Web Analytics — It’s Surprisingly Simple “WebSideStory knows analytics
In its latest annual report entitled “State of Online Retailing 6.0,” the key finding by Shop.org was that profitability has
and continues to push the
become the norm in online retailing, with more than 70% of
boundaries of innovation.”
retailers reporting positive operating margins.
— David Silversmith, Chief Technical Officer, CARFAX
Improved marketing effectiveness was key to this success. Specifically, participants reported that customer acquisition costs decreased by 60% from 2000 to 2002, while order conversion rates improved by 45% in the same period. Marketing optimization plays a central role in helping businesses improve their online operations and reaching their business goals. Fortunately, Web analytics makes online marketing optimization a surprisingly simple process. This document serves as a guide for using simple and proven Web analytics techniques to help site managers improve their online marketing initiatives.
On-Demand Web Analytics™
|
www.WebSideStory.com
What is Web Analytics? Web analytics refers to the collection, analysis and reporting of Web site usage by visitors. The information helps site managers better understand the effectiveness of their online initiatives and helps them optimize their Web site. This optimization process could occur in a number of areas, including site content, media and promotional mix, merchandising, site process designs, maximizing the value of site initiatives — such as the internal search tool — and much more. While traditional Web analytics packages have concentrated their analysis on the clickstream behavior of users — page by page or sequential navigation of Web site visitors — the return on such analysis has proven to be questionable given the high number of clickstream permutations that could develop. This shortcoming has led the way to a new generation of Web analytics solutions that have moved away from raw data reporting and have instead focused on business reporting. Examples of such reports include marketing campaign analysis, ROI measurement, process analysis and more. This publication serves as a guide for Web analytics best practices, with emphasis on proven ROI. Web analytics can be an integral part of marketers’ goal to optimize their customer acquisition, conversion and retention strategies. By minimizing the guesswork, marketers can better plan their activities and achieve more accurate forecasts than before. Where Can Web Analytics Help? Web Analytics can be effectively applied to optimize the performance of four site types. Most Web sites will fall within one or more of these site types. The following is a brief explanation of each. ■ Commerce: In this environment, the goal is to get customers to buy directly online. Examples of such sites include Amazon, Best Buy and proflowers.com. Web analytics can be centered on the site goal of online purchase and should tie clickstream data to online purchases for detailed business-level reporting. ■ Lead Generation: In this model, the goal of the site is to get visitors to submit their contact information so that they can be contacted by the company’s sales force. Examples of such sites include New York Life, Sun Microsystems
Web Analytics — It’s Surprisingly Simple
|
Page 2
and Northwestern Mutual, which specialize in sales of specialized products and services. In this case, web analytics should be centered on lead capture or form submits to help marketers understand how they can increase their lead conversion. ■ Content: Content sites revolve around the advertising business model. Examples of such sites include ESPN.com, abcnews.com and foxsports.com. The goal is to get visitors to keep coming back. Contents should therefore be refreshed at appropriate rates for such sites to be successful. ■ Support/Self Service: The support or the self service model revolves around providing customers with the ability to find the answers they need regarding their products. This model revolves around cost savings associated with deflection of call center volumes. For a Web site to benefit from Web analytics, it should include one or more of site types mentioned above. For example, the Web site of a software company may include e-commerce for direct sales of software, lead generation for direct selling by the sales force, as well as a self service site for helping with customer support issues. Finally, it is important to note that each one of the site types mentioned above enjoy a different level of Web analytics adoption. For example, content or media sites enjoy the highest adoption level because many vendors have been able to provide specific metrics to them early on. Figure 1-1 is a representation of the adoption curve portrayed by Geoffrey Moore’s “Crossing the Chasm.” Within the figure, you’ll notice that the content sites enjoy the highest adoption of Web analytics, followed by commerce, lead generation and self service models. This publication serves as a guide for best practices in Web analytics based on site models.
Figure 1-1: Web analytics adoption curve by site type
Web Analytics — It’s Surprisingly Simple
|
Page 3
Section 1: The E-commerce Model The new generation of Web analytics solutions has the ability to provide detailed reports centered on specific business goals and objectives as defined by the user of the tool. This reporting flexibility provides users with the ability to test and measure every single Web site initiative. The downside however, is that they can test and measure every single Web site initiative and end up with too much information to distill. Fortunately there are specific areas of an e-commerce site that provide the highest impact on either top or bottom line of the business. This section will therefore concentrate on those key areas that e-commerce managers should focus on when using Web analytics. More specifically, this section discusses the use of Web analytics in optimizing the customer lifecycle. The lifecycle spans from the moment a company reaches its target market, all the way through to getting them to become loyal customers. For an e-commerce operation, the areas of concentration in the customer lifecycle include: ■ Targeting and Reaching the Right Audience ■ Creating the Right Messages for Marketing Campaigns ■ Maximizing Home Page Effectiveness ■ Designing and Optimizing Visitor Conversion Processes ■ Optimizing the Checkout Process ■ Improving Product Placement ■ Increasing Customer Conversion and Retention Using Segmentation This section will then cover some specific Web analytics best practices around each one of the seven areas mentioned above. It is important to point out that the practices mentioned in this publication are using the HBX solution from WebSideStory and may not be applicable to all Web analytics solutions.
Web Analytics — It’s Surprisingly Simple
|
Page 4
Targeting and Reaching the Right Audience The first thing your prospects have to do in order to become your customer is to find you. There are a multitude of channels available to e-commerce managers today in order to help them reach customers. These include search engines (organic or paid placements), banner advertising, promotions on shopping engines such as Shopping.com and mySimon, offline promotions, catalogs, affiliate and partner programs, and more. Marketers often rely on a promotional mix that includes some or all of the channels mentioned above. While in the offline world, determining the optimum marketing mix is more art than science, the situation is quite different in the online world where each promotional element can be tied to the final conversion event. The marketer’s goal in the online world is then to analyze each source based on their performance, and allocate budget accordingly. If a certain promotional channel is underperforming, the budget can be moved away to one with better conversion results. To be specific, each source of acquisition should be observed in terms of acquisition (number of visits or click-throughs generated to the site), conversions (purchases and associated revenue) and retention (lifetime value of those customers). To illustrate the effectiveness of this technique, refer to the example below. Example: Pay-Per-Click Keyword Campaign A manager of a large retail site is in charge of keyword placement at paid search engines. To analyze the value of each keyword, the manager will need to assess the performance of each in terms of acquisition, conversion and retention. To correctly assess his/her goals, it is critical for the Web analytics tool to measure and report lifetime valuation of customers tied to acquisition sources. The final analysis of keywords looks like table 2-1, where each keyword is assessed against the four measures which include visits, orders, revenue, and lifetime value of the customer. From these four measures, two more can be derived, which include browse-to-buy ratio and visit value.
