Why apparel stores overspend when they optimise for ROAS

For apparel, ROAS and POAS drift apart for a specific reason: part of the revenue in the ROAS column is going to come back. Google Ads counts the full basket at checkout. Your bank account counts what's left after returns, and your margin has to cover the cost of processing them too.

How the gap forms

Gross revenue in, refunds out, nothing reconciled

Google Ads records conversion value when the order is placed. Returns never reach the account unless you upload conversion adjustments, so a campaign's ROAS stays fixed at its checkout value however much of that value is refunded later. Coresight Research estimates the average US online apparel return rate at 23.4% for 2025, the figure above. On a campaign returning at that rate, close to a quarter of the revenue behind its ROAS never stays in the business.

Bracketing: one order, several sizes

Shoppers unsure of fit often order the same item in two or three sizes and send back the ones that don't fit. Merchants surveyed by the NRF and Happy Returns reported that 26% of apparel and footwear purchases in the previous year included a bracketed item, against 23% across all categories.

A bracketed order looks like a high-value conversion. Target ROAS bidding optimises toward conversion value, so it can learn to favour exactly the searches and audiences that produce these baskets, and those baskets are built to be partly returned. The bidding is doing what it was told. It was told the wrong number.

Markdown mix

Sale and clearance units sell easily and can post a strong ROAS. But their margin is a fraction of full-price margin, so the same ROAS earns much less profit. An account that averages full-price and clearance products into one target will tend to spend toward whatever converts most easily, regardless of what it earns.

What the gap does to one campaign

Working out your own gap

You don't need the industry average once you have your own numbers. Shopify's sales reports show returns alongside gross sales, so for any period you can work out the share of revenue that came back.

• Kept ROAS is reported ROAS multiplied by (1 − your return rate by value).
• POAS is kept revenue minus the cost of the goods kept and the cost of handling the returns, divided by ad spend.

Measure the return rate on orders old enough to have passed your return window. A rate taken on last week's orders will look better than it is, because many of the returns haven't arrived yet.

How to close it

1. Measure POAS net of returns, judged on orders old enough to have passed your return window.
2. Upload conversion adjustments for returns. Google Ads accepts retractions for fully refunded orders and restatements for partial ones, so the bidding learns from kept revenue rather than checkout revenue.
3. Split full-price and markdown products into separate campaigns, so one margin figure isn't averaged across both.
4. Set ROAS targets from return-adjusted margin. The break-even ROAS for a store that keeps less revenue is higher than its gross margin suggests.

The benchmark above is an industry average. Use it to estimate how much of your reported revenue is at risk before you have your own figure, then replace it with your store's rate from Shopify.