Target CPA vs Maximize Conversions: Which to Run and When to Switch
They are the same bidder with one extra constraint. Once you see that, the switch rule becomes obvious.
The short version
- Target CPA is Maximize Conversions with a price ceiling, not a different algorithm.
- Stay on Maximize Conversions until you have roughly 30 conversions in 30 days and a stable conversion rate.
- Set your first CPA target at your current actual CPA, then tighten 10 to 15 percent at a time with a two-week read between steps.
- Clean your search terms before blaming the bid strategy: bad queries poison the conversion-rate model that both strategies depend on.
- Always read CPA and conversion volume together; a great CPA with collapsed volume is a loss.
Run plain Maximize Conversions when the campaign is new or thin on data, and switch to Target CPA once you have a stable conversion rate and a CPA you actually want to defend (roughly 30+ conversions in the last 30 days is a reasonable floor). The two are not different algorithms. They are the same conversion-predicting bidder, and Target CPA just adds a price ceiling. That single fact tells you when each one helps and when it quietly wastes money.
They share one bidder. Target CPA just adds a ceiling
Both strategies use the same underlying machinery. For every auction, Smart Bidding predicts the probability that this specific click converts, given the signals it can see: query, device, time, location, audience, and the rest. It then sets a bid proportional to that predicted value. Higher predicted conversion rate means a higher bid.
Maximize Conversions takes that prediction and spends your full budget as aggressively as needed to win the most total conversions. It has no opinion about cost per conversion. If your budget is 200 dollars a day, it will find the combination of bids that drains that 200 dollars for the maximum conversion count, even if that means paying 90 dollars for the marginal conversion.
Target CPA is the exact same bidder with one added rule: do not bid so high that the expected cost per conversion exceeds the target. It pulls back on auctions where the predicted conversion rate is too low to hit your number, and leans in where the math works. That is the whole difference. Once you internalize that Target CPA is Maximize Conversions plus a constraint, every behavior below becomes predictable.
When plain Maximize Conversions is the right call
Use Maximize Conversions when the bidder does not yet have enough data to estimate conversion rate reliably. A new campaign, a new landing page, or a campaign that just changed its conversion action is flying with a blurry model. If you clamp a CPA target on top of a bad prediction, you get bad throttling: the system pulls back on auctions it wrongly thinks are expensive and misses volume you needed to learn from.
Maximize Conversions also wins when budget, not CPA, is your real constraint. If you have 500 dollars a day and you genuinely want every conversion that money can buy, a target only gets in the way. This is common for lead-gen accounts where sales can absorb whatever volume comes in, and for promotions where the goal is to spend the allotted budget hard for two weeks.
The honest limit here: Maximize Conversions will spend your full budget every single day. If your daily budget is set higher than you actually want to spend, this strategy will find a way to use it. Your budget line is your only cost control, so set it deliberately.
When Target CPA pays off, and the numbers to hold before flipping
Switch to Target CPA when two things are true: the campaign has a stable conversion history, and cost efficiency matters more than raw volume. Stability is the key word. The ceiling constraint is only as good as the conversion-rate prediction it constrains, so you want the model well-fed before you ask it to hedge on price.
A workable floor: at least 30 conversions in the trailing 30 days, and ideally a conversion rate that has not swung wildly week to week. Below that, the target does more throttling harm than efficiency good. When you do turn it on, set the initial target near your recent actual CPA from Maximize Conversions, not your dream number. If you have been paying 45 dollars per conversion and you set the target to 20, you have not asked the bidder to be efficient. You have asked it to stop bidding on most of your traffic, and volume falls off a cliff.
Tighten in steps, not leaps. Move the target down by 10 to 15 percent, wait roughly two weeks or one full conversion cycle (whichever is longer) for the change to settle, then read the result. If CPA held and volume is acceptable, tighten again. If volume collapsed, back off. The reason for the wait is mechanical: Smart Bidding re-enters a short learning period after a target change, and reading results mid-adjustment tells you nothing.
The failure mode nobody warns you about
The trap is switching to Target CPA to cut costs on a campaign whose costs are high because of bad queries, not bad bidding. If your search terms are full of irrelevant or low-intent matches, tightening the target does not fix that. It just makes the bidder pull back everywhere, including on the good traffic, because the model's average predicted conversion rate is dragged down by the junk it is being forced to bid on.
This is why query hygiene changes bidding outcomes directly. Every irrelevant search term you let through is a data point that muddies the conversion-rate model. Clean search terms produce sharper predictions, and sharper predictions let Target CPA discriminate between the auction that will convert at 25 dollars and the one that will cost 200. Go to Campaigns > Insights and reports > Search terms, add negatives for the terms that are clearly off-intent, and let the model relearn before you blame the bid strategy. Tools like QueryCut exist to make that review faster, but the mechanism is what matters: fix inputs before you constrain outputs.
A second, quieter failure: Target CPA set too tight will report a great CPA and almost no volume, and a busy manager will call that a win. It is not. You bought a handful of the easiest conversions and left the rest of the market to competitors. Always read CPA and conversion volume together. A target that halves your cost per conversion while cutting total conversions by 70 percent is usually a loss, not a win.
A concrete decision path for this afternoon
Start by pulling your last 30 days of conversions for the campaign. If it is under 30, stay on Maximize Conversions and do not touch the bid strategy. Your job right now is data collection and query cleanup, not efficiency tuning.
If you are over that floor and your cost per conversion is stable but higher than you want, switch to Target CPA and set the target at or just below your current actual CPA. Then tighten in 10 to 15 percent steps with a two-week read between each. Stop tightening the moment volume drops faster than you can accept, and record that number as your practical floor for the campaign.
If your CPA is already fine and you simply want more of it, stay on Maximize Conversions and raise the budget instead. Adding a target to a campaign that is already efficient just introduces a constraint that can only cost you volume. The strategy you pick should follow your actual constraint: budget-bound campaigns want Maximize Conversions, efficiency-bound campaigns with real data want Target CPA.
Questions people also ask
How many conversions do I really need before turning on Target CPA?
A practical floor is 30 conversions in the trailing 30 days, with a conversion rate that has not swung wildly week to week. Below that, the model's conversion-rate predictions are too noisy for the target constraint to help, and it will throttle traffic based on bad estimates. The number is a heuristic, not a hard rule: if your conversions are highly consistent you can start a little lower, and if they are erratic you may want more.
Can I set Target CPA lower than my current cost per conversion to force efficiency?
Not in one move. If you set the target well below your recent actual CPA, the bidder responds by pulling back on most auctions, and volume drops sharply. Set the target near your current actual CPA first, then tighten in 10 to 15 percent steps with a two-week read between each. That lets you find the real efficiency floor without killing the campaign.
Why did my volume crash after switching to Target CPA?
Almost always because the target was set below what the campaign can actually achieve, so the bidder stopped competing in auctions it predicts will exceed the ceiling. Raise the target back toward your prior actual CPA and volume should recover. If volume was already borderline, also check your search terms, since low-intent queries drag down the average predicted conversion rate and make the target bite harder than expected.
Does Maximize Conversions always spend my full budget?
Yes, that is its design. It sets bids to win the most conversions your budget can buy, with no cost ceiling, so your daily budget is your only real spending control. If you do not want to spend the full amount, either lower the budget or use Target CPA to introduce a cost constraint.
Should I fix search terms before or after choosing a bid strategy?
Before, and continuously. Both strategies depend on the same conversion-rate prediction, and irrelevant search terms feed that model bad data, which degrades every bid it makes. Review Campaigns > Insights and reports > Search terms and add negatives for off-intent matches, then let the model relearn before judging whether a bid strategy is working.