Web Analytics — It’s Surprisingly Simple
|
Page 5
Visit Value Rank
Keyword
Lifetime Value
Orders
Revenue
Browse To Buy Ratio
2,115
31
$3,829.48
1.47%
$1.81
-0.08%
$125.29
883
28
$2,686.31
3.17%
$3.04
67.88%
$206.20
23.47%
1,218
20
$1,776.65
1.54%
$1.46
-19.51%
$189.43
13.43%
Visits
Visit Value
% Above Average
Revenue Per Customer
% Above Average -24.98%
1
luggage
2
samsonite
3
american traveler
4
wallets
824
19
$1,884.05
2.31%
$2.29
26.17%
$177.21
6.11%
5
leather wallet
1,487
19
$1,960.91
1.28%
$1.32
-27.23%
$176.30
5.57%
6
leather bags
509
17
$2,066.15
3.34%
$4.06
124.00%
$176.67
5.79%
7
hand bags
540
16
$1,337.77
2.96%
$2.48
36.71%
$181.10
8.44%
8
travel luggage
354
14
$1,075.42
3.95%
$3.04
67.64%
$180.94
8.35%
9
traveling bags
282
14
$1,555.38
4.96%
$5.52
204.36%
$172.34
3.20%
durable luggage
360
13
$1,769.91
3.61%
$4.92
171.30%
$163.23
-2.26%
10
Table 2-1: Acquisition, conversion, retention assessment of keywords
Definition of Table 2-1: Table 2-1 was produced using HBX Report Builder, a Microsoft® Excel® plug-in that lets users create ad-hoc reports from within the Excel environment. The application lets users create any custom report using their HBX Web analytics data and automate or schedule the delivery of their custom reports to various users within the organization. In this example, top keywords and their corresponding visits, orders, revenue and lifetime revenue per customer have been plotted. Browse-to-buy ratio is then derived by applying the ratio of orders to visits. Browse-to-buy ratio is sometimes also referred to as conversion rate. Visit value is derived using the ratio of revenue to visits. This is an important ratio that effectively shows the value of each visit from a referring source in terms of actual revenue generated. Finally, both visit value and lifetime revenue per customer are compared to overall site average and the differences are shown in either green or red colors using the Report Builder conditional formatting functionality. The analysis in table 2-1 becomes an integral part in the media buyer’s decision making process. If the company’s marketing goal is to increase branding, then high acquisition (visits) keywords are critical to their success. If customer conversion is the key factor, then browse-to-buy ratio (ratio of orders to visits) and visit value are the important indicators. Finally, if increasing customer retention and lifetime value are the end goals — as they often are — then customer lifetime value should carry the heaviest weight in the decision-making process. One often overlooked observation among managers is the comparison between the above data and the overall site average. For example, if the goal of a keyword campaign is to increase conversions, then keywords with visit values above the site average should be given first consideration in the purchase process. In the
Web Analytics — It’s Surprisingly Simple
|
Page 6
example above, the keyword “traveling bags” has a visit value that performs 204% above site average, but a lifetime customer value that’s only 3.2% above normal. In other words, it will result in high conversions and average retention rates. The same exercise can and should be conducted for all other acquisition mechanisms. This will result in the attraction of the most targeted audience and generation of the highest possible ROI from marketing campaign expenditures.
Creating the Right Message for Marketing Campaigns Once your target audience has been identified using the process mentioned in “Targeting and Reaching the Right Audience,” the next step is to use a messaging or promotional mix specific to that audience or segment. In order to better demonstrate this point, consider the example of a visitor who is searching for ski boots. After finding a Web site using a search engine, the visitor is taken directly to the site’s home page where prominent displays of summerwear and swimsuits highlight the site’s summer collection. On the other hand, if the same visitor was taken to a specific landing page highlighting ski boots and related accessories, there would be a higher likelihood of visitor conversion. Personalizing the user experience helps visitors more easily find the right product and often results in higher conversion rates. There are a number of easy steps that you can take in order to personalize the experience of the visitors you acquire through marketing campaigns, two of which are demonstrated below: ■ Search keywords: Use search keywords, especially those resulting in high click-through rate to customize the user experience. With your Web analytics solution, identify the actual products purchased by visitors who find you with these keywords. Once identified, create gateway pages for high-traffic keywords that specifically include the items of interest. See Figure 3-1 for a screen shot of this example. This technique will provide the biggest impact for keywords with high click-through rates and lower than expected conversion rates as they represent traffic with high potential for improvement.
Web Analytics — It’s Surprisingly Simple
|
Page 7
Figure 3-1: Determining products sold based on originating keywords. This is done by starting from the keywords report and cross-correlating to the actual products sold by keyword.
■ Banner placements: Once you identify a certain property that you want to advertise in, you should test and measure various messages, banner designs and placements. The analysis should include the response or click-through rate, as well as the number of conversions and the conversion rates for each. The resulting data will tell you which promotion or message is most effective for each marketing campaign. Let’s say you are running a banner campaign and you want to see which banner is most effective — the skyscraper banner or the 120 x 90 pixel design, as demonstrated in Figure 3-2 — in this case refer to your Web analytics solution and see which design you should emphasize on in order to increase the ROI of your marketing expenditures.
Web Analytics — It’s Surprisingly Simple
|
Page 8
Figure 3-2: Analysis of banner design. Key metrics to measure include responses or clickthroughs, conversions, conversion rate and resulting revenue from each banner design.
Once the proper marketing channels have been identified, the next step in opti mizing the acquisition process is to customize the message to the audience. Use your Web analytics solution to identify conversions by campaign attributes — such as a message, design, or a placement for banner campaigns, individual links for e-mail campaigns, and offers for sponsorship campaigns — and product performance of each campaign and attribute. Push the high-converting products within the campaign gateway pages and customize the content to provide the desired user experience.
Maximizing Home Page Effectiveness Your home page is undoubtedly the most valuable real estate on your Web site. Traffic to the home page of a typical e-commerce site varies from 4% to 20% of the overall site traffic. It is the first impression and point of interaction that most visitors will have within a site, and often determines their online experience. The home page can therefore make or break the site success at converting visitors into buyers. For these reasons, maximizing the effectiveness of your home page as it relates to conversion has become one of the most cost effective areas of improvement for e-commerce managers.
