Meta Ads ROAS on Shopify: The Profit Reality Check

A 3x ROAS sounds like a healthy paid social program. On most Shopify stores it is actually break-even on first-order economics, and a meaningful percentage of the orders are not even incremental. The gap between the ROAS the Meta dashboard shows and the actual profit per acquired customer is the single most common reason Shopify brands grow revenue while quietly losing money. This guide walks through the math, the attribution overstatement, and the target setting that ties ad spend to actual contribution.

The Meta Dashboard Shows Revenue, Not Profit

Meta Ads Manager reports purchase value, which is gross order total before discount, COGS, fees, and shipping. A 1,000 dollar ad spend that drives 3,000 dollars of purchase value reads as 3x ROAS, and that is the number most teams optimise against. The same blind spot we cover in how to calculate true profit on Shopify applies to that 3,000 dollar number: it is revenue, not profit.

Run the math through. Of that 3,000 dollars:

• Discount codes used at checkout pull off roughly 8 to 15 percent on average for a typical brand. Call it 300.
• COGS at 35 percent of net revenue takes another 945.
• Payment fees at 3 percent of customer-paid total take roughly 80.
• Actual shipping cost runs 6 to 8 percent of revenue, around 200.

What remains is approximately 1,475 of net contribution. The ad spend was 1,000. The real first-order profit ROAS is about 1.5x, before any fixed cost allocation. Whether that is healthy depends on what the customer does next.

The LTV Argument and Its Limits

The case for running first-order ad spend at near break-even is repurchase. If the acquired customer comes back two, three, or four times, the contribution margin on those subsequent orders pays for the thin first-order economics. The arithmetic looks like:

Customer-level profit = (First-order net contribution) + (Expected repeat orders × Average repeat contribution) − (CAC)

The number that matters is how confident you are about the expected repeat orders. Three common ways the LTV argument breaks:

• Repeat rate optimism. A category average repeat rate gets applied to a paid-acquired cohort whose repeat rate is materially lower. Calculate the repeat rate on your own paid-acquired customers, not the blended store rate.
• Discount-acquired customers churning faster. Heavy first-order discounting attracts deal hunters whose retention is structurally worse. The repeat math built on full-price customers does not apply.
• Payback window mismatch. If the LTV argument needs 18 months of repeat to break even and your cash conversion cycle is 60 days, the model is technically profitable and operationally illiquid.

Setting an honest LTV target requires the per-cohort lens we describe in profit per order vs revenue. The blended store number is too coarse for paid acquisition decisions.

Attribution Overstatement on Retargeting

Meta's default attribution window captures clicks for 7 days and views for 1 day. On prospecting campaigns to new audiences, the attributed conversions are usually close to incremental. On retargeting campaigns to existing site visitors, a significant portion of conversions would have happened anyway.

Two practical adjustments:

• Separate prospecting and retargeting in your reporting. Run them as different objectives with different targets. Retargeting profit ROAS that looks 6x is often 2x once you remove the customers who would have bought without the ad.
• Periodic geo holdout tests. Turn off retargeting in one region for two weeks while keeping it on elsewhere. The lift difference is your real retargeting incrementality. Most brands that run this test cut retargeting spend by 30 to 50 percent without losing meaningful revenue.

Setting Real ROAS Targets

The target depends on what you are buying with the ad spend. Three common modes:

• Pure new-customer acquisition. Target a profit ROAS slightly above break-even on the first order, betting on repeat. Most brands land at 2x to 2.5x revenue ROAS, which works out to roughly break-even contribution after costs, justified by an expected 1.5 to 2 repeat orders.
• Mixed prospecting plus retargeting. Higher revenue ROAS, often 3x to 4x, but the profit ROAS depends heavily on the mix. Use a blended target only if the underlying mix is stable.
• Repeat-focused campaigns. These should run at a much higher profit ROAS, often 4x or above on contribution, because the LTV argument no longer applies; the customer is already in the LTV count.

The trap is treating revenue ROAS as the target across all three modes. A 3x revenue target is generous for prospecting and absurdly low for repeat campaigns. The mix has to be measured and the targets set per mode.

Connecting the Ad Spend to Per-Order Profit

The operational improvement that most changes the conversation is tagging every order with its real profit at checkout, then attributing those orders back to the originating campaign. Two things become visible:

• The actual profit ROAS by campaign, not the headline revenue ROAS.
• The orders that hit a profit floor for the wrong reasons (discount stack on a low-margin SKU, expensive shipping zone, payment fee outlier) and should never have been counted toward a positive ad ROAS in the first place.

The same flagging logic we describe in how to identify unprofitable orders before they ship applies; the only thing that changes is that the flag is now an input into a paid social attribution model. And the discount levers that show up in the ad creative interact with the same cost structure we cover in how discount codes are killing your margin, except they hit at scale on every campaign send.

Brands that run this loop tend to cut total ad spend by 10 to 20 percent in the first quarter while holding revenue flat or growing it, because the spend that gets cut is the spend that was buying revenue without buying profit.

Frequently asked questions