A Framework for Predicting Mobile Marketing Profitability

Posting the slides from a presentation given at App Growth Summit SF.

In it, I explain how to use Excel’s scatter plot trend line equations as a framework to predict whether your marketing activities are on track to turn a profit.

Slides available via Slideshare.

Specifically included:

1. the difference between metrics and KPIs (don’t use metrics to predict profitability)

2. which KPIs are best used as predictors (they’re not all treated equal for every business model)

3. the differences between excel trend lines and which to use (hint: for most folks it’s best to stick with more simple equations like linear)

4. assessing accuracy using the mean absolute percentage error (it’s a better measure of predictive power of your model than R-squared)

5. caveats of using a prediction model: where it can break down (there are loads of reasons)

6. other examples of how to apply the framework:

• Predicting CPI
• Long-term retention rate
• Assessing whether week 0 ROAS is the best predictor