Web Analytics — It’s Surprisingly Simple
|
Page 9
To maximize the effectiveness of your home page, a few critical questions need to be answered, including: 1. How many visitors are browsing past the home page? 2. Where are they going next? What is the click-through rate of each link? 3. What is the conversion associated with each link, and the associated amount of revenue? The first question above is surprisingly powerful and yet very simple to answer using the “single-access ratio” measurement. This is the ratio of visits that do not go past the landing page and is measured by dividing “single-access pages” with “entry pages.” If the single-access ratio is high, then the home page should be optimized through either design changes or placement modification with the goal of decreasing it to acceptable levels. There is no standard as to what the right singleaccess ratio should be. The benchmark should be based on business objectives and site conversion goals and should be established by business managers. The next step is to optimize the page design and content. This can be done by understanding the actual links that people click on within the home page. This type of information is readily available within most Web analytics solutions and should be utilized, as it reveals areas of interest among site visitors — see Figure 4-1. Ideally, the home page should include easy access links to areas of high interest to visitors. There are a number of ways to determine these areas, which include analysis of traffic on various sections of the site, assessment of internal search terms typed by visitors and product category purchases. Once identified, these areas should be made more prominent on the home page to ease visitors’ navigational experience. In addition, underutilized links should be modified or removed.
Web Analytics — It’s Surprisingly Simple
|
Page 10
Figure 4-1: Screen shot of HBX Active Viewing, a browser plug-in that provides a simple, graphical overlay of link information right on your own Web site.
The final step of the home page optimization step is to measure home page links in terms of conversions. Again, the goal is to utilize existing traffic to the home page and measure each home page link based on conversions and revenue, as demonstrated in Table 4-1. Link
ClickThroughs
Resulting Orders
Resulting Revenue
Conversion Rate
Revenue Per Click
Top banner
1,265
18
$956.65
1.4%
$0.76
Free shipping
1,145
22
$1,543.20
1.9%
$1.35
Specials
1,065
21
$1,342.23
2.0%
$1.26
Mens clothing
951
16
$876.32
1.7%
$0.92
Womens clothing
751
12
$801.21
1.6%
$1.07
Table 4-1: Assessment of home page links based on click-throughs and conversion, provided by HBX Report Builder.
The report presented in Table 4-1 makes it easy for decision makers to maximize the effectiveness of the home page, because each link can be easily assessed against revenue. In Table 4-1 for example, the top banner is the least effective at converting visitors and has lower revenue per click than others, despite having a higher click-through rate. One solution could be to revisit the content of the ensuing page and see why it is not successful at engaging the traffic it’s receiving.
Web Analytics — It’s Surprisingly Simple
|
Page 11
Designing and Optimizing Visitor Conversion Processes Once on the site, visitors typically have to navigate through various site offerings in order to complete their transactions. This requires a series of steps that should be easy to find and complete. In the past, many Web analytics users have tried to measure the effectiveness of their site processes using clickstream or path analysis. The return on such efforts have been questionable because there are a large number of different permutations that visitor clickstream can produce on any given site. Instead, managers should take a process measurement approach in mind with the goal of increasing conversions. The process measurement framework is extremely simple, as illustrated below. 1. Managers start by defining their business goal. 2. Once the goal has been defined, they lay out the steps or the process that visitors have to take in order to achieve that goal. This includes the sections of the site that they have to traverse through in order to reach their goal. 3. The Web analytics solution will measure the effectiveness of the goal and process described above, with the objective of identifying major bottlenecks along the way. The benefit of this approach is that measurement is centered around site goals and objectives and users can immediately identify roadblocks along the way. To better illustrate this framework, consider the following example of internal search process measurement. Example: Internal Search Process As an e-commerce manager, you’re interested in assessing the effectiveness of your site’s internal search engine. More specifically, you’re interested in understanding its usage and conversion rate, its ability to drive people to the shopping area of your site — such as the product pages — and ultimately, the checkout section. In addition, you want to identify the specific products of interest to users of the tool. To do so, you start by defining your goal, followed by the process involved, as defined in Figure 5-1. Once the process has been defined, the Web analytics solution can pinpoint the areas of concern. The solution identifies what the process conversion is and at what stage of the process — or funnel of conversion — visitors are abandoning.
Figure 5-1: Defining a conversion process based on a business goal
Web Analytics — It’s Surprisingly Simple
|
Page 12
Even more specific, the solution will pinpoint the actual pages that result in dropoffs, as demonstrated in figure 5-2. The information then allows you to either remove bottlenecks or address the resulting issues in order to streamline the process. Rather than randomly browsing through site clickstream data with the hope of revealing patterns of behavior, consider a more systematic approach in identifying your navigational roadblocks.
Figure 5-2: Presentation of clickstream patterns from conducting a search on the site to conversion, presented as a funnel of conversion. The presentation provides a simple way to find highest drop-off levels. Zooming into the level provides a break-down of highest abandonment pages.
Figure 5-2 is a graphical representation of the internal search process discussed earlier. The first screen portrays the entire conversion process that visitors have to undertake. The process starts when visitors conduct a search, they then decide to get further engaged with the site and browse the site’s product pages. Notice that although e-commerce sites have hundreds or thousands of product pages, they should be treated as only one step of the process, or one level of the conversion funnel. The visitors continue their conversion process by adding product(s) to the cart, initiating checkout, and finishing the checkout process, which includes shipping, billing and the final confirmation page. The funnel view quickly demonstrates that most visitors abandon the process while in step 2 of the process. You can effectively drill down into that level and identify the individual pages that are causing highest drop-offs, as demonstrated in the superimposed screen. This measurement approach provides a quick and easy way to optimize conversion processes without the need to analyze individual clickstream patterns.
Web Analytics — It’s Surprisingly Simple
|
Page 13
Optimizing the Checkout Process In order to complete an online transaction, today’s e-commerce sites include a checkout process which consists of a series of sequential forms that visitors have to fill out. It is through the checkout process that the site collects the necessary information in order to fulfill the transaction. The optimization of this process has proven to be one of the most effective ways for managers to increase site conversions and their bottom line. By concentrating their efforts on a handful of pages, businesses can achieve great returns on their site optimization initiatives. The checkout process is essentially another site conversion process that should be constantly monitored and optimized using the framework described in “Designing and Optimizing Visitor Conversion Processes.” It should be measured at each step of the process so bottlenecks can be identified. Figure 6-1 represents a conversion funnel representing the checkout process. It helps managers pinpoint the most and least effective sections of the process.
Figure 6-1: Site checkout process presented as a conversion funnel.
One of the many advantages of the outsourced Web analytics solutions today — besides data accuracy — is that they can actually reveal the precise form fields that result in visitor abandonment. The information tells you how checkout forms should be designed including, which fields should be omitted or which should be made optional based on abandonment or conversion rates. Many companies have the tendency to ask for too much personal information from their prospects during the crucial checkout process. Web analytics solutions help pinpoint visitor abandonment of the checkout process and provide feedback to help you pare down, refine or re-focus checkout forms to optimize conversions. Figure 6-2 represents such analysis, in which you can observe the actual form fields resulting in visitor drop-offs. Finally, it is imperative that you measure the effect of any changes to the site us-
Figure 6-2: Analysis of form fields resulting in abandonment. The information may be trended to assess the results of changes on checkout forms.
ing Web analytics. More specifically, you should observe the effect of your changes on your conversion rates. For example, does the conversion rate between steps 2 and 3 of the checkout process increase or decrease after you make changes in the second step? You can easily observe your changes by either trending the conversion rate between the two steps or comparatively see the conversion rates for before and after the change.
