How Spillmark Estimates the Amazon Sales Caused by Meta Ads: A Geo Holdout Method
This technical note explains how Spillmark estimates incrementality: the Amazon orders caused by Meta (Facebook and Instagram) ads, not the sales an ad platform credits. The method is a US geo holdout. Matched groups of states are chosen from ship-to order history, the design is locked with a hash before ads pause, and lift is reported with a 95% range. The note restates the public method, including the estimator, error range, minimum detectable effect, and limits, and points to the canonical page on the Spillmark site. A short public tool, how_lift_is_calculated, is also available on the read-only MCP at https://spillmark.com (path /mcp). Sample figures in the note are labeled as made-up or simulated and are not a brand's measured result.
Authors
- Harry Burman
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-20
- DOI
- https://doi.org/10.5281/zenodo.24410524
- Primary Topic
- Digital Marketing and Social Media
- Type
- article
- Field-Weighted Citation Impact
- 0.00