Seven Reasons Your Website Analysis Belongs in a Data Warehouse

October 11, 2009
203 Views

Tipping PointThe state-of-the-art in web analytics has changed, and today’s current generation of SaaS web analytics tools may never catch up.  Today’s SaaS tools (Omniture, Coremetrics, WebTrends, Google Analytics, etc.) are easy to deploy and solve 80% of the challenges faced by most web analysts, but solving the remaining 20% is what unlocks the lion’s share of the business value web analysts seek. The market is at a tipping point – and I’m not alone in noticing it – Gary Angel of Semphonic also wrote recently about seeing this trend at his X Change conference: companies are increasingly moving beyond the capabilities of traditional web analytics tools and moving their clickstream data into data warehouses to achieve increased actionability and advanced analytics in their website optimization initiatives.

In the past, addressing the functional gaps present in most of today’s SaaS web analytics tools has been so difficult that only a handful of the most trafficked websites in the world have dared to try – massive data volumes and high cardinality create significant technical hurdles. But a renaissance of innovation in the data warehousing industry has resulted in new approaches to



Tipping PointThe state-of-the-art in web analytics has changed, and today’s current generation of SaaS web analytics tools may never catch up.  Today’s SaaS tools (Omniture, Coremetrics, WebTrends, Google Analytics, etc.) are easy to deploy and solve 80% of the challenges faced by most web analysts, but solving the remaining 20% is what unlocks the lion’s share of the business value web analysts seek. The market is at a tipping point – and I’m not alone in noticing it – Gary Angel of Semphonic also wrote recently about seeing this trend at his X Change conference: companies are increasingly moving beyond the capabilities of traditional web analytics tools and moving their clickstream data into data warehouses to achieve increased actionability and advanced analytics in their website optimization initiatives.

In the past, addressing the functional gaps present in most of today’s SaaS web analytics tools has been so difficult that only a handful of the most trafficked websites in the world have dared to try – massive data volumes and high cardinality create significant technical hurdles. But a renaissance of innovation in the data warehousing industry has resulted in new approaches to these historical challenges – most notably, appliances whose simplicity of deployment and operation rivals that of today’s SaaS web analytics tools, and whose query performance enables instant answers to the ad hoc questions that web analysts so frequently need to ask. As a result, a new set of website analysis capabilities are now easily within reach, and leading firms are increasingly uniting their online and offline data in data warehouses to achieve increased actionability in the form of:

1) Full touchpoint marketing – Traditional web analytics tools are limited to reporting basic facts about aggregated site usage via reports that have a page-centric perspective. Modern data warehouses can utilize granular record-level clickstream data to create person-centric views of individual visitors in order to optimize every message across every web interaction for every individual visitor – thus personalizing the web experience to best advance them through the appropriate conversion funnel.

2) Integration with actionable systems – Exposing web data to actionable systems (email marketing, CRM, site personalization, etc.) is clearly critical to maximizing the value captured from website optimization initiatives. Unfortunately, most SaaS solutions have pretty limited integration capabilities. And perhaps more importantly, while the sampled and aggregated data that traditional web analytics tools utilize can provide understanding, it can’t drive action. After all, your traditional web analytics tool can’t tell you to email a visitor if that visitor wasn’t in its sample. A data warehouse can let you get rid of sampling, use a complete data set in your analyses, and dramatically simplify integration with actionable systems.

3) Distributed reporting – Complex, automated distribution of hundreds of customized reports to various constituencies is exceedingly difficult, and often impossible, with traditional web analytics tools. Vendors like Business Objects and MicroStrategy, however, have offerings designed for this purpose that plug right into modern data warehouse platforms.

Warehousing web analytics data also enables firms to go beyond the capabilities of traditional web analytics tools and utilize advanced analytics including:

4) Behavioral segmentation – Segmentation is the single most important capability in marketing, and yet the type of segmentation that traditional web analytics tools provide is really nothing more than primitive filtering. Real segmentation involves advanced analysis that goes well beyond the capabilities of traditional web analytics tools in uncovering highly marketable audience segments.

5) Predictive modelingPredictive modeling enables marketers to score visitors based on their likelihood of response to various offers and tune their messaging strategies to maximize return in ways that are simply not possible with the traditional web analytics tools.

6) Time series analysis – It is the changes in visitor behaviors that signal opportunities. Unfortunately, with traditional web analytics tools, these opportunities are missed. For a simple example, suppose I told you that visitor X has viewed BMW product information on your website 30 times over the past 12 months. That does not sound like a heavily engaged user. But suppose I then told you that 25 of those page views occurred within the last two weeks! Suddenly we see that visitor X may have recently become an in-market auto buyer. And that is very interesting, because visitor X’s value to automotive advertisers has just increased significantly.

7) Data visualization, exploration, and discovery – Traditional web analytics tools are no match for tools that are custom-built for data visualization rather than for web analytics, like SAS and Tableau. A data warehouse makes it possible for an analyst to utilize any best-of-breed visualization tool they choose.

The new technological landscape in data warehousing presents a compelling opportunity for web publishers to create competitive advantage. And the seven reasons listed above just begin to touch the surface of the topic (there are actually more than seven reasons). For additional detail, I recommend this white paper authored by Gary Angel and June Dershewitz of Semphonic – it’s an exceptionally good read (though I am biased since I work for Netezza and collaborated with Gary and June to create it): “Warehousing Web Analytics”.  You can also check out this recording of a webinar we did on the topic here:  “Warehousing Web Analytics: Transforming Your Business with a Best-of-Breed Approach”.

Photo credit:  Max Z