Web Analytics — It’s Surprisingly Simple
|
Page 14
Improving Product Placement Effective product placement is a key ingredient in every successful e-commerce operation. This includes optimal use of merchandising on the home page or other key pages with high traffic, display of appropriate products in marketing campaigns and promotions, and efficient use of real-estate for maximizing cross-sell and up-sell opportunities. Optimization of home pages was discussed earlier in “Maximizing Home Page Effectiveness” and the same set of practices can be applied for other high traffic pages, such as the discounts page and top category and brand pages. Use of product placement for marketing campaign optimization was also discussed in “Creating the Right Message for Marketing Campaigns.” Another key aspect of merchandising optimization available to merchandisers is the use of crosssell opportunities. If a visitor is in the process of purchasing a camera, it is much easier to get them to add a camera case to their shopping cart than to have them purchase the case from scratch. As you can imagine, product pages are some of the most effective merchandising real estate on the Web site. The identification of strong cross-sell opportunities and effective placement of such products on individual product pages based on that information has proven to yield high ROI. In addition to the points mentioned above, make sure you’re using high converting products to your advantage. Some products have a higher propensity to be added to cart but have low conversions, while others have the opposite effect. The product purchasing information can help derive site merchandising decisions. Low converting products are sometimes indicative of hidden problems such as high shipping costs. High converting products should instead be utilized to entice purchase. Specifically, products should be monitored at all purchase levels, including product views, number of times added to cart, checked out and purchased. Low conversions from checkout to purchase could be indicative of high shipping fees. Low conversion from view or cart to purchase may indicate that the items require additional marketing efforts, such as discounts or free shipping. An example of a merchandising conversion matrix is illustrated in Table 7-1.
Web Analytics — It’s Surprisingly Simple
|
Page 15
Metric
What Is It
Why Does It Matter?
Browse To Buy
Ratio of items bought to product views
High converting products should be utilized to entice purchase. They should be displayed more prominently on high traffic areas of the site such as the home page.
View To Cart
Cart Conversion
Checkout Conversion
Ratio of cart adds to product views
Ratio of orders to carts
Ratio of order to checkouts started
Low converting products can benefit from additional promotional efforts (price discounts, volume purchases, etc.) High ratio products are those that are engaging your visitors. This is not the final conversion metric you need, but you should utilize these products in engaging your audience. Low ratio products are unsuccessful at engaging visitors. Analysis of cross-sell data may reveal clues, such as the ability to promote the product along with complementary items. High ratio: once added to cart, purchase likelihood is high. Promote these products heavily as they can increase your top line. Low ratio: items that require additional selling, such as free shipping or discounts. Explore promotional sales in conjunction with their highest cross-sold items. High ratio: easy converting products, demonstrating value to the customer. These are products that you can use to attract visitors to your site. Low converters may require additional selling, such as free shipping, discounts or coupons.
Table 7-1: Using conversion rates to determine product placement strategies
An often overlooked area among many e-commerce sites is that of category and brand conversions. Some product categories and/or brands have a higher propensity to convert your browsers into buyers. In this case, you should use keyword and media buys around those product categories and brands to attract qualified visitors; prominently display them on your high traffic pages and utilize them to remarket to your existing client base. This information can be easily made available from Web analytics solutions and should be used as drivers of product and merchandising placements among e-commerce sites.
Web Analytics — It’s Surprisingly Simple
|
Page 16
Increasing Customer Conversion and Retention Using Segmentation Not all visitors are created equal. This is a common knowledge among marketers and is the basis for segmentation. Segmentation may be accomplished using a number of criteria such as demographics or behavioral aspects. However, a number of best practices have been developed in the Web analytics field that you can explore. ■ E-mail subscribers: This is one of your most captive audiences. These are visitors who opted-in to receive special offers from you. Visitor segmentation can be effectively used to separate your e-mail subscribers’ behavior from others. You can find out specifically what products and product categories spark their interests and send special offers that provide the highest likelihood to purchase. Figure 8-1 provides an example of such reporting. ■ Partner traffic: The traffic that you receive from your partners or key affiliates can be segmented in order to help you understand their behavior and interests in great detail. For example, if you’re in a partnership with DealTime, mySimon or Epinions.com, you may wish to separate their respective traffic
Figure 8-1: Analysis of product pages browsed by e-mail subscribers. Treating e-mail subscribers as a visitor segment allows for reporting specific to that group.
so that you can see specifically what type of merchandising you can promote to each. This more targeted marketing will result in higher conversion rates per partner. You’ve identified the right people, used the right messaging and promotional mix to acquire them, optimized your home page, checkout process and product placements with the aim of converting them. The next step is to retain them for subsequent purchases. Repeat customers are often the most profitable segment for marketers. They have lower order acquisition costs and spend more money per transaction than first-time buyers. Successful marketers capitalize on this fact to increase their bottom line. Today’s Web analytics solutions help you improve retention activities by identifying the actual customers in a specific segment. This powerful functionality lets you conduct targeted marketing campaigns to a captive audience. Three simple examples are up-sell promotions, geographic targeting and repeat purchase patterns. ■ Up-Sell Promotions: By combining cross-sell reports and individual customer identification, you can create highly targeted campaigns. Target the buyers
Web Analytics — It’s Surprisingly Simple
|
Page 17
of a certain product, identify the cross-sell opportunities for the product that they bought, and send them discounted offers to entice repeat purchases. ■ Geographic Segmentation: Segmentation can be used to target customers based on their geographic profile. Web analytics solutions can be utilized to identify differences of behavior and purchase by geographic locations. The information can then be used to create targeted marketing campaigns in the form of e-mail or direct mail to the customer base or registered users. ■ Repeat Purchase Patterns: Segmentation can help identify behavioral differences between new and repeat buyers. For example, if a product has a higher propensity to be bought by existing customers than first-time buyers (table 8-1), then it should be utilized in customer promotional pieces.
Product Name Running Shoe (Mens) Running Shoe (Womens) Tennis Shoe (Mens) Basketball Shoe (Womens) Basketball Shoe (Mens) Tennis Shoe (Womens)
All Buyers Units 215 106 141 55 59 62
Repeat Buyers Units 20 5 28 7 1 18
% 9.3% 4.7% 19.9% 12.7% 1.7% 29.0%
Table 8-1: Identifying products with high purchase propensity from repeat buyers
Segmentation techniques are now well developed and understood among marketers. The same techniques apply to visitor segmentation. Today’s Web analytics solutions make it easier to effectively target specific segments of site visitors or buyers. Although the technical implementation may vary from one vendor to the next, the capabilities provided by the more mature solutions should allow you to develop meaningful segments that will help you in your conversion and retention efforts.
