By Tom Kató & László Bali · ppcout.com · 15 minute read
Ask ten store owners how they set their ROAS target and you'll hear the same answers: an industry benchmark from a blog post, a round number that "felt right," or whatever the previous agency left behind. Yet the target ROAS is arguably the single most consequential number in an e-commerce account — it's the instruction the bidding algorithm executes thousands of times a day, deciding which auctions to enter, how much to pay, and ultimately how big your business can grow.
A target set too low burns margin on unprofitable sales at scale. A target set too high quietly strangles the account: the algorithm retreats to only the safest auctions, volume collapses, and growth stalls while the dashboard shows a beautiful efficiency number. Both failures come from the same root cause — a target that was never derived from the store's actual economics.
This guide walks through deriving your break-even ROAS from real margins, choosing where on the growth–profitability curve to operate, why different products need different targets, and how to manage targets over time without constantly resetting the algorithm's learning. If you want a second opinion on whether your current targets match your economics, that's a core part of a professional Google Ads audit.
A ROAS target is not a performance goal you aspire to — it's an instruction the algorithm obeys. Set 400%, and smart bidding will enter exactly the auctions it predicts will average 400% return, and skip the rest. That's why a borrowed number is so dangerous: the algorithm executes it faithfully whether or not it has anything to do with your economics.
"A good ROAS is 4:1" is the most repeated and least useful sentence in e-commerce advertising. A 400% ROAS is comfortably profitable for a store with 50% gross margins and catastrophically loss-making for one running on 15%. Industry benchmarks average across wildly different margin structures, business models, and brand strengths — the number that's right for you can only come from your own unit economics. Any target you can't trace back to your margin is a guess wearing a percentage sign.
The deeper problem is that ROAS itself measures revenue against spend, and revenue isn't why you run the store. Two campaigns with identical 500% ROAS can have completely different profitability if one sells high-margin products and the other moves discounted stock. Everything in this guide treats ROAS as what it is: a proxy that only works when it's anchored to margin. The stores that get this right think in terms of profit per euro of spend, then translate that back into the ROAS language the platform speaks.
Wrong targets announce themselves, if you know the signs. A target set too low shows up as strong top-line growth with shrinking bank balance: revenue climbs, the dashboard celebrates, and the accountant asks where the money went. A target set too high shows the mirror image — pristine ROAS, collapsing impression share, spend that can't get out the door, and a "limited by budget" flag that never appears because the target itself is the brake. If your account shows either pattern, the fix isn't in the campaigns; it's in the one number they all obey.
Break-even ROAS is the return at which an ad-driven sale makes exactly zero profit — the floor below which every sale loses money. The formula is simple:
Break-even ROAS = 1 ÷ gross margin
A store with a 40% gross margin breaks even at 1 ÷ 0.40 = 250% ROAS. At 25% margin, break-even is 400%. At 60% margin, it's 167%. Notice how enormous the range is — this is exactly why borrowed benchmarks fail.
The formula is only as good as the margin you feed it. Gross margin here should be revenue minus cost of goods, payment fees, fulfillment and shipping costs you absorb, and — critically for some categories — returns. A fashion store with a 30% return rate has meaningfully thinner real margins than its product-level numbers suggest. Compute the margin on the same revenue definition your conversion tracking uses (net of VAT, consistently), or the whole calculation quietly shifts under you.
A single store-wide break-even is a fiction for most catalogs. Accessories at 70% margin break even at 143%; bulky electronics at 18% margin need 556%. Run the formula per category or margin band, and you'll usually discover that some products can never be profitably advertised at your current blended target — and others are being held to a standard far stricter than they need. This per-category table becomes the foundation for campaign structure later.
A worked example makes it concrete. A store sells a product for €100 net. Cost of goods is €55, payment and fulfillment absorb €7, and 10% of orders come back — call it €3 of expected return cost per order. Real gross margin: €35, or 35%. Break-even ROAS: 1 ÷ 0.35 ≈ 286%. If this store had copied the popular "4:1" benchmark and targeted 400%, it would be leaving profitable growth on the table; if it had targeted 250% because a competitor supposedly runs there, every ad-driven sale would lose money at scale. Ten minutes of arithmetic beats every benchmark article ever written.
