So, you're building a web or a mobile app and looking forward to getting your first users onboard. Before launching, you might decide to put tracking in place to measure how users are working their way through your funnel. And although it can be tempting to track every possible interaction, the following guide to building an analytics specification highlights how reducing your event tracking can make for more reliable data and a more robust product.
After designing a new feature, teams often feel compelled to track every possible click and screen view. Because we might need the data later, right? Likely not. Rather than acting as a safeguard, including every conceivable interaction in your analytics spec actually creates a bunch of pitfalls. Here are some we prepared earlier:
Analysis paralysis: Having swathes of data can be overwhelming. With lots of events to describe a single funnel step, it's easy to become confused about what's important.
Time and money: The more events you want to track, the more time required for implementation and testing, begging the question, ‘Would this time and/or money be better spent elsewhere?’.
Maintenance: Given every product update has the potential to affect existing analytics, the more events in place, the greater the risk. And the greater the QA burden.
Data integrity. Analytics data is pure gold when you are confident in its integrity. But if you notice an event isn't what it should be, it can rattle your confidence in all of your data.
So, instead of beginning with potential user interactions, always start with your goals. Not only will it help you define how best to measure them, but it will also keep you focused on reaching them.
With your goals clearly defined, you can now assess the best analytics tool for the job, which may not be the all-singing, all-dancing tool du jour. For example, whilst Mixpanel is an excellent solution for tracking clicks and journeys on a native mobile or custom web app, it requires pretty hefty setup. On the other hand, an off-the-shelf product like Google Analytics is perfectly suited to tracking web page views and does not require much specialist resource. Reducing event tracking allows you to interrogate your tool belt to ensure you’re always getting that good, goal-orientated data.
Once your tracking specification is tried and tested, be sure to maintain it as a source of truth. Because knowing that your specification is correct will allow you to make future product decisions with confidence. To preserve your source of truth, always consider the chain reaction of new features on existing events. Including analytics considerations and a QA check in your new feature spec will serve as a good reminder.
So, when it comes time to write your next analytics spec, remember less is more. Because there is nothing to stop you from adding events later to fill in gaps or investigate a drop-off in your funnel. And by keeping your first implementation granular, you’ll be rewarded with speed, clarity, and, most importantly, data integrity. Because in data, we trust.