Web Analytics — It’s Surprisingly Simple
|
Page 18
Section 2: The Lead Generation Model The lead generation model in many ways resembles the e-commerce model in a sense that the goal of the site is to direct visitors towards a final conversion event. In this case the conversion event is that of leads, which are often passed to the organization’s direct sales team. Such organizations often use a number of standard KPIs (key performance indicators), including browse-to-buy ratios and CPL (cost per lead). Many of the same points discussed in Section 1 can be applied to the lead generation model, however this segment trails the e-commerce segment due to cross-channel challenges associated with analyzing leads captured online and converting offline through the direct sales force. There are a number of key steps that marketers can take in order to optimize the lead generation process, including: ■ Maximizing Campaign Effectiveness ■ Optimizing Site Navigation ■ Optimizing Online Forms ■ Maximizing Content Effectiveness ■ Increasing Lead Generation Using Segmentation ■ Closing the Loop Between Online Lead Generation and Offline Conversion The ensuing chapters in this section will cover a number of specific Web analytics best practices around each one of the areas mentioned above.
Web Analytics — It’s Surprisingly Simple
|
Page 19
Maximizing Campaign Effectiveness In order to increase their lead generation, marketers often rely on a marketing mix that may include banner ads, keyword placements, affiliate marketing and offline initiatives such as direct mail, radio and TV advertising. In the lead generation model, the goal is to get prospects to fill out an online form. Marketers often rely on a number of fulfillment pieces to entice lead submittals. Examples of fulfillment pieces include download of white papers or trial software, free consultation or discounted offers. The goal of the Web analytics application in this case is to help marketers measure the effectiveness of each one of the marketing campaigns, and correlate their success along with each fulfillment piece or offer being used. In addition, the marketing manager should pay close attention to the campaign return on investment, as demonstrated in figure 9-1. In this example, the marketer can analyze the campaign payback time (the point in time at which the campaign starts generating positive returns). This can be an important analysis since companies can accelerate their returns by investing in campaigns with quick payback.
Figure 9-1: Campaign ROI. This view provides the ability to see the campaign break-even point. Depending on the campaign, the breakeven point may occur sooner or later in the campaign lifecycle.
This type of information can then be used to determine the best marketing mix, the best use of offers and positioning and the right amount to invest for each marketing campaign.
Web Analytics — It’s Surprisingly Simple
|
Page 20
Optimizing Site Navigation The site navigation is one of the critical components that can help managers reach their lead generation goals. The most effective way to assess site navigation is to use funnel analysis as discussed in “Designing and Optimizing Visitor Conversion Processes.” An example of such funnel may consist of visitors entering the site through the home page, continuing their navigation by visiting one of the product pages and exploring descriptive pages such as features and benefits, continuing on to the online form and filling it out so that they can be contacted by the company. This is demonstrated in figure 10-1, which demonstrates such behavior. Again, the benefit of such analysis is that one can immediately discover the bottlenecks and address them accordingly. As discussed in “Designing and Optimizing
Figure 10-1: A funnel representation of a lead conversion process. The information can be used to determine where within the process visitors are abandoning the conversion process.
Visitor Conversion Processes,” the user can get detailed analysis on each level of the conversion funnel, which may consist of multiple pages. Another way to assess site navigation against lead generation is to simply correlate page behavior with lead generation. For example, say you have a hypothesis that your product benefits page is positively impacting lead conversion. To test this hypothesis, you can trend and compare the two pages in question, in this case, the product benefits page and the lead confirmation page — see figure 10-2. By trending pages or other elements and comparing them, you can immediately see whether there’s a correlation between two or more elements. The information can then be used to make changes to pages that are negatively impacting lead conversion.
Optimizing Online Forms In order to contact a company, most sites require visitors to fill out an online form. Lead generation sites should utilize the form abandonment functionality discussed in “Optimizing the Checkout Process” to optimize their online forms. As dis-
Figure 10-2: Trending of multiple pages across any time frame. The graphical representation can be used to determine if there’s any correlation between the trended elements.
cussed earlier, form abandonment reveals the actual form fields that result in visitor abandonment. The information tells you how online forms should be designed, including which fields should be omitted and which ones should be made optional. It is a common mistake among marketers to use online forms to gather more information about their prospects. The thinking is that by adding a few fields on forms that people are going to fill out anyway, they can get valuable data that can be used for marketing purposes. Unfortunately, data often reveals that this is not the case. Gathering of survey data such as demographics should be kept separate from the online conversion forms, as they may and often do turn visitors away.
Web Analytics — It’s Surprisingly Simple
|
Page 21
Figure 11-1: Trending of form elements — provides the ability to study the effect of changes on site forms. In this case, a change was made to the site form (represented by the red “notes” icon) resulting in lower abandonment rate.
The information in the form abandonment reports can then be used to make changes to online forms and the effect of the changes should be carefully observed. This can be done using the trending functionality shown in figure 11-1. Trending lets users observe the effect of their changes on form conversions. Because users can concentrate their optimization efforts on a number of Web pages, form optimization has shown some of the most dramatics improvements on Web initiatives.
Maximizing Content Effectiveness Web sites spend a great deal of time and resources on content to help attract visitors, and capture leads. Content analysis is one of the core measurements of Web analytics solutions today. More specifically, the Web analytics solution should have the ability to measure content effectiveness based on content hierarchy defined by the user of the tool. To demonstrate this hierarchical approach, assume you’re a provider of software applications. Your Web site hierarchy may first resemble your site structure, which includes products and services, corporate infor-
Web Analytics — It’s Surprisingly Simple
|
Page 22
mation, and contact information. Once on the “products” section, users will have the ability to browse by product, and once in a specific product area, they can browse through product information, features, benefits, product demonstration, specifications and more. This is further demonstrated in figure 12-1, where the user of the Web analytics solution can start from the top level, and work his/her way down the content hierarchy.
Figure 12-1: View of site content hierarchy, represented as folders. Within each folder, users can drill down into sub-folders for more detailed analysis.
The content hierarchical approach lets the user analyze site content in the context of how visitors see site. This user-centric approach allows site managers to accurately see visitor behavior and immediately see what site sections and sub-sections enjoy the most or least popularity among visitors. The information will then help managers know exactly what investments should be made in each area of the site. By putting more resources on popular content areas and less on unpopular ones, managers can align their content investments with their visitors’ interests.
