Small steps to fight that big product analytics mess
No secret, that data is the most valuable resource in the business world. A lot of product companies desire to be data-driven and are trying to collect product data. But only a small portion of them is gaining real value from their data and making use of it to improve their product.
The sad truth is: implementing an efficient product analytics process and creating a transparent, data-driven culture across an entire team continue to be extremely hard problems for companies.
While working with various startups, no matter if it’s a young app with one team or a big product with many functional teams — I’ve noticed that all of them have a common enemy: a mess in analytics. That means, they’ve really made an effort to set analytics tracking up and use its data, but there’s no structure, no single source of truth, no one understands what’s really tracked, what means what and how to find the data needed.
This mess generates a chain of other consequent problems:
- Teams don’t spend enough time or pay enough attention to their product analytics. That’s understandable, they need to iterate on the product and deliver fast and just can’t afford to get stuck. But in reality, it’s just that mess, that makes everyone mistakenly feel product analytics is too difficult and time-consuming.
- The other problem that the mess produces — data and analytics feels overwhelming. When there’s no plan and structure, people start collecting all the data possible and eventually get lost in this data ocean. In reality only 20% of what’s tracked may be needed. Remember the Pareto principle? 20% of data can answer 80% of questions.
- A big challenge for many teams is a lack of visibility into the state of their data collection. They rely on manual testing, broken charts, and instinct to validate their product analytics. This results in teams lacking trust in their analytics.
- Consequently, there is the wrong perception that you need a specialised resource like a data analyst or a data scientist to get started.
Imagine, such problems could have never existed. The key to this is considering how to organise analytics to be tracked consistently and maintain clean data from the get-go. Of all the product teams I’ve worked with, only one didn’t have these issues— the one, for which we implemented structured product analytics process and data tracking from scratch, so they just haven’t had a chance to meet the analytics mess. After a couple of months the team was using their data for improving their product instead of spending precious time and resources for rebuilding and restructuring their tracking.
The last thing you want after investing time, energy and money into an analytics platform is for the data to become too messy and untrustworthy to use. But if it did happen, there are still some small but useful things you can do now with your team to make your analytics things better.
- Bring order and consistency to how you name your data endpoints/events. Think of a naming format that will be used for all endpoints. Write it down in some place where everyone in the team could see it. For example:
[Object] + [Action in the past] + [optional: Context]: File Uploaded To Profile. Then use it every time a new data endpoint is about to be added. I had one client who had 3 different events for one action of user signing up in the app: completedRegistration, UserRegistered, FirstSignUp. When I asked what’s the difference between them if there is any, no one in the team could answer me. That also leads us to the next point. - Review the list of events and parameters you are tracking, deactivate duplicates and ambiguous endpoints you won’t likely use. Remember that less is more. You don’t have to be overwhelmed with your data. The best approach is to align the data you track with the current product goals and metrics.
- Structure your tracking list — add there descriptions, sources and other info you may need to enable everyone to have a clear overview of the data collected and to understand what means what. Some analytical tools, like Amplitude or Mixpanel have data management feature inside, but if that’s not your case, a well-organised Excel sheet or a Notion page will work better than nothing.
- Brainstorm with your team and establish the process of adding new endpoints. Put you product analytics process on stream. It can be useful to add statuses next to events to track the implementation progress so everyone would be on the same page.
- Grow data-driven culture in a right way. Don’t use charts and data alone — share them with the rest of the team. For example, you can set up regular sharing of important charts to Slack or other workspace. Let the whole team discover how the product is doing and let them see the value from product data and the work that’s been made to set it all up. Also someone might notice some insights you couldn’t even think about.
To make these exercises, there’s no need to dig deep into product analytics, leave all your other tasks and spend hell of time on that. The hack is to start small but meaningful. You’ll soon see how it will ease the struggles with product analytics and bring your team closer to the goal to be truly data-driven and make the most use out of your product data. Let’s put the mess in analytics to an end!
Wish everyone good-quality data and valuable insights 📊
