Your retargeting campaign is reporting 8x ROAS. Your store barely broke even last month. According to Cometly’s industry analysis of holdout test results, 75% of conversions credited to retargeting campaigns would have occurred without the ads. Platform attribution doesn’t measure whether your ads caused the sale—it measures whether your ads were nearby when the sale happened. Those are very different things. The good news: you can test this for free using Meta’s own tools, without a developer or a data science team.
Why Retargeting ROAS Looks So Good—And Why That’s the Problem
Retargeting works by targeting people who already showed intent: they visited your product page, added to cart, or abandoned checkout. These are your highest-intent visitors. Here’s the structural problem: many of them were going to buy anyway.
When someone abandons a cart and comes back two days later to complete the purchase, the retargeting ad they saw in between gets the credit. Your platform reports a conversion. Your ROAS climbs. But the question you actually need to answer is: would they have come back regardless?
For high-intent shoppers, the honest answer is often yes. They bookmarked the page. They’re comparison shopping. They’re waiting for payday. The ad may have been irrelevant—or even annoying—and the purchase was happening either way.
68% of multi-touch attribution models over-credited digital channels in 2025 (MarTech Series, 2025). Retargeting is the most over-credited channel of all.
This isn’t a Meta or Google problem specifically. It’s structural. Any platform that shows you ads and also measures conversions has a financial incentive to take credit for as many as possible. Attribution windows (typically 7 days for clicks, 1 day for views) are set wide enough to catch organic conversions. The math almost always favours the platform.
The eBay example from a Search Engine Land field experiment makes this visible at scale: when eBay paused brand search ads, sales remained largely unchanged. Those ads weren’t driving incremental demand—they were capturing it. The same dynamic plays out in retargeting every day, just harder to see when you’re inside the dashboard.
You may be interested in: Every Multi-Touch Attribution Tool Is Built for Shopify: What WooCommerce Store Owners Should Use Instead
What Incrementality Testing Actually Measures
Incrementality testing doesn’t ask which touchpoint gets credit. It asks whether removing the ad would change the outcome.
Definition: Incrementality testing is a controlled experiment that compares two groups—one exposed to ads (the test group) and one held back (the holdout group)—to measure the causal lift your marketing actually creates, not just correlates with.
The result is your incremental ROAS (iROAS): return on ad spend calculated only from conversions that genuinely would not have happened without the campaign. It’s always lower than platform-reported ROAS. In Cometly’s analysis of holdout tests across retargeting campaigns, the true incremental ROAS averaged around 2x—versus the 8x figure platforms typically report. That means 75% of credited conversions were organic purchases the ads had nothing to do with.
Let that sink in. If your retargeting shows 8x ROAS and your holdout test reveals 2x iROAS, you’re spending the majority of your retargeting budget to reach people who were already going to buy. Every pound or dollar redirected from over-saturated retargeting to prospecting or brand campaigns grows your customer base instead of harvesting existing intent.
Attribution tells you which touchpoints get credit. Incrementality testing tells you whether those touchpoints caused anything.
How to Run a Free Holdout Test on Meta
Meta’s Conversion Lift tool is built into Ads Manager and costs nothing beyond your existing ad spend. Here’s how to run a basic holdout test for a WooCommerce store spending $3K–$20K/month on retargeting.
Step 1: Find the Experiments Tab
In Ads Manager, navigate to the top navigation and select Experiments. Choose “Conversion Lift” from the available test types. You’re looking for the option to create a ghost ad holdout—Meta withholds ads from the holdout group entirely, rather than showing them a different ad.
Step 2: Set Your Holdout at 10%
A 10% holdout is sufficient for most WooCommerce stores at this spend level. Meta will randomly exclude 10% of your retargeting audience from seeing ads for the duration of the test. This group becomes your control. Don’t change anything else—same creative, same budget, same targeting.
Step 3: Run the Test for 14–21 Days
You need enough purchase events to reach statistical significance. For stores with 50+ purchases per month from retargeting, 14 days is usually enough. Lower-volume stores should extend to 21 days. Do not pause or adjust the campaign during the test—any change breaks the comparison.
Step 4: Read the Incremental Results
Meta’s results screen shows you purchases in the exposed group vs. the holdout group, normalised for audience size. The difference is your incremental lift. If the holdout group converted at nearly the same rate, your retargeting wasn’t doing much. A Meta study of incremental optimisation found a 24% reduction in CPA when advertisers shifted budget toward incremental conversions only (Incrmntal/Meta, 2025).
