News for hackers. Stuff that matters.
Re:Show HN: I created an algorithm to help me achieve Produc... by mygo on Wednesday October 30, 2019 @10:03AM (#21397818)
define a cohort.

Say, people who signed up during a given month = 12 cohorts per year

Or people who signed up during a given week = 52 cohorts per year

Then, given a set of cohorts, what percentage of people per cohort are retained over time?

Re:Show HN: I created an algorithm to help me achieve Produc... by StavrosK on Wednesday October 30, 2019 @10:06AM (#21397860)
Ah, so basically the GP is saying "your users should be sticking around"?
Re:Show HN: I created an algorithm to help me achieve Produc... by koonsolo on Wednesday October 30, 2019 @11:34AM (#21398863)
At some point, no more users should leave. For example after registration week 1, retain 60%, week 2 30%, week 3 10%, week 4 1%, week 5 0.1%. In that case, you don't have PMF. Unfortunately my numbers look a lot like this :,(

But assume week 4 has 10% and week 5 has 10%, that curve flattened out at 10%, and so you have product market fit.

Re:Show HN: I created an algorithm to help me achieve Produc... by mygo on Wednesday October 30, 2019 @06:09PM (#21403394)
> Unfortunately my numbers look a lot like this :,(

Fortunately you have a mathematical equation at your disposal that helps tell you this so that you can feel confident in trying something different and seeing how that effects PMF