Increasing Lead Generation Using Segmentation Database marketers are all too familiar with segmentation. Segmentation techniques have been used for decades to identify clusters of customers or prospects that can then be more directly targeted. Traditional segmentation techniques include demographic, psychographic and RFM (recency, frequency and monetary) segmentation, where those in the same clusters are marketed together, resulting in higher conversion rates. On the Web, segmentation can occur in a number of
Web Analytics — It’s Surprisingly Simple
|
Page 23
ways, including behavioral segmentation. Behavioral segmentation lets users of the Web analytics solution segment site visitors based on a common behavior. Examples of behavioral segmentation include registered/unregistered users, leads, prospects, or more web specific behaviors such as visitors who downloaded a white paper or viewed a product demonstration. Once segmented, marketers can then look at activity specific to that segment, or even the tendency for a segment to convert to a more desirable segment such as a lead. Consider a design firm specializing in both print and Web development. By segmenting their visitors based on their two core competencies, they can then see which segment (print or Web) is more likely to contact them and become an online lead — see figure 13-1. The information helps managers understand which segment has the greatest potential in terms of online lead generation. The success of that segment can then be replicated to increase online leads
Figure 13-1: Conversion of segments. In this example, the user can immediately see the likelihood of one segment to convert into the desirable “lead” segment.
from the other segment. Another way to use segmentation is to get traffic breakdown by visitor segment — see figure 13-2. The information will help site managers better understand whether their content is being properly used by the right segment. For example, consider a firm that provides services to both the automotive and aerospace industries. If the firm is spending a great deal of resources developing content for their automotive services, but that content is mostly being viewed by their aerospace audience, they may want to reconsider their content investment strategy and spend a higher proportion of their resources on the aerospace segment. Figure 13-2: HBX Active Viewing allows for an easy break-down of page traffic by segment.
The information can be used to better utilize content for lead generation activities.
Another approach that the firm mentioned earlier can take is to assess the conversions from the aerospace or automotive segments into the lead segment. The information can then be used to conduct more targeted marketing efforts to help increase site conversion rate.
Web Analytics — It’s Surprisingly Simple
|
Page 24
Closing the Loop between Online Lead Generation and Offline Conversion The online lead generation is only a step in the customer acquisition process. Companies selling complex solutions requiring customer education often rely on their direct sales force to manage the customer interactions. In order to help their sales force increase their productivity, the company’s marketing team is in charge of generating leads. Web sites are the preferred channel today for lead generation because of their low cost nature. Leads generated online are then sent to the sales force using some sort of sales force automation (SFA) tool. Unfortunately, once the lead is passed to the SFA tool, the marketer often lacks the necessary metrics and infrastructure to tie those generated leads to final sales conversion. Often times, to overcome this shortcoming, marketers rely on law of averages. Here’s an example: The average cost per lead (CPL) is $20. One out of every 20 leads is converted, and the average selling price is $1,000. Therefore, for every $400 spent, the company generates $1,000 in revenue. Although this may be acceptable on the surface, it does not address the fact that the company is likely spending too much on some campaigns and too little on others. To address this shortcoming, WebSideStory has teamed up with salesforce. com, the leading outsourced SFA. Through the partnership, marketers can close the loop between online lead generation and customer conversions through the salesforce.com tool. All that is needed from users is subscriptions to both services.
Web Analytics — It’s Surprisingly Simple
|
Page 25
Figure 14-1: Graphical representation of cross-channel campaign reporting. While the lead generation occurs online, the sales conversion could occur offline, and the centralized reporting lets marketers intelligently allocate marketing budget.
This kind of closed loop analysis is shown in figure 14-1, which portrays the campaign and customer lifecycle. For example, consider an email campaign. The campaign funnel shows the email performance, including number of emails sent, email open rate, click-throughs, leads generated, and finally sales conversions that occurred from within the salesforce.com system. Through this closed loop analysis, one can clearly see the actual value of leads generated and put more emphasis on campaign with high sales conversions.
Web Analytics — It’s Surprisingly Simple
|
Page 26
Section 3: The Content/Media Model The Content or Media model revolves around serving content to site visitors. The nature of the content could vary anywhere from sports, news, lifestyles or other subjects. The revenue in this model is derived from one of two ways, through advertising or through visitor subscription fees. The goal of the manager in this case is to both increase subscription rates and to increase sales of advertising inventory. Web analytics solutions are used therefore to help in the following areas: ■ Managing Advertising Inventory ■ Optimizing Site Content ■ Matching Inventory to Advertiser Needs ■ Increasing Subscription Rates A number of specific Web analytics practices can help managers increase their effectiveness in each one of the following areas. The ensuing chapters will discuss these in more detail.
Web Analytics — It’s Surprisingly Simple
|
Page 27
Managing Advertising Inventory Inventory refers to the amount of space or banners that a site can sell in any one period. Measured in impressions, there’s usually a one-to-many correlation between page views and inventory. For example, on any specific Web page, there may be a number of ad spaces with different formats and sizes. Each format will then command a different CPM rate (cost per thousand impressions). In order to increase inventory, site managers rely on a number of mechanisms, including increased site traffic and increased visit duration. A number of metrics can be used as benchmarks for this purpose. These include page views (a preferable method since inventory is often directly correlated with page views), visits, page views per visit and time spent on site. Each may serve different purposes, but a number of KPIs can be derived from such reports. Figure 15-1 is an example of how site managers may use Web analytics to measure their inventory. A KPI that can be used in this case is “percent of inventory sold.” The percentage can be based on both impressions and monetary value. Inventory 125x125 Page Name
120x90
468x60
Impressions
Banners
CPM
Banners
CPM
Banners
CPM
/Sports/Football/NFL
125,003
3
$25
4
$35
2
$45
/Sports/Football/NFL/Teams
119,342
3
$25
4
$35
2
$45
/Sports/Football/NFL/Teams/San Francisco
115,321
3
$25
4
$35
2
$45
/Sports/Football/NFL/Teams/Green Bay
109,321
3
$25
3
$35
2
$45
/Sports/Football/NFL/Teams/Dallas
102,123
4
$25
3
$35
2
$45
/Sports/Football/NFL/Teams/NYG
97,321
4
$25
3
$35
2
$45
/Sports/Football/NFL/Teams/Miami
94,564
4
$25
3
$35
2
$45
/Sports/Football/NFL/Teams/Chicago
87,629
4
$25
2
$35
2
$45
/Sports/Football/NFL/Teams/San Diego
82,103
4
$25
2
$35
2
$45
/Sports/Football/NFL/Teams/Oakland
67,354
4
$25
2
$35
2
$45
Total Inventory (impressions) Total Inventory (vallue)
3,531,337
3,122,823
2,000,162
$88,283
$109,299
$90,007.29
Table 15-1: Example of Inventory Management using HBX Report Builder
Table 15-1 is an example of type of inventory reports that users can run using HBX Report Builder. In this case, a manager of the football section of the media site has a listing of football pages, along with the number of impressions or page views. Following the traffic information is the inventory on each page, classified by banner size, along with the CPM rate. The total available inventory is then calculated in both impressions and monetary value. The manager can use these figures
Web Analytics — It’s Surprisingly Simple
|
Page 28
to measure the percentage of sold inventory in any given period. More specifically, this methodology provides the ability to see what inventory should be promoted more aggressively.
