Our standard NPS calculations feel off somehow. Users rate us 9-10 but still churn within weeks.
Starting to think we need a different approach to measuring actual satisfaction rather than just the basic promoter minus detractor formula.
What weightings or adjustments have worked for you?
Try asking the NPS question after they use the main feature instead of randomly.
NPS means nothing without tracking what people actually do.
Had the same issue with a subscription app. High NPS scores but people kept leaving after the trial.
What helped was tracking NPS alongside behavior data. Found that users giving us 9s were barely using core features. They liked the concept but weren’t getting value.
Started weighting NPS by usage frequency. So a 9 from someone who used the app daily counted more than a 9 from someone who opened it twice.
Also added a follow-up question about likelihood to use the app next month. That predicted retention way better than the standard recommendation question.
The math got messier but the insights were actually useful for reducing churn.
Standard NPS timing is usually the problem. Most apps ask too early when users are still in the honeymoon phase.
I wait until someone has completed their first real task or hit a key milestone. The scores drop but they actually match who sticks around.
Also helps to segment NPS by user type. New users versus power users will rate you differently for valid reasons.
Stop asking about recommendation likelihood. Ask about perceived value instead. Most users will say they’d recommend something even if they barely use it. But they won’t lie about whether it’s actually solving their problem. I switched to asking “How much value does this app provide you?” on a 1-10 scale after users complete specific actions. Then I weight those scores by revenue per user. A paying customer rating value at 7 matters more than a free user rating it 10. This approach predicted our retention and expansion revenue way better than traditional NPS.