Geo Holdout Test for Facebook Ads for Amazon Sellers

Spillmark draws 4,000 random splits of the US states and keeps the one whose groups sold most steadily together. It tries 30%, 40% or 50% holdouts and 4 to 12 weeks, then recommends the cheapest option that can detect the lift you care about.

It also states limits plainly: under about 10 Amazon orders a day, most tests can't see a realistic lift, and leakage across state lines makes results conservative.

The design you approve is the one that gets scored. Spillmark records the states, the dates, and the holdout share, then compares Amazon orders by ship-to state. Nothing pauses until you say so.

1. Upload your reports, or connect. Start with the Seller Central, Amazon Ads, Meta and Shopify reports you already download, or connect with read-only access.
2. See profit by channel. Agents show sales minus Amazon fees, product costs and ad spend, with the math behind every number.
3. Plan a holdout you approve. Spillmark checks your order history to see whether a geo holdout can answer at your size. If it can, agents plan a test that pauses Meta in a few US states for 4 to 12 weeks and compares Amazon orders by ship-to state. Nothing pauses without your approval, and each test design is locked before it starts.

• Dashboard (free): report imports, profit by channel, and a pre-test check.
• Starter ($49/mo): for brands doing $0–$50k a month on Amazon. One holdout test at a time, locked experiment ledger, alerts.
• Growth ($199/mo): for brands doing $50k+ a month. Unlimited holdout tests, branded Amazon PPC tests, scored predictions.

Paid plans start with a 7-day trial for $0. Cancel anytime. Check current prices on spillmark.com.