Process
The hypothesis-driven development loop
Six steps for testing a product idea before you commit to building it
The six steps of hypothesis-driven product development, from spotting a real problem to deciding whether to persevere, pivot or kill an idea. Product managers and founders can use it as a checklist for any experiment. The key habit is setting the success metric before the data comes in. Koji's AI-moderated interviews can help with the qualitative steps.
Where this appears
What's in this process
- Observe a problem: Start from a real signal: interviews, tickets, analytics, NPS comments.
- Write a falsifiable hypothesis: We believe X for this segment will result in Y. We'll know when Z.
- Design the cheapest experiment: Climb from interviews to MVPs only as far as the decision needs.
- Run it with a pre-set metric: Define success before you see any data.
- Analyze with humility: Read results as evidence about the hypothesis, not the team.
- Decide: persevere, pivot or kill: Then write down what you learned for the next decision.
Use this image
Free to use in your docs, decks and posts under CC BY 4.0, as long as you credit Koji with a link to this page. To embed it, paste this into your page:
html
<a href="https://www.koji.so/visuals/hypothesis-driven-development-loop-six-steps"><img src="https://www.koji.so/visuals/hypothesis-driven-development-loop-six-steps/hypothesis-driven-development-loop-six-steps.png" alt="Process diagram of the six-step hypothesis-driven development loop: observe, write a hypothesis, design an experiment, run it, analyze, decide." width="1080" height="1080" /></a>
<p>Source: <a href="https://www.koji.so/visuals/hypothesis-driven-development-loop-six-steps">The hypothesis-driven development loop</a>, Koji</p>
