Stop listening to what data may have to say - this isn’t how it works.
Data rarely speaks for itself, which is why most people feel overwhelmed when opening Google Analytics or any type of report from PowerBI, Dynamics, etc.
Contrary to popular opinion, Data is also biased. This is because there are so many different ways to collect, measure, present, and interpret data to ultimately make decisions and take action. As an illustration, think of how survey results dramatically depend on the way you’ve designed the questions. Because all those biases go into the data itself, it is often risky to reuse data for different purposes.
Data, however, is very useful for telling stories. It is a partial and imperfect picture of the past that we hope can inform the future. It is a means to further enhance and illustrate our understanding of the world and ‘what’ is happening within it.
So instead of glaring at the data you already have and hoping to find some insights from it, a better approach is to start with a thoughtful set of assumptions. It’s all about trial and error, just like any other scientific process. Your entire strategy should be designed to generate and collect specific data to then corroborate, substantiate, or validate your assumptions.
Depending on your project, it may look something like this:
- Articulate your most probable or riskiest assumption(s)
- Identify the few data points which would inform those assumptions
- Ship an MVP based on your (revised) assumptions
- Observe the same data points to measure the success of your MVP
- Rinse, and repeat.
Too many people jump from 1 to 3, failing to define how we could test our hypotheses and measure success later on. The same people also tend to collect way too much data, to ensure they find a few interesting things to say in their report, by chance rather than by design.
Using data as a way to test assumption(s) is not easy. It can be scary, take more time than expected and the results aren’t always clear cut. Yet, aiming to do this in a way that is thoughtful, transparent, and accurate is your best chance to increase the likelihood of your project’s success.
