Footwear Advertising Benchmarks 2026: The Break-Even ROAS a Shoe Brand Actually Has to Hit

A footwear brand at $140 average order value, a 55% SKU-level gross margin and a 25% return rate keeps $0.279 of every $1.00 Google reports as conversion value. That puts break-even at 3.59× reported ROAS — not the 1.82× the standard “one divided by gross margin” rule produces, and not a number you will find in any published benchmark, because no benchmark publisher segments below the apparel bucket.

Footwear accounts rarely fail in the auction. They fail in five specific places between the click and the bank account, and each one has a different fix. Start with the symptom you recognise.

Knowing the category average cost per click tells you what other people paid. It does not tell you what you can afford, and those are different questions with different answers. The number that decides whether to spend is your contribution margin after returns — put your own five numbers in and the rest falls out of the arithmetic.

Five of those inputs are yours and no benchmark can supply them. Your SKU-level gross margin comes from a specific product’s fully landed cost including duty, not from the P&L. Your return rate comes from your own refund data, because the published footwear range spans 17–30% and the choice moves break-even by more than a full turn. Your conversion rate has to come from the non-brand paid segment alone. Your realised discount rate across all orders is a different number from the discount depth on promoted items. And your twelve-month repeat rate has to come from cohort data, not from an average.

For some footwear brands the honest answer is no, and no amount of bidding skill changes it. Five thresholds decide it. Three or more failures means the constraint is the business model rather than the ad account, and the correct next move is merchandising, pricing or channel mix — not a bidding experiment.

Google Ads records an order at its full checkout value and keeps it there unless a conversion adjustment is uploaded when the refund is processed. Value-based Smart Bidding then optimises toward the inflated figure. Because return rates in footwear vary sharply by product, by size, by price point and by customer cohort, the optimisation engine does not merely misreport — it systematically learns to buy more of the traffic that returns most, because that traffic looks most valuable at the moment of conversion.

The correction is a multiplication, not an adjustment of attitude. Contribution-margin ROAS equals reported ROAS multiplied by your contribution margin. At the mid-market archetype’s 27.88%, the dashboard flatters reality by a fixed factor of 3.59× — which is why break-even sits exactly there.

On the return rate itself we publish a band rather than a headline, because the sources genuinely disagree. The NRF with Happy Returns puts online returns at 19.3% of online sales across all categories in 2025 Reported. ECDB models US footwear at 19.9% for 2024 Modelled. Single-brand direct-to-consumer footwear is commonly reported at 18–19% and multi-brand assortments at 25–30%. We plan at a 22–28% band for DTC footwear with the NRF figure as an anchored floor, and we make return rate the first input the reader replaces.

Size and fit is the single largest return driver in footwear, and the root cause is width rather than length. Several pages currently ranking for footwear benchmark queries claim the opposite — that footwear enjoys a structural advantage because most buyers know their shoe size. That claim is unsourced and mutually exclusive with every documented return-reason study we could find. Repeating it would make this report wrong on its central mechanism.

The behaviour that pushes footwear above the all-category online return rate is bracketing: ordering two or three sizes with the intention of sending most of them back. Roughly six in ten surveyed UK fashion shoppers admit to it, four in ten of those buy multiple sizes of the same product, and about seven in ten Gen Z shoppers reported over-ordering sizes or colours in 2024 Reported. Bracketing inflates gross orders — which is what Google counts and bills against — while leaving net orders unchanged. It degrades reported ROAS and true contribution at the same time.

Cutting returns from 25% to 18% lifts contribution per order by 17.7% and loosens the required ROAS by 0.54 turns. Every one of the interventions that gets you there is a merchandising decision the owner controls directly, and none of them is in the ad account.

• ✓

  True-to-size guidance written per last, not per brand — the same brand runs differently across lasts.

• ✓

  Width-specific copy on the product page, stated plainly, because width is the driver.

• ✓

  On-foot and size-specific imagery rather than a single studio shot.

• ✓

  Fit reviews surfaced on the product page instead of buried in a reviews tab.

• ✓

  A post-purchase fit follow-up timed to land before the return window closes, not after.

Footwear is sold as a size run — typically eight to fourteen sizes per colourway. A style is commercially dead once the core sizes sell through, even though the fringe sizes remain in stock and remain in the feed. This is the leak that has no analogue in a generic e-commerce report, and it is invisible in any campaign structure organised by product line rather than by size availability.

