Analytics Project

Google Analytics custom e-commerce tracking

November 2017 | 2 min read
TECH:
  • PHP
  • Javascript
  • Google Analytics
  • Google Tag Manager
  • Vue JS
Next

While employed at The Control Group (TCG) I took an interest in the company’s Google Analytics (GA) account. I was sharing my data findings with my team but was not getting much buy-in because the GA setup and data were not trusted. GA was not a source of truth at TCG for multiple reasons, but the primary was the marketing and optimization teams measured success with different tools. I spoke to the CTO about wanting to update the GA setup, and a few weeks later I was given the challenge and resources to fix it.

Requirements

One requirement was to add e-commerce tracking, but I also wanted to add some additional features. To compile my requirements, I began interviewing the team leaders and noting data they found valuable. I also hired a consultant to assist me with the GA setup and employee training.

Development

GA’s e-commerce tracking requires integration with the application. I wrote a middleware package to expose and format purchase info to the data layer used by Google Tag Manager (GTM). I added offline data with the GA API for tracking subscription renewals and phone cancellations. And finally, I added custom events to be used as goals for monitoring upsell and cross-sell conversions.

Data Audit

After the initial setup, I allowed a few weeks of data to collect before starting a vigorous round of data audits. I worked with the marketing and data team to compare the GA data against other data tools considered truth by TCG. Much to my relief, the new setup was tracking with a below 10% discrepancy (set as my milestone) against multiple sources of data.

The marketing team had a significant interest in the new GA setup and improving the tracking of campaigns and channels. This type of tracking requires the use of UTM parameters on inbound links, and upon the launch of the new setup, 70% of inbound traffic was improperly formatted. To fix this my coworker Andrew Johnson wrote a script that looped thru thousands of links in the redirect platform and fixed all the formatting issues I had observed, and this would eventually reduce the improperly formatted links to below 1%.

I built a serious of dashboards to monitor data discrepancies and created alerts to warn myself and others when/if these events happened. I also created a tool in Vue.js and Node for the marketing team that could validate UTM links with the configured channel groups.

Google Analytics Training

I did several training sessions with various teams, walking them thru the basics of data scope and how to build custom reports.

Conclusion

This project presented many challenges along the way, but I kept chipping away at the tasks. My efforts did not go unnoticed by the leadership team, I was promoted and also selected as a Prell Cup winner, an award given to individuals or teams who you can create a repeatable process for success within the company.