

Discover more from The Product-Led Geek
Why is it important for product and growth teams to find ways to fail fast?
Assumptions:
Failing = learning (don’t bork your experiments)
25% mean experiment win rate (75% of hypotheses disproven)
20% baseline metric
5% mean relative uplift for winning experiment
Team A runs 1 test per month
Team B runs 2 tests per month
Let’s see what happens over the course of 24 months.
It’s an overly simplistic example, but illustrates the point:
Team A lifts the metric from 20% to ~27%
Team B lifts the metric from 20% to ~36%
What’s the lesson here?
Prioritise learning
Experimentation is relatively expensive. You may or may not have sample size volume to go faster with formal testing. So diversify the testing techniques you use to get valid learnings faster.
Minimum viable tests (Painted door etc)
Observations
Prototype studies.
Surveys
…
Ask yourself - what’s the cheapest, fastest way we can test our hypothesis?