Measuring the impact of advertising

Jampp Geeks
Sep 19, 2017 · 17 min read

Introduction

The metric can vary depending on the type of campaign:

A statistical minefield

Not enough significance

Bias Selection — Naive approach: Exposed vs Unexposed

Noisy Data — A/B Classical Test

Bias in the delivery of Ads — Normal ad delivery with PSA in Control

The Jampp Method

Let’s define a useful tool we will use from now on:

Pre test Stats — Estimate MDE

Running the Test

Sample Size Re-Estimation

Test Monitoring

Test Evaluation

Next Steps

References

Appendix

Derivation of MDE calculation from the sample size formula:

To calculate the initial MDE

jampp-engineering

Creating technology that helps mobile companies grow

Jampp Geeks

Written by

jampp-engineering

Creating technology that helps mobile companies grow