Your Target ROAS Is a Math Problem, Not a Google Setting
Someone typed a number into your Target ROAS field. Ask where it came from and you get one of two answers: Google recommended it, or the finance person wants a five-times return. Neither answer involves your margin after returns, and neither involves what your auction can physically produce at your current click cost. Your target is a number you compute from your own unit economics and then test against what your account can actually deliver. Until you do both, tuning campaigns is decoration.
ROAS means revenue per dollar of ad spend, so 300% ROAS is three dollars back for every one you put in. CPA means what you pay to get one new customer or lead. Break-even is the point where you neither make nor lose money on that spend.
The number in that box is already spending your money
Google's own help documentation on Target CPA bidding explains the mechanics well: you set an average cost you are willing to pay per conversion, and the system bids to hit that average across the campaign. It is accurate and it is complete on mechanics. It says nothing about where your number should come from, which is the only part you cannot outsource.
Set the target too low and the system happily spends, buying you conversions at a price that loses money on every sale. You scale, revenue climbs, and the bank balance drops. Set it too high and the opposite happens quietly: Google bids down, impressions thin out, and the campaign starves itself while you sit in a meeting asking why volume collapsed. Both failures look like a platform problem. Both are arithmetic problems that were decided before the campaign launched.
The first calculation takes thirty seconds. Break-even ROAS is one divided by your profit margin. A 40% margin breaks even at 250%, meaning every dollar of ad spend has to return two dollars fifty before you have made a cent. Wijnand Meijer, a paid search practitioner since 2006 and co-founder of TrueClicks, laid this out in Search Engine Land on July 22, 2026, and it is the cleanest version of the math I have seen published.
| Your effective profit margin | Break-even ROAS | What that means at the till |
|---|---|---|
| 20% | 500% | Five dollars back per dollar spent just to stand still. |
| 30% | 333% | Anything under 333% is spending to lose money faster. |
| 40% | 250% | The classic retail number, and the one most often quoted from gross margin instead of real margin. |
| 50% | 200% | Two dollars back per dollar, and every point above it is profit. |
Lead generation runs on the same logic with different inputs. Break-even CPA is profit per customer multiplied by your close rate. If a closed client is worth $1,000 in profit and your sales team closes 20% of leads, your break-even CPA is $200, because you buy five leads to get one customer. A clinic paying $260 a lead with a 20% close rate is losing $300 on every new patient and will not see it in the ads dashboard, which reports leads, not patients.
Your margin is lower than the number you just used
Most owners reach for gross margin, because it is the number on the P&L summary and the one they can recall in a meeting. It is the wrong input. Meijer's example is a retailer running a 40% gross margin with a 25% return rate, which lands the business near 30% once returns and fees come out. That single correction moves break-even ROAS from 250% to 333%.
Payment processing, shipping subsidies, the discount code that half your buyers apply at checkout, the refunds your customer service team approves without telling marketing. Strip all of it out before you divide. Take last quarter's revenue from the channel, subtract cost of goods plus every one of those deductions, and divide what is left by the revenue. That percentage is the margin you divide into. The gap between headline margin and effective margin is where most "profitable" accounts quietly live, and it is invisible from inside Google Ads because Google reports revenue, not what you keep.
Timing matters here too, because this number is about to start doing more work. Google's change on August 17 makes the number you typed and forgot matter more, not less, and the August 17 target bidding change explained covers what changes and when. This one covers what the number should be.
"Set it at last month's average" is a circle, not a calculation
The most widely repeated advice on this is Storegrowers', and it is stated plainly: "Set your initial target CPA at or slightly above your average CPA from the last 30 days." It is safe, it avoids sudden volume shocks, and it is completely circular. Last month's average CPA is the result of last month's bids. Using it to set this month's target means you have anchored your profitability to an accident.
Ask what that advice would tell a business that spent last month acquiring customers at a loss. It would tell them to keep doing it, slightly more expensively. The number never touches margin and never asks whether the auction could deliver something better.
The second wrong path is the handed-down target. Finance says five times return, the number arrives in an email, and the account manager types it in. In ten years running Google Ads, across a client base of 300-plus businesses in the USA, Canada, the UK, Singapore, Australia and New Zealand, the most common version I see is exactly this: an inherited target nobody can source, defended by nobody, governing real money. When I ask who calculated it, the room goes quiet. That silence is worth more than most audits, because it tells you the number was never a decision.
Two numbers decide it: what keeps you profitable, and what your auction can hit
Break-even tells you the floor. It does not tell you your target, because breaking even on customer acquisition is not a business model. The target comes from deciding what share of the customer's value you are willing to spend to acquire them. Target ROAS is one divided by margin multiplied by acquisition share. Multiply your margin by the share first: 0.40 times 0.50 is 0.20. Then divide one by that: a 500% target. Spend 70% of the profit instead and the math gives 1 divided by 0.28, a 357% target, which buys considerably more volume at lower profit per sale.
Meijer cites George Michie's square root rule, which puts the profit-maximizing acquisition share for most accounts between 50% and 70%. Below that band you are leaving growth on the table. Spending past that share buys revenue that costs more than it returns.
A target is a request, and the auction has to be able to fill it. Achievable ROAS is your conversion rate multiplied by average order value, divided by your cost per click, with conversion rate and click cost taken from the campaign's own reports and order value from your store or CRM. At a 2% conversion rate, a $120 average order and an $0.80 CPC, every click returns $2.40 of revenue against $0.80 of cost, which is 300%. That is the ceiling. Not a guideline, a ceiling, set by three numbers that live in your account and your checkout.