Break-even is a floor, not a target. The target you actually set expresses a strategic choice, because ROAS and volume trade against each other along a curve: the higher the required return, the fewer auctions qualify, and the less you sell.
Picture the auctions available to your store, ranked from most to least efficient. A very high target restricts the algorithm to the top slice — great returns, tiny volume. Lowering the target admits progressively less efficient auctions: each step adds revenue at a worse marginal return. There's no "correct" point on this curve; there's the point that matches your goals. A cash-constrained store maximizing near-term profit sits high on the curve. A store prioritizing growth and market share deliberately operates lower, accepting thinner per-order profit for scale. The failure isn't choosing either — it's not knowing you chose.
The curve looks different once repeat purchases enter the math. If your average customer orders 2.5 times, a first order at break-even isn't zero profit — it's an acquisition that pays back across the relationship. Stores with strong retention can rationally set targets below single-order break-even for new-customer acquisition, funding growth their single-order-focused competitors can't match. The discipline required: actually know your repeat rate and payback window from data, and hold new-customer economics to that standard rather than using "LTV" as a slogan to excuse any loss.
A common confusion: treating budget as the growth lever and the target as a quality setting. In a target-ROAS world they're one system. With an achievable target and an open budget, the algorithm scales until the marginal auctions no longer meet the target — the target, not the budget, becomes the natural limiter. A capped budget on top of a healthy target means you're refusing profitable orders; an uncapped budget on a too-loose target means you're buying unprofitable ones at maximum speed. Decide the target from economics first, then give it the budget headroom to find its true volume.
A well-set target has a sentence behind it: "Our margin is 42%, so break-even is ~240%; we target 320% on the core catalog for profitable growth, and 240% on new-customer campaigns because repeat purchases pay back within five months." If you can't produce that sentence for your current target, the number is decorative — and the algorithm is executing a guess.
One blended target across a mixed catalog produces a predictable pathology: high-margin products effectively subsidize low-margin ones. The algorithm, told to hit 350% on average, happily overspends on cheap-to-convert low-margin items and underspends on the products that actually build your profit — and the blended number looks fine throughout.
The clean solution: group products by margin band using custom labels in the feed, split campaigns along those bands, and give each its own target derived from its own break-even. Your 60%-margin accessories might run at a relaxed 220% target and scale aggressively; your 18%-margin electronics need 600%+ or shouldn't be advertised cold at all. This is where feed work and target-setting meet — the labels in the feed are what make margin-based targets structurally possible.
A sale to a new customer and a sale to a returning one carry different strategic value, but revenue-based ROAS treats them identically. Where your data supports it, run acquisition-oriented campaigns with an LTV-adjusted (lower) target, and hold remarketing and returning-customer traffic to stricter single-order economics — it needed less convincing, so it should cost less. New-customer goals and value rules in Google Ads make this distinction operational rather than theoretical.
One caveat governs all this segmentation: data volume. Every split divides your conversions, and a campaign that drops below roughly 30–60 conversions a month gives smart bidding too little signal to hit any target reliably. Small stores should start with two or three margin bands at most, and only fragment further as volume earns it. A slightly blended target on a learnable campaign beats a perfect target on a starved one.
Searches for your own brand convert at rates cold traffic never will — those customers were already coming. Letting brand and non-brand share a target lets easy brand conversions mask expensive cold-traffic inefficiency, and gives the algorithm every incentive to lean on the easy wins. Separate them structurally, hold brand to a very high ROAS standard, and read your non-brand target as the true price of growth.
A target isn't a set-and-forget number. Margins shift, competition moves, seasons turn — and the right target moves with them. But how you change targets matters almost as much as what you change them to.
Large target jumps reset smart bidding's learning and cause erratic delivery. The working rule: adjust in steps of roughly 10–20%, then give the system one to two weeks of stable data before judging or adjusting again. If you need to travel a long way — say from 250% to 450% — plan it as a staircase over several weeks, not a single leap. Constant tinkering is the other failure mode: an algorithm that never gets a stable period never leaves the learning phase, and the frantic activity feels like diligence while actively hurting performance.
Peak periods change the math. During high-conversion seasons, the same target yields more volume naturally; deliberately lowering the target slightly during your strongest weeks can capture disproportionate scale while marginal returns remain acceptable. Conversely, in dead seasons a stubborn target forces the algorithm into an impossible brief. Use seasonality adjustments for short, sharp events (a weekend sale) rather than manual target swings, and always plan the return path — the weeks after a peak, when you gradually restore normal targets, are where much of the season's profit is kept or lost.