It costs money twice. Continuing to advertise a broken size run means paying full cost per click for traffic that structurally cannot convert: on the mid-market archetype every one of those clicks is a pure $1.17 loss, and the exposure is the entire ad spend running against out-of-core-size styles. Then a return arriving late in a season lands into that same broken run and is far less likely to resell at full price — which is why the model carries a recovery rate of 0.85 for year-round sneakers and 0.80 for seasonal boots, and why late-season contribution is worse than early-season contribution on an otherwise identical order.

Cost of goods is fixed in dollars while the discount comes entirely out of margin, so discount depth destroys contribution far faster than it reduces price. This is arithmetic rather than opinion, and it is the single most under-modelled fact in footwear paid media.

Two discount numbers get conflated constantly and they are not in conflict. ECDB’s 9.5% US footwear discount rate is the blended realised discount across all orders Modelled; the 25–30% figure quoted elsewhere is depth-when-discounted on the promoted subset. Both can be true of the same brand in the same year.

Seasonality belongs in this section rather than its own, because it is the same arithmetic with the calendar attached. The demand spine is public: the US Census Bureau’s monthly retail sales series for clothing and accessories stores, and the shoe-store series beneath it, show December running roughly twice January with a March-to-July trough Reported. That series is official, monthly, reproducible and permanently citable, and none of the articles currently ranking for these queries uses it.

The cost side moves asymmetrically across channels, and that asymmetry is a budgeting instruction rather than an observation. Meta CPMs inflate by up to 66% across the holiday season and hit $17.70 on Cyber Monday, 138% above the $7.43 annualised average Reported, while Google Ads CPC rose only about 3% at Black Friday. Conversion partly offsets it: fashion converts at 3.3% in November and December against 2.4–2.6% in January and February, a swing of roughly 30% Reported. Shift Q4 budget toward search and away from paid social reach buys, and read the offset before you panic about CPMs.

Three footwear brands bidding on the same non-brand keyword are not playing the same game, and the gap between them is not skill. Order value multiplied by contribution margin sets a hard ceiling on what each one can pay, and the spread across ordinary footwear price points is an order of magnitude.

The premium boot brand can afford 10.4× the acquisition cost of the value sneaker brand, and at an identical 3.0% conversion rate roughly 10.3× the cost per click. They bid on the same words. Against published apparel CPCs — Google Shopping around $0.66, apparel across all campaign types around $1.64 — the value sneaker’s $0.32 ceiling sits below even the Shopping average. That segment is structurally excluded from non-brand paid search on first-order economics, and no amount of account work changes it.

Below roughly $85 average order value, fixed per-order logistics consume over 15% of revenue on their own and paid acquisition at market cost per click is arithmetically closed. The curve is steep below $140 and flattens above $180, which means order value is a far more powerful lever at the bottom of the ladder than anywhere else.

One claim circulating widely is worth refuting directly, because it is arithmetically false rather than merely optimistic: that fashion brands profit at just 1.5× ROAS. Profit at 1.5× requires a contribution rate of 66.7%, which is above the gross margin of every footwear brand on the public markets. At a 65% SKU margin with 25% returns and standard logistics, break-even is 2.80×.

Here are the benchmarks, in full, with their denominators attached. They are evidence for the argument above rather than the product of this report — but they are the reason the argument holds, and every one of them is worth reading with its period and its panel in view.

The reason nobody segments below the bucket is worth stating plainly. The most-cited benchmark publisher in paid search segments only to “Apparel, Fashion & Jewelry” — a bucket containing local retailers and engagement-ring advertisers — and the headline number it reports for that bucket is a cost per lead of $101.49 Reported. The largest e-commerce panel we could find covers twelve industries and has no footwear row. The one genuinely footwear-specific advertising report we located covers the twelve largest UK retailers and sits behind an agency lead form.

One reconciliation nobody else appears to do: paid-search conversion rate for apparel is reported up roughly 15% year on year while retail sitewide conversion is reported down 5.5% year on year across 99 billion sessions Reported. Both are true. One measures paid-campaign conversion rate, the other measures sitewide session conversion. Presenting them side by side without that reconciliation would look like sloppiness rather than nuance.

