Warning: Creating default object from empty value in /home/chadsummerhill/chadsummerhill.com/wp-content/themes/canvas/functions/admin-hooks.php on line 160

DIY Web Analytics

In Web Analytics an Hour a Day, Avinash Kaushik says,

“…a cornerstone of an optimal web analytics strategy is having a data warehouse environment that allows you to have more-complex data in an easy-to-report-from environment.  Data warehouses are typically very flexible when it comes to importing data from external sources and hence empowering you to do the required analysis with a view into the end-to-end customer experience.”

Before Avinash published his book, my company was a few years into using big-brand web analytics solutions.  Overall, we were just very underwhelmed by the feature set and the flexibility of the software available at the time.  So, we decided to add web analytics to our existing data warehouse.

In House Web Analytics Diagram

In House Web Analytics Diagram

Top Reasons We Built Our Own Web Analytics Solution:

  • We needed data in the right format to use with PPC bid-management (bid-management platforms were not flexible or powerful enough for us)
  • Data integration with our existing data warehouse and outcomes data
  • Access to data (inflexible table relationships—couldn’t ask the questions we wanted to ask)
  • The cost of missed opportunity while we fought with our third-party software.

Being able to ask any question you want with all available data sources is powerful and liberating.  Now we can ask questions at the speed-of-thought without having to go back and create a new sub-relationship table with our vendor (for more money).

It is important to note that the cost of the software was not a deciding factor for us and while it has cost us less to build our own solution we would have gladly continued with a third-party had it not been such painful experience.

What using third-party software did do for us was teach us what we needed (good data & data access) and what we didn’t need (pretty front-end).  It also gave us a good perspective on web analytics in general and a foundation of knowledge that we could improve on.

Some Cool Stuff We’ve Done with Our Web Analytics Data Warehouse

  • Built robust and flexible Bid-management Software (custom designed to take advantage of Google Conversion Optimizer).
  • In-house Landing Page Testing with unlimited segmentation and targeting
  • In-house Conversion Funnel Testing
  • End-to-End view of our customer (first-click, any clicks, last clicks, calls, revenue data, demographic data, etc.)
  • Predictive analytics studies
  • Tracking across our multiple websites
  • In-house Email Testing
  • When someone asks us if we can track something we get to say yes (videos, flash, form-field interactions, downloads, outbound links, etc.)
  • Using Tableau for analyzing our web analytics data
  • Web Analytics data integrated into executive dashboards along with all other important business data.
  • Unlimited Events & Custom Variables

Other Considerations

Of course, building your own web analytics solutions isn’t without its own challenges.  We have to keep up with the crawlers, PPC API changes, maintenance, etc.  Even with these types of challenges, we wouldn’t change our decision to build over buying.

We also run Google Analytics in parallel to our own solution.  Google Analytics is great for answering easier questions and visualizing visitor trends, etc.  It’s also a good place for our less-savvy users to start learning about web analytics.

Final Thoughts

If you are an organization that is struggling with your current web analytics solution and you have already invested in business intelligence and data warehousing, web analytics isn’t that much different from the data problems you are already solving.

It is possible, with the right people, to build your own web analytics solution and have it customized to your specific business needs.

I do believe that the vendors have become more flexible, when it comes to data integration and access, over the past couple of years, but most of them still can’t compete with my data warehouse and Excel or Tableau (when it comes to our business problems anyway).



Even if you don't visit my site on a regular basis, you can get the latest posts delivered to you for free via RSS or Email:

One Response to “DIY Web Analytics”

  1. Mark Rippy September 8, 2010 at 11:20 am #


    This is an excellent explanation of the evolution of in house solutions.


Leave a Reply