A realistic target CPA or ROAS comes from your own margin and average order value, not an industry benchmark: work out what you can afford to pay for a sale or lead once profit is protected, then set Smart Bidding to that number — not one copied from a blog post.
Book a free strategy callWhen an account isn’t spending its budget, or conversion volume has quietly dried up, the first place I look isn’t the keywords or the ad copy — it’s the target CPA or target ROAS sitting in the bidding settings. More often than not, that number was set once, months or years ago, copied from an industry-average figure someone found online, and never touched again.
Smart Bidding doesn’t treat your target as a rough guide. It treats it as an instruction to only bid on auctions it predicts will land at or better than that number. Set the target too tight — a target CPA that’s lower than what it actually costs to win the auction, or a target ROAS that’s higher than your account genuinely converts at — and Google’s algorithm won’t fight to hit it. It will simply bid less, less often, and your traffic and conversions taper off, sometimes with no obvious cause in the interface.
“What’s a good CPA for [industry]?” is the wrong question, because the honest answer is that it depends entirely on what your business can afford to pay — and that’s not something a benchmark table can tell you. Two businesses selling the same product at the same price can have completely different affordable CPAs, because one has 60% margin and the other has 20%, or one has a customer who reorders five times a year and the other has a one-off transaction.
A target CPA or target ROAS is really just your business economics translated into a number Google’s bidding algorithm can act on. So the starting point isn’t Google Ads at all — it’s your own numbers: what a sale is actually worth to you, and how much of that you can hand over in ad spend before the sale stops being worth winning.
An arbitrary target — one pulled from an industry benchmark, a competitor’s rumoured numbers, or just what “feels” like a good CPA — doesn’t fail loudly. It fails quietly. Smart Bidding starts declining to bid on auctions it predicts won’t hit that number, volume drifts down, and because there’s no error message or obvious red flag, it’s easy to read the drop as “the market’s gone quiet” rather than “the target is throttling the campaign.” I’ve seen accounts where loosening an unrealistically tight target — because the real economics supported it — brought volume back within days, with the CPA still landing at a profitable level.
A target CPA or ROAS only works if the budget behind it is big enough for Smart Bidding to actually learn. If your daily budget caps out before the algorithm has explored enough auctions, you’re not really testing whether your target is achievable — you’re testing whether your budget is big enough to find out. The two settings work together: the target tells Google what a good outcome looks like, and the budget determines how much room it has to go and find those outcomes.
Google has confirmed that from 17 August 2026, budget-constrained target CPA and target ROAS campaigns will stop over-performing their stated target and instead track much more closely to it — so any mismatch between target and budget that used to get masked by better-than-target results will start showing up directly in performance. Worth checking your budget-limited campaigns against their targets before that date lands.
If you’re not sure whether a target is unrealistic or the budget is simply too constrained to prove it either way, that’s usually the first thing worth separating out before touching the target itself.
Margins change, average order value shifts with product mix, and lifetime value assumptions get more reliable the longer you’re trading. A target CPA or ROAS set a year ago on last year’s numbers is working off out-of-date economics even if nothing about the account itself looks broken. Treat it as a figure to revisit periodically alongside your actual margin and repeat-purchase data, not a setting you configure once and leave alone.
There isn’t one that applies across businesses, even within the same industry, because it depends on your margin, average order value and repeat purchase rate — not the product category. Two competitors selling the same thing can have very different affordable CPAs depending on their own economics. Work it out from your numbers rather than an industry average.
A tighter target CPA tells Smart Bidding to only chase cheaper conversions, so it becomes more selective about which auctions it enters. If the new target is below what it actually costs to win a reasonable share of your auctions, volume falls because the algorithm is declining to bid rather than bidding and losing.
Target CPA suits businesses where most conversions have a similar value — leads, or products at a fairly consistent price. Target ROAS suits accounts with a wide spread of order values, like ecommerce stores selling both low- and high-ticket items, because it lets the algorithm bid more for higher-value baskets rather than treating every conversion the same.
Even a rough, honest estimate is better than ignoring it. If you know a meaningful proportion of customers reorder, even a conservative uplift on first-purchase value gives you a more realistic target than calculating purely off the first transaction, which tends to under-value the acquisition and set the target too tight.
Yes. If the daily budget limits how many auctions Smart Bidding can enter, it may never get the chance to prove a target is achievable. Before concluding a target itself is wrong, it’s worth checking whether the budget is capping the campaign’s ability to explore enough auctions to hit it.
Whenever the underlying economics change — a margin shift, a change in average order value, new data on repeat purchase rate — rather than on a fixed schedule. A target set on old numbers can quietly throttle or overspend the account even when nothing else about the setup has changed.
I've seen plenty of accounts where the target CPA or ROAS was set once, years ago, and never checked against what the business could actually afford. Book a free strategy call and I'll give you an honest read on whether your targets match your real numbers.
Call 07410 907 104 Book a free strategy call