Follow Datanami:
March 26, 2012

Of Tags and Targets: When SaaS Falls Short

Nicole Hemsoth

When it comes to drilling down to the coveted real-time “nitty gritty” of customer interaction data, or to spinning up swift, tailored action to usher in new engagement, web-based businesses are faced with no shortage of software options—and accordingly, endless potential for ugly complexities.

As the scale of mega-retail and content-driven sites continues to drive ever-upward, in part due to the added influx of diverse data types from the mobile web, the need for fine-tuned tools to analyze site movement and engagement are more present, but harder to enact for a number of reasons. This is where tools that have been present for years are finding a new lease on life and viability, including tools like tag management systems.

In a recent whitepaper, “The Future of Digital Measurement and Personalization” Gary Angel, President and CTO of Semphonic, an independent web analytics consultancy, discussed the convergence of two trends in big data—the shift to SaaS-based analytics and the need for a more cohesive, organized data management and warehousing strategy to accommodate real-time analysis demands.

 ;In the middle of these two movements are evolving technologies, including the aforementioned tag management systems, which help analytics gurus navigate the finer spaces between customer and use interactions that Angel feels current web-based analytics solutions aren’t providing room for.

According to the Semphonic CTO, “Most web analytics solutions don’t deliver the capabilities necessary for many forms of advanced analysis. Predictive modeling, data-driven segmentation, regression and correlation are unavailable in off-the-shelf web analytics tools.”

When it comes to making use of customer data to drive engagement and interaction, there are limitations when the data is not aggregated beyond basic levels. Angel says that many organizations are finding that both analytics and targeting are best served when online data is combined with offline data and enriched with third party data sources. In other words, while it may be tempting to look at a data warehousing project with an eye on simplicity via simply running the feed from the current web analytics operation to get that nice neat report each night to the servers, there could be drawbacks.

As Angel notes, the biggest problem with this “simple” approach is that real-time uses of the data are limited. While he says that “this has very limited analytic impact (except in a few special cases such as retail offer optimization) it has a profound impact on your ability to do targeting, personalization and remarketing.” Best put, you can’t use the data to its maximum potential if there is a 24-hour lag time that follows it.

In addition to this and the problems of moving data with flaws that are obscured in aggregated reporting systems (“bad data”) into the warehouse, there are also problems with specifying the nature of the data for richer analysis. As Angel writes, “Web analytics tags are early all page-based and capture little or no information about key elements like scrolls or form field fills. These limitations are unfortunate when it comes to some of the most important targeting and re-marketing functions in the warhouse.”

Building data warehousing approaches isn’t easy, especially because of the diversity and volume of the data. As Angel points out, however, “building a good data model for digital data is non-trivial. It’s a task that is, if anything, hampered by the reliance on big data systems and the belief that all data should live in its most detailed format. Data in this form is very hard to use and understand and is, of course, totally unsuitable for real-time decision making.”

Angel provides these insights in the context of tag management and enterprise measurement, but nonetheless, even without this layer of specificity, these points could very well resonate in any number of enterprise scenarios.

According to Angel, the tag management system market has emerged in full force to accommodate these lags in SaaS-based analytics, especially since tags can be difficult to implement, integrate into the existing vendors landscape (and it’s not like switching vendors is a simple task) and finally, hard to manage and create true value from without massive governance measures in place.

This led us on a bit of a quest ; to investigate just how robust the tag management market is compared to say, a few years ago wherein we stumbled on this excellent overview of the challenges both created by and addressed by tag management systems from Ensighten’s Josh Manion:

In line with what Manion is stating (if you were able to take him at all seriously what with the wig) Angel says that SaaS-based analytics vendors are focused on providing “the best set of reports that meet the needs of their clientele—a kind of least common denominator approach to measurement. As a consequence, web analytics tools provide little or no access to detail data, little or no customer-level analysis and little or no ability to do advanced analytics or customer-based testing.”  ;

He says that many organizations that could benefit from the real-time turnaround of this specific information are looking into the options of tag management systems….and that this in turn is driving more interest in further personalization and targeting programs that can come of such segmentation.

According to tag management system vendor Satellite, tags have become a “monstrous issue” for enterprise users. The company, which caters to big web-based retail and content businesses says that deep analytics and meaningful ROI can only come from in-depth understanding of an audience, “to see where they are struggling, and put changes or tests in place to improve user experience and conversion propensity.” They claim users can’t get that information from standard analytics tools from media data and claim to have built a solution that will be able to load any type of tag or analytics beacon on any user interaction within a page, no matter how many tags are involved.

The thrust of Angel’s exploration of this broad topic in the whitepaper is really on the use of tag management systems. However, in the midst of his statements about the role of tags as central to an alternative collection infrastructures for more robust collection and data use is a much more important message—enterprise analytics, especially at the SMB level—can benefit from some fresh approaches to mature technologies, including the use of tags in architecting a data warehouse strategy.

Related Stories

Startup Wrangles Machine Learning for Retailers

Teradata Rakes In New Retail Wins

Revolution Analytics Sparks New Uses for R

Datanami