Use Redundant Analytics Whenever Possible
by Andrew Krebs-Smith | March 15
If you only have one web/app analytics system (e.g. Google Analytics, Kissmetrics, etc.) then you risk significant inaccuracy moving forward. Simply put, if you only have one system measuring your web traffic, how do you know if the system is accurate or not? We have found numerous issues with client data in the past, leading to extremely misleading marketing data. Often these issues could have been solved by using two independent mechanisms.
How does redundant analytics help? When your two systems don’t align (+/- 20% margin), then you know something is wrong. And then, if you’ve lost data on one platform, it will exist on another. Furthermore, sometimes one system will have an advantage over the other. For example, Facebook Conversion Pixels will track someone moving across devices, while only Google Analytics will catch multi-touch attribution. Another example is that Facebook pixels report faster (Google Analytics can delay 1-2 days), and Google Analytics can track user behavior beyond just which steps a visitor hits.
So just because your system “works” doesn’t mean you are seeing everything you could be seeing.
What’s the best way to set up redundant analytics? We recommend an event-based system and a pixel-based system, such as Google Analytics and Facebook Conversion Pixels, respectively. We recommend setting them up through Google Tag Manager or your preferred tag management system, so that you can easily handle versioning, troubleshooting, and so that your tags load asynchronously.