Build A Hypothesis As Your Guide

When you're planning to optimize a north star metric through a new feature, it's useful to build a hypothesis to guide your thinking.

Let's say your product is an electronics classifieds app & you're trying to drive up ad listing volume.

Think about:
- the ideal outcome in measurable terms
- the target user segment
- the value you will deliver
- the proposed solution

So, you're hypothesis could look like:

We believe that simplifying the ad creation process to just taking a picture, recognizing it via AI & auto-generating the details,

will help TV sellers, save time in creating an ad, resulting in an increase of 10% in ad listings on the platform in 3 months.

Great, testable hypothesis!

Not so fast.

Can you spot the glaring assumption there?

That the ad upload process is what's causing friction in listing volume.

It could well be related to demand, category popularity, resale value of TVs etc.

Therefore, you need to first de-risk your hypothesis by validating the assumption.

You do this by developing a cheap MVP.

It could be anything from a survey, data analysis or even a prototype.

Once you attain confidence that the assumption holds true, that's when you start framing sprint tasks for the hypothesis.