Managing the Analytics
Now let’s turn the table around and examine how a web CMS package can improve your web analytics effort.

To do that, we should first review some recent trends and changes in web analytics technology. In the early days of the web, nearly all analytics tools were “log analyzers.” Webserver software generates log files for each page or image that gets served out to the user. Log analyzer tools parsed those entries and generated basic reports.

This has changed quite a bit with the rise of using page tags for data collection as opposed to (or in addition to) webserver logs. Page tags are either metatags or JavaScript snippets (or both) that collect visitor data and write it to a database, typically in conjunction with placing and reading cookies.

Tag-based solutions are not without their drawbacks, but they collect a potentially richer set of information than webserver logging, and they opened the possibility for hosted (SaaS) web analytics solutions, which increasingly dominate the marketplace. Even WebTrends, the venerable log analyzer of yore, now principally sells a tag-based solution that you can purchase as a hosted service. Google Analytics, arguably the most popular analytics package (because it’s free, and it’s Google), is a hosted, tag-based solution.

Page tags are powerful, but they present substantial implementation and maintenance challenges that customers tend to underestimate. In contrast, webserver logs are relatively maintenance-free.

I know one federal agency that, two years after implementing a web analytics tool, cannot take full advantage of the reports because they can’t guarantee that all their pages are tagged, or that they are all tagged properly. And like nearly every other analytics customer, they underestimated the level of effort required to make essential changes to the tags after the initial setup: to account for new categories, new content, and new or improved reports.

Employing templates can really help here. Templates allow you to bake the proper tag code onto every page as a template element. This allows for central control, and central updating upon modification.

In addition to templating, a CMS can add value to an analytics effort by dynamically generating suitable tag attributes (e.g., JavaScript variables) based on any information the CMS knows about, including the location on the site, A/B testing, the author, or whatever.

Of course, this requires coding in your CMS. You could also modify your CMS to provide an interface for non-analysts to update variables, including tags for specific pages, templates, or areas of a site.

Working Together
You may also want to look for ways to leverage information in your web CMS repository to enrich your web analytics reports. Consider these examples:

What are the titles of the pages in that traffic report?
We’ve all seen reports that demonstrate that “page.jsp?contentID=2345” received more traffic than “page.jsp?contentID=5678.” Okaaaaay. Unless you keep encyclopedic knowledge of your system’s content IDs in your head, those reports are frustratingly opaque. A good analytics tool should be able to look up the title field against content IDs and tell you the real page names.

Whose articles attract the most new visitors? Or lead to the most transactions?
Now things get interesting. Web analysts know that gross visitor numbers often mean little. They look instead for key performance indicators, like new visitors or transactions (even if that’s just downloading a white paper). Perhaps certain contributors’ postings do better than others in this regard. Or, to make it less personal, perhaps some categories of content do better than others. And where are authors and categories stored and mapped to content items? In your web CMS database, waiting for retrieval by your web analytics tool.

Does Google really prefer our newer content, or is that a myth?
Understanding how Google and other public search engines see your site and flow visitors to your pages is a key concern of any web marketer. Your web analytics package can tell you which queries send visitors to which pages, but you might want to perform deeper analysis. As your content ages, does it stop attracting search traffic, and if so, at what pace? Which categories of content seem to attract more searchers? To answer those questions, your web analytics package may need access to CMS data, if only to associate category IDs to real names.

Doing these sorts of analytical mash-ups is rarely simple. Unlike the web CMS market, most web analytics solutions today are hosted platforms, which means you have to deal with network and security issues in transferring potentially sensitive CMS data to a remote analytics tool. As I mentioned above, you need to figure out where the data warehouse resides, and whether your analytics tool is capable of aggregating other data.

And perhaps most importantly, it takes developer resources to build, manage, and maintain these sorts of integrations.

Finally, the biggest hosted platform—Google Analytics—doesn’t have an API at all for integrating non-Google datasets. Presumably the company is working on this, since even its target audience of small and mid-sized customers are increasingly recognizing the importance of integrated analytics.

If you’re like most, and you’ve struggled to get your web CMS up and running, you might not want to hear that your work has only just begun. It’s very nice to have an automated publishing system, but if you can’t measure the results in terms of business value, what’s the point?

OK, so you understand the value of good analytics. Yet what really matters is what you do with your traffic and conversion reports. And here you should consider getting metrics directly to web publishers in the environment where they can actually act on them: their CMS.

Sidebar: The Web Analytics Marketplace

CMS Watch divides the web analytics supplier marketplace into two different categories, based on the licensing/ service model.

SaaS-only Vendors
The following vendors provide their analytics solution only as Software as a Service (SaaS),

  • Google: Google Analytics
  • Coremetrics: Online Analytics
  • Digital River: Fireclick Advanced Warehouse
  • Nedstat: Sitestat
  • Omniture: SiteCatalyst
  • VisualSciences: HBX Analytics

Hybrid: SaaS and Traditional Software
The following vendors provide their analytic solution both as SaaS and as a licensed software. In the latter case, you purchase the software, host it on your servers, and maintain it internally.

  • SageMetrics: SageAnalyst
  • Unica: Affinium NetInsight 7.1
  • WebTrends: Analytics 8
  • 24/7 Real Media: Open AdStream Analytics
  • AuriQ: RTmetrics
  • VisualSciences: Visual Site
  • ClickTracks: ClickTracks


Companies Featured in this Article

CMS Watch

Articles that are part of this Special Section:

[Note: Full access to web content is restricted to paid print subscribers.]

Mashing Up Web Analytics and Web Content Management
Managing content and measuring results have been two distinct domains. Increasingly, the two are converging, as site publishers and marketers seek to take action based on measurable results.

Content That Travels Well: The Global Content Management Challenge
Companies are only as good as the content they deliver, so it pays to go past thinking globally and start structuring content that can play as well in Pakistan and Paris as it does in Peoria.

The Courtship of CRM and CM
Customer relationship management and content management have traveled parallel paths inside the enterprise. Though the two systems have a great deal in common, they rarely interacted, until now.

It’s All About Community: Prerequisites for Web 2.0 Content Management
As companies seek to create stickier, more interactive online experiences, they look to Web 2.0 approaches, creating new CM challenges.

Expert Tips on Implementing an Enterprise Content Management System
Insights into the difficult process of ECM purchase and implementation.