Optimizing Site Content One of the biggest challenges that media sites face is the optimization of site content. More specifically, content managers are interested in finding out what type of content they should invest in and what refresh rate the site should adhere to. As expected, Web analytics plays a major role in helping answer these questions. First, the Web analytics solution should be implemented so that the content-level reporting matches the site content structure. This was discussed earlier in “Maximizing Content Effectiveness.” Say that as a media site, you serve a number of different editorials such as sports, news, weather and classifieds. Under the sports section, visitors can read articles about football, basketball, tennis and soccer. Under each sport, they have the option to get more detailed articles such as teams, scores, player statistics and so on. The Web analytics solution’s ability to match its reporting to user-centric content hierarchy lets the managers get a true understanding of what editorials are popular and what are not. They can then use this information to put more editorial budget on popular content and sub-content areas.
Figure 16-1: The frequency index measures elapsed time between consecutive visits. The Recency index measures elapsed time from visitors’ last visit.
Web Analytics — It’s Surprisingly Simple
|
Page 29
Another area in which content managers can benefit from Web analytics is in determining the content refresh rate. The content refresh rate is the amount of time that should elapse before the site content is updated. The Web analytics reports that help in this area include the frequency and recency indices, as demonstrated in figure 16-1. The frequency index is a measure of the elapsed time between site visitor’s consecutive visits, while the recency index measures the time elapsed since the visitor’s last visit. While the frequency index determines how often site content should be refreshed, Recency index provides insight on how deep into the cycle the site is at any point in time. A content/media site that refreshes its content more frequently than its optimum refresh rate runs the risk of overspending in its editorial budget. Conversely, a site with a lower refresh rate than its optimum rate will result in visitors’ viewing of old content which will hurt visitor retention. By matching the content refresh rate with the frequency index, content managers can optimize their site content initiatives.
Matching Inventory to Advertiser Needs A major source of income for content/media sites is through sales of available inventory. The inventory is sold the by the media site’s ad sales team to potential advertisers who want to reach the site’s audience. The CPM rates that Web sites command has usually a direct correlation with the click-through rates that the marketing campaigns generate, where Web sites with higher click-through rate can command higher rates. While increasing traffic will help increase inventory, increasing the CPM rate can have the highest impact on the site’s top and bottom line. Content/media sites can command higher CPM rates by directing advertisers to more targeted audiences. They do so by utilizing visitor segmentation techniques. Say you’re a seller of video games and have identified your target market as men between the ages 25 to 35. When buying advertising impressions, you want to make sure that your ads are as targeted as possible. Advertisers are willing to pay a higher CPM rate for their impressions if the audience is highly targeted. The media site in this case can use the reports generated from their Web analytics package to direct the advertiser to site areas with the highest concentration of the buyer’s target audience — see figure 17-1.
Web Analytics — It’s Surprisingly Simple
|
Page 30
Figure 17-1: view of content hierarchy by a segment of visitors (male visitors in this example). The information can be used to direct advertisers to targeted content areas. More targeted advertising command higher CPM rates.
The figure above is a report that media sites can run to direct advertisers to targeted content. In this case, the user can select a specific segment of visitors and see what the most popular site areas are for that segment are. The information can then be used to command a higher premium from the advertising inventory.
Increasing Subscription Rates Another source of income for content/media sites is through subscriptions. Not all content/media sites use the subscription model, but those that do require visitors to pay a reoccurring subscription fee in order to get access to prime content. The subscription revenue is then used to fund content providers such as staff, authors and third-party vendors. In this model, revenue is derived from two components: new subscriptions and recurring revenue from existing subscribers. To optimize new subscriptions, marketers can rely a number of Web analytics reports, including click-stream and funnel analysis, as discussed in “Designing and Optimizing Visitor Conversion Processes.” The information can be used to streamline the conversion processes and understand what offers work best at converting visitors. Because of the recurring revenue from the subscription model, retaining the subscribed users is often as critical as acquiring new ones. For this purpose, one of the biggest values that Web analytics packages offer is to provide detailed
Web Analytics — It’s Surprisingly Simple
|
Page 31
insight on the behavior of the subscriber base. This is achieved using visitor segmentation, where subscribers are segmented separately from the rest of the site visitors and their behavior is reported in detail — see figure 18-1. In this example, a report is generated to assess the most popular pages by the subscriber segment.
Figure 18-1: Analysis of subscriber traffic. By determining the most popular pages among subscribers, content managers can invest in appropriate content to guarantee subscription retention.
The information will provide content managers with insight as to which content the subscribers are finding of interest. In this case, more emphasis should be put on the content areas that are popular among the subscriber base, resulting in higher retention rates.
Web Analytics — It’s Surprisingly Simple
|
Page 32
Section 4: The Support/Self-Service Model Enterprises today face the daunting challenge of simultaneously maximizing customer satisfaction and minimizing costs. The Web-based support/self-service model revolves around providing fast, accurate answers to end-users’ questions. The result is reduced support costs for the organization. A study by Forrester reveals that the average cost per transaction using the traditional phone channel or call center is as much as $33 compared to $1.17 for the Web self-service site. Based on this data, enterprises can achieve tremendous cost savings by deflecting phone traffic to the self-service channel. In this case, the optimization of the self-service Web site should revolve around serving the appropriate content to help deflect support calls. The remainder of this section will discuss topics on: ■ Measuring the Self-service Site Volume ■ Identifying Top Customer Issues ■ Utilizing Web Analytics to Deflect Call Center Volume This section will then cover some specific Web analytics practices around each one of the areas mentioned above.
Web Analytics — It’s Surprisingly Simple
|
Page 33
Measuring the Self-Service Site Volume One of the first things that managers can look into is the high-level metrics associated with their support or self-service Web site. To do so, they have to treat their support site as its own entity so that they can accurately measure key metrics and benchmark them when needed. These metrics are often the same used in other site types, but one thing to consider is that the evaluation criteria is very different in this case. For example, the user still needs to look at some key traffic trends such as visitors, visits, page views, and time spent on site, but often the goal is to lower some of these benchmarks, not increase them. For example, for a content/ media site, the goal is often to increase visitor loyalty and time spent on site. For a support site however, the goal is to help customers as fast as possible, therefore a lower loyalty index and shorter time spent on the site are desirable. Once the benchmarks are set, the goal is to monitor the benchmarks. The HBX Report Builder application can be used to create dashboards that include all the necessary KPIs for this business model. For example, as a manager of a support site, you want to make sure that customers get answers to their questions in as short a time as possible. High-level measures to capture such trends include short visit duration, low loyalty and frequency indices, and a high percentage of satisfied compared to dissatisfied users. The latter can be achieved using the visitor segmentation functionality of the Web analytics solution. By segmenting users based on their satisfaction level, one can trend the satisfaction level with the Web site as shown in figure 19-1. It provides an example of a critical satisfaction index that the site manager can use to quickly measure and benchmark against past site behaviors. A downward trend (development of more dissatisfied users compared to those satisfied) should be used as an alarm that overall customer satisfaction is deteriorating and should cause further investigation into the support site.
