Formulating hypotheses

DId you define the real problem you are trying to solve? Then it’s time to think about ideas to solve this problem.

It’s a good idea to start with what you already know about the problem and it’s context. Do you have results from earlier analyses or experiments, are there research studies done by others around this topic? Use this knowledge to build on, instead of reinventing it for yourself.

A hypothesis (in this context) is what you think will happen if you do something. It should be formulated something like this:

If I do…  (the thing you will change), this will cause … (the thing that will happen), for …. (this group, or this situation), because … (the reason why you believe this will happen) and this will lead to … (the benefit this will have)

Make sure that the thing you will change and the effect are concrete. It’s also important to check if the idea and the expected impact make sense. Is a small change you make in one process for a small group of customers really capable of changing the overall customer satisfaction of your company with 20%?

What’s also really important is to think about how you can prove that you are right or wrong. How can you test your idea? What will you measure? Running experiments is a great way to test if your idea is really as excellent as you believe it is. More on experimenting in a different place.

Two final important things:

  • Formulate in advance when your hypothesis is valid. Is a 5% increase good enough? Any significant result? If you don’t do this in advance, it is very hard to do it once you have your results.
  • Define what you will do when the hypothesis is true and if it’s false. If there’s actually no difference, why bother?

A company called Strategyzer has some great material on this topic, including this video about their ‘Test Card’:



Leave a Reply

Your email address will not be published. Required fields are marked *