📐Attribution & Testing
Proving what worked: attribution, incrementality, experiments and marketing ROI · 15 terms
Attribution
7Sharing the credit for a sale among the ads and visits that came before it.
Deciding which ads and visits get the credit for a sale. It also says how much each one gets.
The rule for sharing credit for a sale. It splits it among the touches that came first.
The first ad or visit gets all the credit for a sale.
The last ad or visit gets all the credit for a sale.
Credit for a sale is shared by many ads and visits. No one gets it all.
How many days a sale still counts for an ad. It starts at the click or view.
A sale after someone saw an ad but did not click it.
Incrementality
5Finding out what marketing really caused, by comparing people or places that saw it with ones that did not.
The extra sales that ads really caused. It leaves out sales that would come anyway.
People who do not see a campaign, on purpose. They show what it really added.
A test that runs ads in some towns but not others. Then it compares sales.
A model of past sales and spend. It shows what each channel adds.
How long an ad keeps working after it stops running.
Experiments
3Tests that show versions of a page to random groups and keep the one that wins.
Two versions, shown to two random groups. You keep the one that does better.
Testing a few changes at once, in every mix. You find the best one.
A test of a page against an exact copy of itself. It checks the test tool works.