In this case study we look at a recent web analytics project where a few things were required:

• Migrating tag manager from Adobe DTM to Adobe Launch
• Set up cross domain tracking across 5 sites of the same company
• Export Adobe Analytics data for processing and visualization in Tableau server

As usual we did an initial audit of the implementation and found a few more things that need work:

• inconsistent data layer and event triggers
• critical reports missing data (unexpected unspecified)
• broken campaign and form attribution paths
• unused reports consuming significant amounts of server calls.

The fixes and migration took about 3 months + 1 of setting up new QA and deployment processes to protect the new implementation. A lot of the time involved waiting and coordinating with the different dev teams since they all had different release times. By the end of the project, the stakeholders were paying $20,000 a year less in server calls, had accurate critical reports in Adobe/Tableau plus a bunch of new reports and integration possibilities. This helped them get a new perspective on the global and regional performance of the sites and reinvest the savings in QA tools to prevent the reports from breaking again. They selected Observepoint for the regression testing and wrote a case study just on that. Besides a happier relationship between the marketers and developers, being able to integrate web analytics into the release cycle meant that product owners didn’t have to worry about code conflicts affecting users in the production site.

I then learned that the agency that was hired to do the original implementation took twice the time and money. It was one of those big global firms and insisted that the reports were accurate, which made the stakeholders skeptical of the potential for improvement. It was only until we fixed the first critical report that was losing 15% of users that they understood the situation. The moral of the story is that if you can’t explain all the data in the reports you are using, keeping an open mind to second opinions and professional audits can help. Do not accept answers like ‘it is like that for everyone’ when it comes to data gaps and always keep improving.