David Raab Article Archive

Next Steps for Web Analytics
David M. Raab
DM Review
January 2007

It wasn’t so long ago that Web site analysis was an esoteric specialty whose practitioners focused on the arcane details of log file analysis, ask cookies and page tags. The primary outputs were funnel reports that tracked customer paths through a Web site and source analysis that captured the origins of site visitors. A really impressive system linked the two to measure conversion rates by source—critical for an accurate evaluation of search marketing and email campaign return on investment.

But that was back when the Web was young. Today, pilule the Web is a dominant marketing channel for many companies and an important channel for all. Web activities are increasingly integrated with core marketing programs rather than being handled on their own by a separate staff. This means that Web analytics must be integrated with other marketing analytics as well.

This presents Web analytics vendors with a classic dilemma. They are experts within their niche, now exposed to new competitors with broader functions. Do they expand their own scope to match, or deepen their functionality to remain the best within their narrow field?

This may seem like a purely rhetorical question, since specialist vendors faced with a similar choice nearly always expand their scope. The reasons are well understood: it’s easier to sell new products to existing customers than to find new customers for existing products; company value is increased since potential revenue per customer is higher; and customers prefer integrated suites to best-of-breed products they must knit together on their own. Stand-alone products are also attractive acquisition targets for larger firms seeking to extend their own scope—and while a buy-out is appealing to some entrepreneurs, most prefer to remain in charge of their own business.

But the answer actually isn’t so clear. Because Web analytics is still a fairly new field, the vendors still have considerable room for growth without leaving the niche. Broader scope expansions are certainly possible, and probably inevitable in the long run. But they can reasonably be deferred if a vendor prefers.

A quick review of leading Web analytics vendors shows they are pursuing both approaches, sometimes simultaneously.

Within the realm of improved Web analytics, new functions include:

– offer targeting: since the Web analytics vendors are already gathering information about visitor behavior, a logical next step is to use this information to help select the offers and other treatments most appropriate to specific individuals. This is a somewhat difficult extension, since the predictive modeling techniques used for behavior-based targeting are quite different from the techniques needed for traditional Web analytics. Firms moving in this direction are likely to buy specialists that already possess this technology, rather than trying to build it themselves.

– Web content management: this may seem rather distant from analytics, but the Web analytics tools need to be aware of existing content in order to report intelligently on visitor behavior. So a logical connection is present.

– in-site search integration: this is search within a site, used to help visitors navigate more efficiently. Understanding Web content gives the analytics vendors one connection with in-site search, and understanding customer behavior gives them another. The goal is to return more useful search results by better understanding what visitors are really looking for. In-site search can apply to knowledgebase searches such as customer support questions. But the real financial value comes from product searches. Making it more likely that customers will find products they want obviously increases they chances they will buy. In-site search also draws on technologies similar to offer targeting, so vendors investing in one area can hope to apply the results in both.

– enhanced campaign reporting: Web analytics has always included conversion rates by visitor source, which is the heart of campaign analysis. But marketers have much more sophisticated requirements, including evaluation of multi-step campaigns, real-time reporting, multi-variate testing, summaries by product and key word groups, and tracking of non-revenue events. This is one area where the Web analytics vendors may be able to learn from non-Web marketing analytics. It can also provide the non-Web analytics vendors with a bridge into the Web analytics niche.

This bridge runs both ways. Campaign analysis is equally the most obvious path for Web analytics vendors seeking to cross into other channels. Marketers desperately need to understand how uncoordinated campaigns in different channels are impacting their customers and how intentionally multi-channel campaigns are working. Web campaigns are a critical element in both types of analysis. In addition, there is case to be made that because Web campaigns generate larger quantities of data than off-line campaigns, it’s easier to append non-Web data to a Web analytics tool than to expand a non-Web analysis system. It’s surely true that Web analytics itself must increasingly include data in different formats—not just conventional page hits, but also email and SMS messages, search queries, rich media interactions, social network activities, RSS feeds, and whatever comes next. A system that can do all that shouldn’t have much trouble with direct mail coupons. So even as Web analytics extend their capabilities within the Web marketing world, we can expect them to try little forays into non-Web campaign analysis and perhaps eventually make a serious move in that direction.