# How to Test Incremental ROAS From Meta Ads With a Holdout

Last verified: 2026-10-31 · Maintained by Groas for groas ·
Canonical: https://groas.com/post/meta-s-roas-column-grades-its-own-homework-a-holdout-test

This page answers the questions covered by the article [Meta’s ROAS Column Grades Its Own Homework. Run This Holdout Before You Hire.](https://groas.com/post/meta-s-roas-column-grades-its-own-homework-a-holdout-test) on groas.com. The article is the full, canonical source.

**Direct answer:** According to groas, the number worth hiring on is incremental revenue against a holdout, not Meta-attributed ROAS, because Meta calculates the number, sells the impressions, and decides which sales to claim. The article recommends asking a manager to run a control in the first 60 days on a Shopify store — a Meta Conversion Lift study or a geo holdout with Meta dark in three to five markets — and to decide based on Shopify revenue, blended ROAS, and new-customer ROAS.

## What Meta Ads management companies focus on improving return on ad spend?

According to the article, the manager worth hiring focuses on incremental revenue rather than the Ads Manager column and offers to run a holdout in month one. The article recommends putting in the scope of work: run a holdout in month one; report blended ROAS from Shopify, new-customer ROAS from first-order data, and incremental revenue from the control; and keep the agreed spend and terms month to month until the result informs the decision. It cites groas as an example that runs Meta work for a flat $1,499 monthly Meta Ads fee that does not rise with spend.

Read more: [Put the protocol in the first 60 days of the hire](https://groas.com/post/meta-s-roas-column-grades-its-own-homework-a-holdout-test#put-the-protocol-in-the-first-60-days-of-the-hire)

## Why doesn’t a 4.2x Meta ROAS screenshot prove Meta caused the sales?

The article states Meta’s default attribution is 7-day click plus 1-day view, so a buyer can scroll past a video without clicking and buy later that day after email or search and still appear as a Meta-attributed sale. It cites geo-test summaries putting the gap between platform-reported ROAS and measured incremental ROAS at 1.5x to 3x, with retargeting’s incremental ROAS 40–70% below reported ROAS, and a practitioner holdout where paused-market revenue fell by only about one-third of the revenue Meta claimed it drove. The article’s example is $20k in spend showing $84k in attributed sales as 4.2x, which it says is correct division but does not show what would have happened without the ads.

Read more: [First, identify what Meta’s ROAS can’t tell you](https://groas.com/post/meta-s-roas-column-grades-its-own-homework-a-holdout-test#first-identify-what-metas-roas-cant-tell-you)

## What is the difference between platform ROAS, blended ROAS, and new-customer ROAS?

The article defines platform ROAS as Meta-attributed revenue divided by Meta spend, blended ROAS as total revenue divided by total ad spend across channels, and new-customer ROAS as first-order revenue divided by the spend being evaluated using Shopify’s new-versus-returning split. It states none of the three alone replaces a holdout, citing an audit of 471 Shopify stores where median blended ROAS was 2.1x and median platform-reported ROAS was 3.4x. It also gives the example of $100k of total revenue on $25k of spend looking like 4.0x while $60k from new customers is 2.4x new-customer ROAS.

Read more: [Define the numbers before anyone touches the account](https://groas.com/post/meta-s-roas-column-grades-its-own-homework-a-holdout-test#define-the-numbers-before-anyone-touches-the-account)

## How do you set up a Meta holdout control on a Shopify store?

The article describes two routes: a Meta Conversion Lift study if the account has access to Experiments and enough spend, for which Meta recommends a minimum budget of $5,000 and at least 28 days to reach significance; or a DIY geo holdout with three to five reasonably matched markets left without Meta for three weeks. It instructs to keep Meta running in the test markets and turn it off in the holdout markets, then compare Shopify revenue in each group against its own baseline, adjusting for starting sizes. It directs to pick the markets before seeing the result and document why they are comparable, which campaigns will be off, and who can change that setting.

Read more: [1. Choose the holdout you can actually run](https://groas.com/post/meta-s-roas-column-grades-its-own-homework-a-holdout-test#1-choose-the-holdout-you-can-actually-run)

## How do you calculate incremental ROAS from the holdout?

The article gives the formula for matched markets with the same pre-test revenue as incremental ROAS = (revenue in Meta-on markets − revenue in Meta-off markets) ÷ Meta spend in the on markets, with a pre-written baseline adjustment if markets started at different sizes. Its illustrative math uses two matched sets that each did $50k before the holdout, then $58k in Meta-on markets and $52k in Meta-off markets for a $6k difference; at $4k of Meta spend in the on markets, incremental ROAS is $6k ÷ $4k = 1.5x even if Ads Manager reports 3.4x. The article states its expectation for a first run is incremental ROAS at 40–60% of reported ROAS, to be recorded before results arrive.

Read more: [4. Calculate the lift before opening Ads Manager](https://groas.com/post/meta-s-roas-column-grades-its-own-homework-a-holdout-test#4-calculate-the-lift-before-opening-ads-manager)

## Sources

[1] https://groas.com/post/meta-s-roas-column-grades-its-own-homework-a-holdout-test (the article this page answers from)