how to
use it
Bluetooth analytics Positive • More signals •
The penetration is growing
•
Same as Wifi
Drawback •
Coverage around 15 - 35 %
•
Slightly higher precision between 13m
•
Range smaller
Door Counting Positive •
•
Different technologies
•
3D camera – 99%
•
Thermal camera - 90 %
•
Infrared camera – 85 %
Precision up to 99.3%
Drawback •
High installation costs
•
High cost investment
•
Missing returning visitors – entity identification
Sensors Positive •
Small size
•
Long lasting battery
•
High accuracy
•
Highly customizable
•
Connectivity flexibility
•
Camera – can provide features
Drawback •
Higher custom tailored development
•
Higher Maintenance
•
Specific domain
Scrapping data Product list •
Product details
•
Product prices and demand
•
Product review
•
Product recommendation
•
New products
•
Categories, groups and sub-groups
What you can get? External
WiFi
Analytics
Bluetooth Analytics
Door
Counters
Sensors
Internal data
Cash register
Product list/catalogue
Warehouse/Inventory
Retail Pyramid Pricing & Promotions
Optimization and Recommendation
Scoring models
Per category Per store Products, Categories, Vendors and Stores
Customer segments
Based on purchases, time and patterns
Assortments
Basket Analysis
Product, multi-category
Brands
Revenue bases ABC, XYZ analysis Top Brands, revenue, volume, profit
Products
Top products, revenue, Volume, profit
Categories
Analysis ABC analysis
VED Analysis:
Pareto analysis into three categories (A, B, and C)
relative importance of certain items to other items
•
'A' items are very important,
•
Vital – inventory that consistently needs to be kept in stock.
•
'B' items are important,
•
Essential – keeping a minimum stock of this inventory is enough
•
'C' items are marginally important.
•
Desirable – operations can run with or without this, optional
XYZ analysis
HML Analysis
classify SKU according to variability of their demand.
how much a product costs/its unit price
•
X – Very little variation:
•
High Cost (H) = Item with a high unit value.
•
Y – Some variation, seasonality
•
Medium Cost (M) = Item with a medium unit value.
•
Z – The most variation with fluctuations
•
Low Cost (L) = Item with a low unit value
FSN Analysis:
SDE Analysis
classifies inventory based on quantity, rate of consumption and
•
Scarce (S) = Items which are imported and require longer lead time.
frequency of issues and uses
•
Difficult (D) = Items which require more than a fortnight to be available,
•
Fast Moving (F)
•
Slow Moving (S)
•
Non-Moving (N)
but less than 6 months’ lead time.
•
Easily available (E) = Items which are easily available
Product and Brands Top performers by:
Drill-down by:
•
By revenue
•
Store
•
By volume
•
Channel
•
By profit
Time-based data analysis
•
ABC analysis
Analysis
Categories • Product analysis per subgroups • Analysis • 8 Steps of Category management • Category Role
Basket Analysis • Basket per receipt • Customer basket • Multi – Categories basket • Per volume, revenue and profit
• Most profitable combinations
Customer segments
Customer segments IoT • Locations • Time pattern • Frequency of visits
• Places of visits • Demographics Receipts or loyal cards
• Products • Categories • Multi – products or categories • Promotions • Demographics
Scoring models IoT
• Customer segments • Churn • Retention Receipts or loyal cards • Products • Categories • Customer segments • Vendors
• Stores
Assortments Receipts
External sources
• Per store
• New products simulations
• Per category
• Old products • Rate to convert
Pricing and Promotion
• Pricing recommendation • Pricing simulator • Promotion simulation • Promotion recommendation
• Product recommendation
ShopUp Analytical platform which combines different IoT technologies relying on Machine Learning in favor of supporting brick and mortar owners to take data driven decision about their venues’ and
staff performance, shoppers’ interests and patterns, seasonalities, cannibalizations, promotions’ effectiveness, shifts optimizations, queue management and simulation on future demand
Bluetooth Analytics
WiFi
Analytics
Door
Counters
Mobile
Applications
External
POS, Marketing, weather data
CATWING Applying Retail pyramid using external databases and Cash Register
Catwing
Demo
Experiment, be like a ninja Thank you!
sergiev.sergi@gmail.com 0888 400 290