Web Analytics — It’s Surprisingly Simple
|
Page 34
Figure 19-1: An example of overall satisfaction measurement using visitor segmentation. The site manager’s task is to increase the overall percentage of satisfied users (blue chart).
Identifying Top Customer Issues One of the most widely used areas of Web analytics in the support/self-service model has been to use clickstream data to identify the top customer issues. This includes determining what products require the most assistance, what are the top issues relating to each product, and understanding where customers are facing the most difficulties. This is an area that can be easily uncovered with Web analytics. For example, by matching the Web analytics content hierarchy to that of the support site, one can easily find what are top categories and sub-categories. This was discussed earlier in “Optimizing Site Content” and “Matching Inventory to Advertiser Needs.” The information provides managers with the insight they need to determine how much budget they should allocate to each product line, what areas they should consider within each product line, and what issues are customers increasingly facing. The information can be used in a number of ways. First, products with high number of customer support issues should be highlighted more effectively. Site optimization initiatives should help make these issues more accessible. Second, the information should be forwarded to the call center so that they are better prepared to address customer calls. Because the Web channel is an effective medium for collection of visitor and customer interactions, it can be leveraged to provide timely feedback to other channels. For example, by observing a spike in traffic to pages relating to a product, one can also expect an
Web Analytics — It’s Surprisingly Simple
|
Page 35
increase in call volumes for that product. Organization can use Web analytics to isolate the issue and be better prepared to address customer calls. A well-known practice by many support sites is to directly ask users for their feedback on the support content — see figure 20-1. This information can be directly captured in the Web analytics package and utilized to assess the effectiveness of individual content. This ranking can then be used in different ways. First, managers can start by determining the average ranking of all support site content. Next, they can assess ranking of individual content and compare it to the overall average ranking. Contents that get below average ranking should then be improved to be more specific to what the user is trying to achieve. This practice should be repeated with the goal of increasing the overall content ranking. As the overall average improves, a new set of content(s) will surface with below average ranking. By optimizing those, you engage in a continuous improvement loop that will provide a
Figure 20-1: An example of question support sites can ask to evaluate site content. The information can be captured by the Web analytics tool to measure the content ranking.
sure way to increase support site effectiveness.
Utilizing Web Analytics to Deflect Call Center Volume The entire support/self-service business model revolves around deflecting calls from the call center. This is done through a better understanding of customer concerns, some of which were discussed in the previous section. For example, we discussed how one could identify the contents with the most negative feedback so that they could be improved. What if such information is not available? What if visitors do not bother providing this information? In this case the information has to be deducted from the click-stream activity. More specifically, the measurement should be around contents that do not provide the users with what they need. To understand this, let’s start with a conversion process as defined in “Designing and Optimizing Visitor Conversion Processes.” A successful visit to the support site can be described as one where the visitor enters the site, conducts a search for a topic of interest, finds the appropriate content and requires no further action. Consequently, an unsuccessful visit to the support site can be defined as one where the visitor enters the site, conducts a search for a topic, browses through a list of options provided, and because the content did not help solve the issue, the visitor contacts the company’s support team for personal assistance — see figure 21-1.
Web Analytics — It’s Surprisingly Simple
|
Page 36
Figure 21-1: Defining an unsuccessful visit process for identification of costly customer issues.
Once the above process has been defined, a conversion funnel can be created to further measure the unsuccessful support visits. The analysis of this funnel will help support managers in their role by answering the following questions: ■ What percentage of support site visits result in a request for more help? ■ What are the top viewed pages or contents that resulted in customers requesting for more help? ■ What products or issues are resulting in users abandoning the support site? In this case, support managers can efficiently determine products for which the support site should be improved, as well as discover the main areas of concern for customers who are not being appropriately served. The end goal of this process of course is to deflect traffic from the call-center, which commands a much higher cost per transaction than the Web site.
Conclusion The Web analytics approach is that of gradual changes and improvements or a continuous improvement loop consisting of three simple steps: Track, Analyze and Optimize. Web analytics optimization can be labeled as continuous evolutionary improvements, and not drastic modifications to the Web site. It is therefore important for users to constantly monitor their site behavior, but more importantly, their site changes and improvements over time, as demonstrated in Figure 22-1. More importantly, site changes should be mapped to their effect on the company’s end goal of customer conversion, retention or deflection of volume from the call center.
Web Analytics — It’s Surprisingly Simple
|
Page 37
Figure 22-1: Assessment of changes on page traffic using comparative views and resulting ramification — pages trending up or down.
Finally, there is a tendency among marketers to compare their performance to others using market research studies or industry surveys. Such data points may not always provide the desired benchmarks. A floral Web site, for example, has a very different audience than an outdoor retailer. Although both may have the same KPIs — such as conversion rate, cart conversion rate or average selling price — the numbers will vary drastically. A better comparison may be that of industry peers, but due to the competitive nature and sensitivity of data, competitors are unlikely to share their KPIs with others. Instead, the goal should be to constantly optimize one’s own performance. Start with an objective — such as increase conversion rate by 5% — and then use Web analytics to see where you can achieve the highest results for your efforts. Apply the changes and then measure the results, until the objective is met. It is through these exploratory actions that you can fully optimize your customer lifecycle using Web analytics.
Web Analytics — It’s Surprisingly Simple
|
Page 38
United States 10182 Telesis Court, 6th Floor San Diego, CA 92121 [P] 858.546.0040 [F] 858.546.0480
EMEA Neptunusstraat 23 2132 JA Hoofddorp, NL [P] +31 (0) 23-5677900 [F] +31 (0) 23-5541011
United Kingdom 212 Piccadilly London, W1J 9HG
WebSideStory is a leading provider of on-demand Web analytics services. WebSideStory’s services collect data from Web browsers, process that data and deliver reports of online behavior to our customers on demand. More than 600 enterprises currently use WebSideStory’s services to understand how Internet users respond to Web site design and content, online marketing campaigns and e-commerce offerings. As a result, WebSideStory’s customers can make more effective marketing decisions and improve the merchandising, sales, support and design of their Web sites.
[P] +44 (0) 207 917 6280 [F] +44 (0) 207 917 6281
For more information, visit www.WebSideStory.com or call 877.2BUY.HBX
© 2004 WebSideStory, Inc. All rights reserved. HBX is a trademark and WebSideStory is a registered trademark of WebSideStory, Inc. Other trademarks belong to their respective owners. PN: WSS-005-03