The Benchmark Spread Index is the highest published value divided by the lowest published value, for the same metric over the same period, across named publishers. We compute it because a brand comparing itself to a benchmark whose denominator it cannot see is not measuring anything, and because the size of the disagreement is more useful than any single number inside it.

Publishing what we refuse to use costs nothing and is more persuasive than describing a standard. Seven figures did not make this report.

1. 1

   WordStream/LocaliQ's $101.49 'Apparel, Fashion & Jewelry' figure used as a footwear CPA. It is a cost per LEAD, from a bucket containing local retailers and engagement-ring advertisers. Misapplying it is the single most common error in the competing articles.

2. 2

   The 18% branded / 82% non-branded split with 1299% branded ROAS against 68% non-branded. It circulates across dozens of pages with no identifiable primary source and no sample size. We could not find a defensible published branded/non-branded split for footwear, so this report does not publish one.

3. 3

   A $10.19B global online footwear market for 2025. Irreconcilable by roughly fifteenfold with a category of that size at 30–35% online penetration.

4. 4

   A Meta-versus-Google clothing CPA gap published in two mutually incompatible forms — '$8.28 lower per order' and '43% below Google non-brand'. On a $109 average order value both cannot be true.

5. 5

   A '30–40% footwear return rate'. It traces to a PPC agency's sales page, and it is roughly double both the NRF all-category figure and ECDB's footwear-specific one.

6. 6

   'A healthy blended MER is 3–5×.' That band is cross-vertical and dominated by high-margin, low-return categories. Applied to footwear it implies marketing at 20–33% of revenue against a 27.88% contribution rate — the spend level associated with a documented public failure.

7. 7

   The programmatic benchmark-content cluster that currently ranks for these queries. Those pages establish that a range exists. They are never an attributed source.

A model built from third-party inputs is worth exactly as much as its agreement with something that actually happened. Allbirds reported a Q3 2025 gross margin of 43.2% and marketing at 35.5% of net revenue Reported. Feed that margin into the formula with footwear-typical returns and logistics and the contribution rate is 18.6% — against marketing at 35.5%, a gap of 16.9 points of revenue before any general and administrative expense, retail rent or headcount.

A scaled, profitable footwear business spends about a tenth of revenue on marketing and carries about a quarter in overhead. If paid media is consuming a quarter of your revenue, the model has already failed regardless of what ROAS the platform reports. Derived from that: the healthy footwear operating band is a blended marketing efficiency ratio of 7–11×, an ad-spend-only ratio of 3.5–4.5×, and paid-attributed revenue at or below roughly half of total.

Everything above resolves into one sequence, and the order matters more than the speed. Each step is worthless until the one before it is done — bidding changes made on unadjusted conversion values compound the error rather than correcting it.

1. 1

   Week one — establish the real number. Export twelve months of refunds by style and by size, get SKU-level margin on your top ten styles from landed cost including duty, and pull your non-brand paid conversion rate as its own segment. Put all five into the calculator above. You now have a break-even you can defend, which almost no footwear brand has.

2. 2

   Week one — stop the measurement lie. Turn on conversion adjustments so refunds flow back to the platform. Until this is running, every other number in your account is describing a business that does not exist.

3. 3

   Week two — reset the targets. Move tROAS above contribution break-even rather than at reported break-even, and accept the volume loss. Expect spend to fall; that is the intervention working, not failing.

4. 4

   Week two — find the broken runs. Rank live styles by depth in the core size band and bid down or pause anything whose core sizes are gone. This is usually the fastest cash saving available and it needs no new tooling.

5. 5

   Week three — attack the biggest lever you control. For most brands that is return rate, not bidding: seven points is worth $6.89 of contribution per order at the mid-market archetype. Pick the three worst styles by return rate and fix width guidance, imagery and fit reviews on those pages.

6. 6

   Week four — write the rules down. Your discount ceiling for paid acquisition, your pass-through rule for landed-cost moves, and how the target changes as a season ages. Decisions made in advance survive the week they are needed; decisions made under pressure do not.

This is the first report in an annual series covering ten e-commerce verticals. The cross-vertical view sits in our conversion rate benchmarks by vertical, and the search-term waste that shows up inside most footwear accounts is covered in the search term module. If a figure here is wrong, tell us — corrections are published with attribution and logged against the edition.