The user growth we’ve seen is all over the place, ranging from 15% growth to 3% decline. Trying to predict revenue three months ahead is nearly impossible.
I’m relying on trailing 90-day averages, but with this level of volatility, it’s not very effective.
Same nightmare here with a food delivery app. Monthly swings made forecasting impossible.
Breaking revenue into buckets saved me - new users, returning users, and power users. Each group acts differently. New users are chaos, but returning users? They follow patterns you can work with.
I ditched trying to predict user growth and tracked revenue per user trends instead. Much more stable. Then just ran conservative, realistic, and optimistic scenarios from those numbers.
This video nails the framework:
Your 90-day average works for smooth growth, but with your volatility? Cohort analysis is better. Track how each monthly cohort performs over time.