Digital Analytics Data Quality- Your First Lesson

February 20, 2017

Digital Analytics Data Quality- Your First Lesson

Make sure you are very confident in the accuracy of your data before you present it to stakeholders.

Eager Young Digital Analyst Meet the Data Quality Mousetrap

Almost 9 years ago when I was first starting out in my career in digital analytics I was tasked with an analysis of an internal search tool. It was one of those sites that had a bunch of filtering options for search. The visitor could select sizes, colors, product categories, and many other product attributes. I dove head first into the web analytics tool and started to explore the data. I quickly found what I thought was a smoking gun. According to the data, the insight I discovered was that the only internal search filtering option that visitors used was the color filter.

I presented two recommendations in my PowerPoint analysis:
1) Simplify the filtering and just keep the color filter
OR
2) Highlight the other filtering options like size, product category etc. to get visitors to use other filters

Read the story of how I found my career as a digital analyst on Craigslist.

Digital Analyst Meet the Debugger

I spent hours putting the presentation together and trying to make the PowerPoint perfect. Thankfully before I presented this to the client I did a run through with my boss Paul Legutko. We came to the slide about the color filter and Paul said “This seems interesting let me take a look at the site with a debugger and make sure the tags are firing on all the filters and the data is accurate.”

If you’ve dealt with data accuracy and tagging issues before you can guess what Paul found…

Digital Analyst and the Not So Curious Case of the Missing Tags

The color filter was the only filter that was tagged. All the other filters were missing tags. So the explanation was not that the color filter was performing way better than the other filters, the explanation was that the data was not accurate for all the other filters due to the missing tags.

Digital Analyst Put the Horse Before Cart- Data Quality then Analysis

I had wasted hours putting together a beautiful PowerPoint for an analysis that was based on inaccurate data. It was important lesson. From that day forward Charles Proxy became my best friend at work and my debugging tool of choice. To this day I make sure that before pulling any data or doing any analysis I first go to the website and click around with the debugger open to make sure that digital analytics data for the site is being captured. I also use Charles Proxy to QA test Adobe Analytics Omniture and Google Analytics web analytics implementations.

Digital Analytics data is certainly messy and has less structure than other data sources. That said I still think this is a lesson that can be applied to anyone working with data. I urge you to make sure you are very confident in the accuracy of your data before diving into any analysis and making recommendations.

I’d love to hear how you deal with data quality issues in digital analytics. What tools do you use for debugging Adobe Analytics and Google Analytics? What is your process? What advice do you have for a new digital analyst?