6 Comments
User's avatar
Wadnes's avatar

Great experiment! December to May seems long. On average, how long do you recommend an experiment should last?

Andreas Just's avatar

Yes you are right! The experiment was way too long but we thought we just keep it there until the product launches and it brought some solid results. On average I would do experiments in some sort of sprinting, as short as it is effective (from 1-4 weeks)

Wadnes's avatar

Makes sense, I would have done the same too. Thanks for the advice. Working on a few experiment ideas so excited to test it out.

Andreas Just's avatar

Nice one! What do you want to test?

Wadnes's avatar

Identifying 1-3 GTM motions to drive customer acquisition. A quarter is a long time to commit to a motion so I want to experiment on quick signals to identify/test to ensure the motion is worth doing.

Andreas Just's avatar

That's smart... And yes, it's much better than testing a quarter and not having any results