Alternative: A Simple Manual Geo-Lift Test
If you’d rather not use Meta’s tool, a geo-lift test is also free. Pause retargeting in two or three smaller markets for 30 days while keeping it active in comparable markets. Compare WooCommerce order data directly. A three-market geo-lift case study by Five Nine Strategy found Meta + Google Ads together drove 13% incremental sales lift—versus just 2.2% for Google Ads alone. The gap told them exactly which channel was earning its budget.
You may be interested in: Why Marketing Efficiency Ratio Is the Only WooCommerce Metric That Doesn’t Lie
The Data Problem That Breaks Incrementality Tests
Here’s where most WooCommerce stores hit a wall they didn’t expect: incrementality testing assumes your conversion data is complete. If ad blockers or Safari’s ITP restrictions are causing your Meta pixel to miss 20–30% of purchases, both your exposed group and your holdout group appear to convert at lower rates—but the holdout group may appear relatively worse, making your retargeting look more effective than it is.
31.5% of global internet users run ad blockers (Statista, 2024). Every browser-blocked conversion is an event your holdout comparison can’t see.
The same problem affects any test that relies on client-side pixel data. Safari’s 7-day cookie limit means returning visitors who don’t convert quickly lose their attribution entirely—skewing your holdout result. If your holdout group contains more Safari users, they’ll appear to convert at a lower rate regardless of whether the ads caused it.
Transmute Engine™ solves this at the source. It’s a first-party Node.js server that runs on your subdomain (e.g., data.yourstore.com)—the inPIPE WordPress plugin captures WooCommerce purchase events and sends them via API to the Transmute Engine server, which then routes them to Meta CAPI and all other destinations simultaneously, bypassing ad blockers and ITP entirely. When every conversion reaches Meta CAPI regardless of browser environment, your holdout comparison works from complete data. The test result you get reflects reality, not a partial picture.
Key Takeaways
- Retargeting ROAS is structurally inflated because platforms credit conversions that happen near ads, not because of them— 68% of attribution models over-credited digital channels in 2025 (MarTech Series).
- Holdout tests reveal your true incremental ROAS. Industry data shows the gap is typically 4–6x: an 8x reported ROAS often reflects a 2x incremental ROAS (Cometly, 2025).
- Meta’s Conversion Lift tool is free and available inside Ads Manager under Experiments. A 10% holdout over 14–21 days is enough for most WooCommerce stores.
- Accurate incrementality testing requires complete conversion data. Client-side pixel gaps from ad blockers and browser restrictions skew holdout comparisons, making retargeting appear more effective than it is.
- The real opportunity is budget reallocation. Shifting spend from over-attributed retargeting to prospecting or higher-iROAS channels compounds growth over time instead of harvesting existing demand.
Incrementality testing is a controlled experiment that compares two groups—one exposed to ads (the test group) and one not (the holdout group)—to measure the causal lift your marketing actually creates. Unlike attribution models, which assign credit based on rules, incrementality testing directly measures what would have happened without the ad spend.
Use Meta’s Conversion Lift tool, available inside Ads Manager under the Experiments tab. Set a holdout of 10% of your retargeting audience, run the test for 14–21 days without changing your creative or budget, and compare purchase rates between the exposed and holdout groups. The difference is your true incremental lift.
Incremental ROAS (iROAS) measures return on ad spend calculated only from conversions that genuinely would not have happened without the ad. Regular ROAS includes all conversions attributed to the campaign—including people who would have purchased anyway. iROAS is always lower than reported ROAS because it strips out organic demand.
Retargeting targets people who already visited your store or added items to cart—the highest-intent audience segment. Many of them would convert organically within days regardless of ads. Platform attribution windows (typically 7 days click, 1 day view) then credit those organic conversions to the retargeting campaign, inflating the reported ROAS figure.
Industry holdout test data suggests 50–75% of conversions attributed to retargeting would occur without ads (Cometly, 2025). For a store spending $5,000/month on retargeting with an 8x reported ROAS, the true incremental ROAS may be closer to 2x—meaning the majority of spend is capturing demand that would have converted organically anyway.
If your retargeting ROAS has always looked too good to question, now you have a free way to find out. Run the holdout test. The result—whatever it is—gives you a decision to make with real numbers instead of platform-reported credit.
Start with complete data. Learn how Transmute Engine ensures every WooCommerce conversion reaches Meta CAPI—before you run a test that depends on it.
