One Journey, Three Attribution Models, Three Different Heroes
Last touch, first touch and linear models can credit different channels for one sale. A hypothetical example shows how the model moves social’s share.
Last touch, first touch and linear models can credit different channels for one sale. A hypothetical example shows how the model moves social’s share.
An MMM output shifts with which channels the analyst includes. Here is what to ask before trusting a case study’s modeled revenue or incrementality figure.
A case study’s brand lift percentage needs a disclosed sample size before it means anything. Here is what that number hides and what to ask for instead.
A four dollar cost per lead sounds great until you learn it bought four leads. Here is how to read a cost-per figure that has no count next to it.
A one-tailed test can push a borderline result over the significance line. Here is when that choice is honest, and when it just manufactures a finding.
Best month ever is often just what a growing metric does. Learn to tell an impressive record from an inevitable one, with a hypothetical 12 month example.
A campaign that gained 50 percent then lost 50 percent ended 25 percent down. Learn why an averaged growth rate misleads and how to check one in a case study.
Revenue growth quoted in raw dollars includes inflation the campaign did not cause. Learn the one division that shows how much of a result is real growth.
Agencies report a per-client average or a total across all accounts. Learn why the client count changes what each means, and which questions to ask first.
A backtested result was never delivered, only simulated against past data. How to tell a modeled case study number from a real campaign that actually ran.