When You Need More Than Clickstream Data

“Data, data everywhere” could be the cry of the merchant who selects a Web Analytics solution without understanding the commitment it requires from technical, marketing, merchandising and other staff.

The same lament might come from selecting a solution that’s not optimized for e-commerce or for a user’s special needs related to Web site optimization. In the long run, it’s better to have realistic expectations of what you want to use a Web Analytics solution for.

Like many companies, Gardener’s Supply tried several Web Analytics solutions before settling on its current application: Omniture. But the merchant’s experience taught it to be realistic about how it would use a Web Analytics tool. Chris Thompson, chief technology officer, says, “The most important thing we do with our Web Analytics tool now is A/B testing, putting a different version of our Web site on each of our front-end servers to see which one performs best. Sometimes the best design is not the best performer,” he notes.

What’s Available

In this article, I’ll take a brief look at several of the most popular Web Analytics tools and indicate the range of functions these applications encompass, including tracking support for campaign management, e-mail management and customer-behavior analysis. Though Web Analytics started out as clickstream tracking, it’s matured into a more robust set of e-commerce analysis and support tools.

* Omniture SiteCatalyst offers a Campaign Manager that tracks active e-mail and online marketing efforts, showing revenue generated by each. You then can drill down to examine what customers purchased. If the system’s thousands of standard reports don’t do the trick, there’s a Site-Catalyst Report Wizard that easily can create new ones, such as sales generated by paid search placement vs. natural searches. You also can put any report on a convenient dashboard with real-time gauges that give you a comprehensive view of site performance.

Omniture is equally adept at producing conversion funnels that show, step-by-step, how customers or prospects navigate a site from arrival to purchase or cart abandonment. If you can eliminate the hurdles or steps that cause the most abandons, you’ll obviously increase revenues, sometimes dramatically. With Omniture, it’s easy to change the time frames and dimensions you want to analyze.

For those who’d rather do their analysis off-line or merge data from other sources, all of SiteCatalyst’s data easily are exported to Excel.

Content producers will find Omniture’s click map overlay an easy-to-use method to show how many people clicked on a page link or what link they clicked on to arrive at the page. The system also offers customer segmentation and cross-sell analysis, which is easy to set up and use. For more, visit: www.omniture.com.

* WebSideStory’s HBX On-Demand Web Analytics is a hosted solution that’s somewhat more complex to set up than Omniture, but it offers a vast array of dashboard gauges measuring keywords, banner ads and other critical functions. Marketers and Web developers can set up their own reports with little difficulty.

It’s hard to beat HBX for report detail, sales reports, navigation measurement and campaign analysis. You can use the system to analyze specific product sales against revenue referenced to campaign, search engines, geographic regions and other variables.

HBX’s funnel analysis is capable of pin-pointing customer behavior down to the level of the field in a form where the customer left the site, as well as all other entry and exit points. Context-sensitive analysis lets you view trend information such as visits per page, reporting on referrers for a given page and how their visits changed over time.

HBX shows which pages were visited most often and how frequently by each visitor. Customer-purchase analysis can show not only which SKUs and brands sold best, but also which promotions were most effective in achieving the sale. And it makes suggestions for cross-sell and upsell offers based on customer purchase histories (i.e., affinity analysis).

The system even can “auto-sense” campaigns, and it supports a number of customer-cost models, including pay-per-click and cost-per-acquisition. And you can do on-the-fly segmentation, for example, query all customers in Canada who browsed for a certain product type and didn’t find it.

Exporting data from HBX to Excel is a snap. You can run exports on a set schedule and automatically e-mail reports of the results. If needed, there are application programming interfaces to allow third-party service providers to access data for specialized analysis. For business-to-business merchants, HBX integrates with Salesforce.com to follow up on sales leads in a timely fashion. For more, visit: www.websidestory.com.

* Another hosted solution, Coremetrics, is strong in marketing and merchandising, and it’s relatively easy to set up and maintain. Smart tags placed by the user, for instance, automatically can sense which marketing campaign drove each visitor to specific Web pages, without predefining campaigns or segments, as most other systems require. A “virtual analyst” helps you set up most reports and explains how they work.

A primary feature of Coremetrics is LIVE, which stands for Lifetime Individual Visitor Experience. This feature stores all clicks and activities for each customer and visitor.

When users log in, they see a user-definable snapshot of how the site is performing (e.g., sales vs. forecasts by product type; month-to-date sales vs. targets). Detailed reports can be aggregated into groups such as top-line metrics, marketing, merchandising and segmentation, each with its own drop-down lists to help you drill into the data.

You also can select the type of data you want to see superimposed on Web pages, from orders by campaign to sales by time period.

A marketing management module allows even more data slicing and dicing, particularly when you’re analyzing conversions by campaign, e-mail conversions and average order value. As with HBX, you can export these data to Excel on a predefined schedule and automatically e-mail it to designated managers.

LIVE is particularly good at letting you use a Wizard to create customer profiles based on multidimensional recency, frequency and monetary value analysis. For example, it lets you tally: all visitors who returned to the site more than x number of times; bought more than a specified value of product types; and came to your site from a Google keyword campaign. As noted above, if you then want to drill down and analyze any one customer’s activity in that group, you can do so. For more, visit: www.coremetrics.com.

* WebTrends NetIQ, which can be hosted or installed in-house, offers many of the same functions as the other packages, although you may find it more difficult to set up. Tracking codes and campaigns must be configured manually, for instance. But you still can schedule data offloads to Excel using SmartReports, which also imports product cost or other data, creates pivot tables, and can update from live data. You also can integrate with third-party applications like e-mail. Standard reports can be clustered and accessed via a table of contents.

And if you don’t want to offload to Excel, NetIQ lets you customize standard reports easily.

The Campaign Drilldown module measures campaign performance, including e-mail campaigns, showing the number of unique visitors per period, conversion data and revenues generated. You can distinguish between paid search and organic search, and even can see your search engine ranking and return on investment (ROI) for selected keywords, and compare that to rankings of your competitors.

Scenario Analysis in NetIQ is a robust tool for funnel analysis. For e-commerce optimization, Product Drilldown aggregates order, sales and revenue figures by product groups, with one-click drill-down for further details. For more, visit: www.webtrends.com.

Conclusion

The more you know about what you’d like a system to do, or at least can prioritize several objectives, the more likely you’ll choose a system that significantly enhances your e-commerce business. Once you’ve implemented it, focus on your original objectives before you take a broader approach — that is, walk before you run.

Costs & Benefits

Prices for Web Analytics systems generally range from $10,000 to as high as $60,000 or more per year, depending on hosted vs. non-hosted, optional modules, support contracts and configuration.

The return on investment (ROI) can be stupendous, though. Companies that are undertaking the following can tell you that the Web Analytics tools they use for these vital activities are well worth every penny:

- fine-tuning keywords and keyword bidding,

- discovering which pages customers and prospects really visit, and from where,

- optimizing links and revenues per click,

- doing product category analysis,

- fine-tuning site search terms,

- looking realistically at conversion funnels,

- diagnosing points of abandonment, and

- taking a hard look at campaign profitability.

Of course, you don’t have to use all of a system’s features, or even most of them, to get an impressive ROI.

Ernie Schell is author of “The Guide to Catalog Management Software” and president of Marketing Systems Analysis, Southampton, Pa., which helps catalog companies specify and select order processing software. You can reach him at (215) 396-0660.