When the two numbers disagree, the target moves, not the auction
Picture a service business whose accountant has decided the ads should return five times what they cost. Reasonable request, stated with authority, and nobody in the room can dispute it because nobody has done the second calculation. Run it and the account tops out near 300%. The target goes in at 500% anyway, Google bids down to chase an average it cannot reach, impressions fall, lead flow halves, and by week six the owner is telling people that Google Ads stopped working for their industry. Google did exactly what it was told. It was told something impossible.
When target and achievable disagree, only one of them is negotiable, and it is not the auction.
| The situation | What Google quietly does | The only fix that works |
|---|---|---|
| Target 500%, achievable 300% | Bids down, volume starves, the campaign looks broken. | Move conversion rate, order value or click cost. Typing a bigger number changes nothing. |
| Target 200% on a 30% effective margin | Spends freely and hits the target, because the target is easy. | Raise the target to at least 333% before adding a single dollar of budget. |
| Target sitting exactly at break-even | Buys customers all day at zero profit per sale. | Decide your acquisition share, 50% to 70% for most accounts, then set the target above break-even. |
Closing a gap means moving one of three levers, and they are not equally easy. Average order value moves fastest, through bundles, minimum order thresholds or a better upsell at checkout, and it costs no media budget. Conversion rate is a landing page and offer problem, and it compounds with everything else you run. Click cost is the hardest, because auction prices are set by competitors bidding for the same people, and no amount of smarter bidding fixes structurally rising click costs on its own. If none of the three will move, the honest conclusion is that this channel cannot hit the target the business wants, which is a real answer and a far cheaper one than six months of blame.
Getting from your current wrong target to the derived one is not a single edit. The way I move accounts is in steps of 10% to 15% every week or two, holding each step long enough to read real results before taking the next. Jump the full distance in one go and you shock volume, the system starts re-learning from scratch, and you lose the weeks you were trying to save.
If an agency runs your account, you do not need to do this yourself. You need to ask one question and listen to the shape of the answer: what effective margin and what achievable ROAS did you use to derive our target, and can I see both calculations? A good partner sends a spreadsheet within a day. A weak one sends a paragraph about optimization and machine learning.
Judge the number on profit, and give it time to be wrong first
A new campaign shows you its worst possible numbers. Quality Scores have not settled, the system has no conversion history to learn from, and click costs sit at their peak. Andrew Goodman of Page Zero Media, writing in Search Engine Land on July 10, 2026, reported on one real account that he "recently saw CPCs drop by 80% between establishing Quality Scores and our optimizations." Set a permanent target off week-one data and you are locking in the most expensive week the account will ever have.
That settling period is also why the popular big-launch strategy backfires so reliably. Pouring budget in on day one buys the maximum number of clicks at the maximum price with the minimum amount of learning, and I have watched it quietly wreck otherwise healthy accounts. A big opening spend burns cash before the account knows anything, and it also poisons the very data you would use to set a sensible target.
Once you are past the settling period, three things are worth watching monthly. Profit per campaign, calculated outside Google using your effective margin, because Google reports revenue and revenue is not yours. Delivery, meaning whether the campaign is spending its budget, since a campaign chronically underspending is telling you the target is above achievable. And your three achievable-ROAS inputs, because when conversion rate or order value improves, your ceiling rises and your target should rise with it. Most accounts never revisit the target after a good quarter, which means they leave the profit sitting in extra volume they did not need.
The metric to distrust is the ROAS number on the dashboard itself. It is only as honest as the conversion values feeding it, and if your values are static placeholders or count the same order twice across overlapping conversion actions, the entire calculation you just did is running on fiction. Fix the inputs before you trust the output, and how to fix broken conversion tracking values covers the checks I run first.
Frequently Asked Questions
What should my target ROAS be if I have no idea where to start?
Start with break-even, which is one divided by your profit margin after returns, refunds, shipping and payment fees. A 30% effective margin means you break even at 333%, so anything below that loses money on every sale. Then decide what share of your profit you are willing to spend to acquire a customer, between 50% and 70%, and divide again. Before you commit, check what your account can actually hit: conversion rate times average order value, divided by average cost per click.
Why did my campaign stop spending after I raised my target ROAS?
Because you asked for a return the auction cannot produce at your current numbers, so the system bids lower and enters fewer auctions rather than buying clicks it does not expect to pay off. That is the campaign obeying you, not failing. Work out your achievable ROAS from conversion rate, order value and click cost, and if your target sits above it, either lower the target or move one of those three inputs. Raising the number again will only starve the campaign further.
Should I use Target CPA or Target ROAS?
Use Target CPA when every customer is worth roughly the same to you, which is most lead generation and service businesses, and derive it from profit per customer multiplied by your close rate. Use Target ROAS when order values vary a lot, which is most ecommerce, so the system can chase the bigger baskets rather than the cheapest conversions. The choice of setting matters far less than where the number came from. A well-derived Target CPA beats a guessed Target ROAS every time.
The number in that field decides whether growth makes you richer or just busier. It deserves a spreadsheet and an afternoon, not a suggestion accepted at 4pm on a Friday. If you want a second pair of eyes on your margin math and what your auction can realistically deliver, book a call and bring your last 90 days of numbers. Most owners I speak with can name their conversion rate to one decimal place and have never once calculated the ROAS their own account is capable of producing.