New campaigns deserve their own on-ramp as well. A fresh campaign with no conversion history can't hit a strict target — it has no data to predict with — so it either barely spends or flails. The standard path: launch on a value-maximizing strategy without a target (or with a deliberately loose one), let it accumulate a few weeks of conversions, then introduce the real target and tighten in steps. Demanding day-one efficiency from a day-one campaign just guarantees it never gets the data to deliver any.
At least quarterly, rerun the break-even math: have COGS, shipping costs, payment fees, or return rates moved? Has your repeat-purchase rate changed the LTV allowance? A target derived from last year's margins is a target derived from a business that no longer exists. The stores that treat this as routine maintenance keep their most consequential number honest — and their algorithm pointed at real profit.
Competitive shifts belong in the same review. When a major competitor enters your auctions or an incumbent retreats, CPCs move, and the volume available at your target moves with them. The target itself may still be economically correct — your margins didn't change — but the growth expectations attached to it must adjust. Reading impression share and auction insights alongside the quarterly margin check keeps both halves of the equation current: what a sale is worth to you, and what the market currently charges to get one.
If you'd like your current targets stress-tested against your actual unit economics — including whether your tracking values support them at all — that's exactly what a professional Google Ads audit covers. Or get a first read with a free Google Ads audit.
There is no universal good ROAS — the number depends entirely on your gross margin. A 400% ROAS is comfortably profitable at 50% margins and loss-making at 15%. The only meaningful reference point is your own break-even (1 ÷ gross margin); a "good" ROAS is one sufficiently above that floor to deliver the profit and growth balance you're aiming for.
Divide 1 by your gross margin. At 40% margin: 1 ÷ 0.40 = 250%. Use an honest margin — revenue minus cost of goods, payment fees, fulfillment costs, and expected returns — computed on the same revenue definition your conversion tracking uses. Then calculate it per category, because a mixed catalog never has one break-even.
Because the target is an instruction, not a wish. A higher required return disqualifies most auctions, so the algorithm retreats to a small slice of the safest traffic. That's the volume–efficiency trade-off working as designed. If you need to raise targets substantially, move in 10–20% steps with stabilization time between, and accept that a higher operating point structurally means less scale.
The critical rule is consistency: your conversion values and your target must use the same revenue definition. Most stores are best served with net revenue (excluding VAT) in both, since margin math is done on net figures. Gross values measured against a net-derived target make your economics look better than they are.
ROAS measures revenue per unit of spend; POAS (profit on ad spend) measures gross profit per unit of spend. POAS is the more honest metric for a mixed-margin catalog, because it stops high-revenue low-margin orders from looking like wins. You can approximate it within Google Ads by sending margin-adjusted conversion values or by splitting campaigns into margin bands with per-band targets.
Usually yes, if your retention data supports it. A first order that breaks even isn't zero profit when the average customer buys 2.5 times — it's paid acquisition with a payback period. Set acquisition campaigns to an LTV-adjusted lower target and hold returning-customer and remarketing traffic to stricter single-order economics. The discipline: derive the allowance from measured repeat rates, not optimism.
Change it when your economics or strategy change, not on a schedule — but revisit the underlying math quarterly. When you do adjust, move in 10–20% increments and give the algorithm one to two weeks of stability before the next move. For short events like sale weekends, use seasonality adjustments rather than manual target swings.
Absolutely — and it's the quieter failure. An excessive target strangles delivery: the algorithm serves only the safest auctions, volume and growth stall, and the account shows beautiful efficiency on shrinking scale. If spend can't get out the door and impression share is collapsing while ROAS looks stellar, your target is functioning as a brake, not a goal.
Online marketing and PPC specialist focused on Google Ads and advertising strategy — the kind that builds not just clicks, but brands. With 10+ years in digital marketing and e-commerce, Tom leads on strategy and measurement, turning strategic scaling and zero-click trends into measurable business results.
in
Performance marketing specialist with deep hands-on Google Ads and e-commerce experience. László leads on campaign execution and growth, building and scaling accounts for e-commerce brands and small businesses — the same senior specialist on your account from day one, not a junior and a dashboard.
in