News for hackers. Stuff that matters.

How long do I run an experiment for?

I'm always reading about "failing fast" and validating an idea, but I can't wrap my mind around how long it takes to validate a concept.

It seems like most startup post-mortems include an element of "incorrect validation" or not asking the right question or not asking that question for long enough.

Are there any good mental frameworks to read about picking the length of a study/experiment?

( 4 comments | 5 points | on HN )
Re:How long do I run an experiment for? by anthony_franco on Monday September 08, 2014 @12:35AM (#8283458)
A great, actionable guide I'd recommend is the book Running Lean.

But knowing is half the battle. By this point everyone should know about Lean Startup and Customer Development. The true test is having the discipline to apply the principles rigorously and not cut any corners. That's where most entrepreneurs stumble, I think.

Re:How long do I run an experiment for? by Ryel on Monday September 08, 2014 @01:01AM (#8283498)
Test the idea until you either find a result (good idea or bad) or you stress your limits (financial/emotional).
Re:How long do I run an experiment for? by Animats on Monday September 08, 2014 @02:39AM (#8283658)
Read "Why Experiments End", which addresses exactly this question.
Re:How long do I run an experiment for? by hganesan on Monday September 08, 2014 @01:27PM (#8286007)
Did you mean "How Experiments End" by Peter Galison? It looks perfect for helping answer this question.