The Ideal Web Analytics Tool

At around the same time I was talking about macro and micro integration in my last columns, an interesting thread appeared in Eric Peterson’s Web analytics forum run by. A poster was commenting on the recent announcement from WebTrends about the ability in the latest release to be able to access the data via a standard ODBC connection. A poster made the comment that “the real ‘value’ in a Web analytics solution is the data collection (tags) and storage/processing (data warehouse)” rather than in the reporting interface.

I must say I agree. A number of vendors responded to the post saying it’s been possible for a long time for their data to be analysed using standard reporting tools as well as via a custom interface. Interestingly, a number of these vendors were some of the early name vendors who had originally built their technology on the foundation of log file analysis in databases sitting on the client’s own servers. Over the past few years, this deployment model has declined in popularity against the “ASP” model, where data are collected, stored and managed by a third party.

In many respects, the ASP model is great. An organization needn’t commit IT resources to managing and storing vast quantities of Web data. That’s someone else’s headache. A Web reporting interface provides easy access to dashboards and standard reports. There is, of course, a trade off between the cost of managing your own data and paying someone else to do it, but there’s another trade off, too: the ability to integrate and analyse your data in different ways.

By the nature of an ASP service, your Web data sits somewhere else. That presents some challenges when it comes to interrogating it in non-standard ways. You either have to set up a data feed to your own databases or you must access the service provider’s databases remotely. Many larger ASP vendors provide a capability to export data from their systems but there’s naturally extra effort, hence costs, involved.

So, as the trend towards data integration continues and organisations want to be able to view their Web data alongside other data sources, where does the value of a Web analytics system lie?

First, value lies in the ability to collect data in the first place. As anyone who’s involved in the field knows, Web data is horrible, messy stuff. There’s lots of it. Data integrity is key to robust reporting and analysis. Web analytic systems are specialist applications for collecting and cleaning Web data whether from log files or page tags.

Second, value lies in the standard reporting of that data in a speedy and reliable manner, in real time, if needed. Web analytic systems have evolved to do that well. But there’s only so far you can go developing dashboards and other reporting interfaces.

Ultimately, the real value lies in the ability to be able to gain insight into visitor and customer behaviour. This requires a number of things:

• the ability to quickly segment and looks at different groups of visitors and customers based on any number of criteria
• the ability to be able to aggregate and summarise data in a way that makes it easier to observe patterns and trends
• the ability to integrate site centric data with other data sources to get a holistic view of the customer

Whilst some Web analytic systems do some of these things quite well, as a rule they tend to fall short on one or more of these attributes. Other tools, such as classical Business Intelligence systems and also specialist data analysis systems, can probably execute these tasks better than the majority of Web analytic systems once the data are in the right format. Therein lies the rub: getting the data into the right format.

As the needs of an organization move beyond the requirement to simply report on Web data and be able to integrate it and analyse it in different ways, the differences between the various systems on the market become greater. Clients I’ve worked with in vendor selection programmes have found it difficult to distinguish between the various vendors until we get beyond the dashboards and standard reports. As your data requirements continue to develop and change, so too may some of the economics around how you manage and access that data.

My perfect Web analytics system wouldn’t be one that could do everything for me. It would be one that would allow me to have fast, reliable access to my core data on the site and also allow me to easily and cheaply access and extract data in different formats. Then I can analyse it in a variety of ways using the best tool for the job. I’ll take a look at some of these other tools for analysing customer data next time.

Till then …