The Number You Set Once and Forgot About
Most Google Ads accounts run on Smart Bidding now. You told Google to Maximize Conversions, or you handed it a Target CPA or a Target ROAS, and the algorithm has been steering the account ever since. For a lot of owners that target is a number they typed in months ago and never touched again.
That number does more than you think. It is not a note in the margin. It is the single most important instruction you give the machine, and Google has been getting more literal about following it. If you have never gone back and asked whether your target is set to the right number, this is the section that saves you money.
Because here is the shift most small advertisers never notice: the target you set is quietly moving from a ceiling you beat into a promise Google keeps. And when it does, "doing nothing" is a decision with a cost.
What Smart Bidding Is Actually Doing
Before the fix, the plumbing. Smart Bidding is Google setting your individual keyword bids in every auction, automatically, aimed at a goal you choose. There are four goals worth knowing in plain English:
- Maximize Conversions. Get the most conversions possible for the budget. No cost target. Google spends the budget to buy as many conversions as it can find.
- Maximize Conversion Value. Same idea, but chasing revenue instead of raw count. Best when a $900 sale and a $30 sale are not worth the same to you.
- Target CPA (cost per acquisition). Get conversions at roughly the average cost you name. Set $40 and Google tries to bring you leads around $40 each.
- Target ROAS (return on ad spend). Chase a revenue-to-spend ratio. Set 400% and Google aims to make $4 back for every $1 spent.
The first two tell Google "get me as much as you can." The second two hand Google a specific number to hit. That specific number is where the trouble, and the opportunity, lives.
Your Target Is an Instruction, Not a Ceiling
Here is the mechanic almost nobody has clocked. When a campaign is limited by budget and it also has a Target CPA or Target ROAS, the algorithm used to play it safe. It knew it could only spend so much, so it cherry-picked the cheapest conversions it could find inside that budget. The result felt like a gift: you set a Target ROAS of 340% and the campaign quietly ran at 430%. You set a $50 target CPA and got leads at $32.
Owners saw that and thought they were winning. In a way they were. But that gap was never something you asked for. It was a side effect of the algorithm being cautious inside a tight budget, and it was always fragile.
Google has moved to close that gap. Now the target you state is the target it aims for, consistently, not a ceiling you drift below. Set 340% and that is what it steers toward. Which means if your budget-limited campaign has been over-performing its target, and you change nothing, that extra performance does not stay. It settles down to exactly the number you asked for.
Read that twice, because it is the whole article. The number you type in is no longer the worst case. It is the plan.
If You've Been Quietly Beating Your Target, Read This
This only touches campaigns in one specific situation: limited by budget AND running a Target CPA or Target ROAS AND currently out-performing that target. If a campaign is not capped by budget, nothing here changes for it. If it has no target at all, same.
But if you are in that spot, the "bonus" performance you have enjoyed is on the clock. Left alone, a campaign running at 430% against a 340% target will trend back toward 340%. To hit that lower target, the algorithm enters more expensive auctions it used to skip, so your cost per conversion rises and your conversion volume falls. You end up paying more for less, and your reports will still say you "hit your goal." Nothing looks broken. It just quietly costs you.
There is a competitive wrinkle too. Imagine ten local competitors bidding on the same keyword, all budget-limited, all comfortably beating their targets. As targets become literal, the advertiser willing to raise their target, to genuinely pay more per lead, starts winning auctions the others used to share. The passive advertisers feel it as rising costs and thinner volume. Your bid target is now a real competitive lever, not a formality. That is exactly why the owner who understands this beats the one who set a number a year ago and looked away.
Smart Bidding Isn't a Wishing Well
Now the guardrail, so nobody reads this and does something reckless.
Setting an ambitious target does not summon ambitious results. Smart Bidding needs two things to hit a target reliably.
First, data. Google's own guidance is to judge these strategies over a window with at least 30 conversions in the last 30 days, and many practitioners want to see 30 to 50-plus a month before they lean hard on a target. Thin conversion data makes the algorithm guess, and a guessing algorithm is a volatile one. If you are not there yet, Maximize Conversions while you build volume is often the saner move.
Second, a target grounded in reality. Your target should track what the campaign has actually done over the last 30 to 60 days, not the number you wish were true. Type in a dream ROAS the account has never produced and you do not get the dream. You get the algorithm strangling your spend trying to chase a ceiling that was never there, and your volume collapses.
One more honest note. The ROAS and CPA that Google reports are not the same as the profit in your bank account. Attribution windows, conversion counting, and the gap between a tracked conversion and a paying customer all mean the platform number is an optimization signal, not your real return. Set your targets off numbers you trust, and treat the dashboard as a steering tool, not a P&L.
The Three Levers
So you found a budget-limited campaign beating its target. You have exactly three moves, and the right one depends on what you actually want.
Lever 1: Let it run free. If the campaign is reliably beating your target and you are happy with that performance, the cleanest move is often to remove the target and switch to Maximize Conversions or Maximize Conversion Value. You are already exceeding the goal, so let the algorithm chase volume instead of throttling itself to hit a number. Alternatively, keep the target and raise the budget, which turns that efficiency into more conversions rather than a nicer-looking ratio.
Lever 2: Lock what you have. If you want to protect exactly the performance you are getting right now, set your target to your real recent actual. If the campaign is running at a $30 CPA against a $50 target, move the target to $30. Now the number you stated and the number you are hitting agree, and there is no gap for the algorithm to give back.
Lever 3: Reach for more, carefully. If your real goal is genuinely higher than the campaign's current performance, do not just type it in. Which brings us to the last move.
Climb the Target, Don't Leap It
The single most common way owners hurt themselves here is jumping the target in one move. They see a campaign running at a 200% ROAS, decide they want 350%, and set 350% today. The algorithm, told to nearly double its return overnight, pulls way back on spend, enters far fewer auctions, and volume falls off a cliff.
Targets respond to steps, not leaps, the same way budgets do. If a campaign is running at 200% and you want 350%, nudge the target to about 250%, give it a week or two of stable data, then step to 300%, then 350%. Each step lets the algorithm re-learn on solid footing instead of panicking. It is slower, and it is the difference between climbing to a better number and blowing up a working campaign to chase one.
Here is the whole thing on one page. Open your account and look at every campaign that is limited by budget and running a Target CPA or Target ROAS. For each one, compare the target you set against what it is actually doing. Where they match, you are fine. Where the campaign is beating its target, pick a lever on purpose: let it run free, lock it in, or ladder it up. The mistake is not any one of those choices. The mistake is leaving a stale number in the box and letting the machine quietly decide for you.
None of this is complicated once you see it, but it is the kind of thing that hides in plain sight while an account slowly leaks performance. If you would rather have someone watching your targets so they never drift, that is what we do all day. Either way, go check that number. It is steering more than you think.
