Guides to measuring what your ads cause
Ad reports tell you which sales your ads got credit for. They can't tell you which sales your ads caused. These guides explain the published methods for measuring that, in plain words, with examples from Amazon brands that run Meta ads.
Start with what incrementality is. To see the methods put together, read how lift is calculated. To try the numbers, use the free calculators. To look up a term like ROAS or iROAS, use the glossary.
Other Spillmark guides stay off this index on purpose. Find them from Google, an assistant, or a link inside a related page.
- What is incrementality?The sales your ads cause, not the sales they get credit for, and why the two differ.Read the guide →
- Attribution modelsLast click, multi-touch, data-driven and view-through: how each splits credit, and what none can see.Read the guide →
- What is ROAS?The formula, break-even ROAS, and how reported ROAS differs from incremental and marginal ROAS.Read the guide →
- Holdout groupsThe people or places that don't see your ads, and how they show what the ads did.Read the guide →
- Geo testing (geo lift tests)Measure ads by turning them off in some regions and comparing sales.Read the guide →
- Matched market testingHow to pick test and control regions that normally sell in step.Read the guide →
- Difference-in-differencesThe before-and-after comparison behind most geo tests, explained with sales.Read the guide →
- Synthetic controlBuild a stand-in for the test region from a blend of others. The method behind GeoLift.Read the guide →
- Marketing mix modeling vs. testingWhen a model is enough, when you need a test, and how they work together.Read the guide →