The Weather Company · B2C · weather forecast app and site
Keep tests running past a quick win until they've seen different weather
Unknownexperimentation
What they did
Every test starts with a clear hypothesis and one well-defined KPI. Tests cover things like button color and trial length. Even though traffic lets some experiments reach statistical significance within 24 hours, they keep them running long enough to see results across different weather conditions and other outside variables that drive purchase behavior, before trusting and shipping the result. Invested in the data foundation first to make testing faster.
In their words
Sub Club by RevenueCat · How to Optimize Trial-to-Paid Conversions — Rachel Chukura, The Weather Company“we know that we have to run experiments for long enough to be able to, because it impacts our business, see what happens in certain weather conditions”
“having a clear hypothesis to start is really important, as well as well-defined key results.”
Related topics
Get tactics like this every Monday
New tactics from the week's founder interviews, each linked to where it was said.
More from this episode
- The Weather Company: Keep the core app free with ads; charge power users for extra radar and hourly dataDrove the subscription business's early growth; no numbers given
- The Weather Company: Score each user's chance to subscribe and use local weather to time offersNo numbers; guest warns untuned models decay as conditions change