What is Maximise conversions?
Maximise conversions is a Google Ads bid strategy that spends the whole budget pursuing as many conversions as possible. Without a target attached it has no cost ceiling, so the budget becomes the only constraint and cost per conversion is whatever the auction produced.
The instruction, read literally
Maximise conversions tells the campaign to produce as many conversions as it can from the budget it has been given. Read literally — which is how a bidder reads everything — two things follow immediately.
The first is that the budget will be spent. The strategy has no reason to leave money unused, because unused money is conversions not bought. Campaigns moved onto it from a target-based strategy routinely start spending at the full daily budget for the first time, and accounts that were quietly under-delivering discover it abruptly.
The second is that there is no cost ceiling unless one is attached. A conversion at a high price is still a conversion, and the strategy will buy it in preference to buying nothing. The resulting cost per conversion is an output, not a setting.
Neither of those is a flaw. They are the strategy doing what it was asked. The mistakes come from expecting a ceiling that was never set.
When the shape is right
The strategy fits a specific situation: the budget is the real constraint, and any conversion the budget can buy is worth having.
That describes more campaigns than people expect. A business with a large gap between what a conversion costs and what it is worth is not price-constrained; it is constrained by how much it is willing to spend in total. Asking such a campaign to hit a cost target below what the auction charges leaves budget unspent and volume on the table, for the sake of an average that was never the binding number.
It also fits a campaign being deliberately scaled, where the immediate question is whether the volume exists at all before the question of what it costs.
And it fits a campaign with conversion volume too thin to support a reliable target but sufficient for the model to work with, where imposing a target adds a constraint the data cannot really justify.
When it is expensive
Three situations.
Conversions are worth different amounts. Counting them equally means the strategy prefers whichever is cheapest to produce. In an account where a trial signup and an enterprise enquiry both count as one, it will find trials. Where value varies, the value-based sibling is the right tool, and the fix is not a target but a different objective.
There is a real ceiling. A business that genuinely cannot pay above a figure should say so with a target, and accept that delivery becomes the variable. Running unbounded and hoping the average lands well is a way of discovering the ceiling after the money has gone.
The budget is larger than the opportunity. Told to spend a budget that exceeds what the audience can absorb at sensible prices, the strategy reaches further down the quality curve to do it. The symptom is a cost per conversion climbing steadily while volume grows slowly, and the remedy is a smaller budget or a wider audience rather than a bidding change.
Attaching a target changes what it is
A target cost per acquisition attached to this strategy converts it from volume-seeking to cost-bounded. The bidder now declines auctions it does not expect to clear the target, which means it may not spend the budget at all.
That is the whole trade, and it is worth stating plainly because the two configurations behave so differently that treating them as one strategy with an optional extra causes real confusion.
Unbounded: budget is the constraint, cost is the output, delivery is predictable and expensive.
Targeted: cost is the constraint, delivery is the output, spend is unpredictable and the average is held.
A business that knows its ceiling should use the second. A business whose ceiling is far above what the market charges should use the first and stop worrying about an average that is not binding on anything.
Running it
Check the conversion first. A count-maximising strategy is maximally sensitive to a conversion definition that counts the wrong thing, because volume is the entire objective. A duplicated tag or a shallow event will be exploited immediately.
Set the budget as the real limit. Under this strategy the budget is not a safety net, it is the instruction. Set it at what the business is prepared to spend, not at a number chosen to leave headroom.
Watch cost per conversion as an outcome, not a target. Track it against the ceiling that contribution and close rate imply. The moment the marginal conversion crosses that ceiling, the campaign is buying volume the business cannot afford, and that is a signal to attach a target or to reduce the budget.
Give it room after a change. Like every conversion-based strategy, it rebuilds its predictions after a material edit and delivers unevenly while it does.
Revisit as volume grows. A campaign that started unbounded because the data was thin can usually support a target once it has a few months of history, and the target is what turns a growing campaign into a controlled one.
A common misdiagnosis
An account switches to this strategy, spend jumps, cost per conversion rises, and the conclusion drawn is that the automation is out of control.
Usually nothing is out of control. The previous configuration was leaving budget unspent, the new one is using it, and the extra conversions bought at the margin cost more than the average did — exactly as they must, because the cheap ones were already being bought.
The right question is not why the average moved but whether the marginal conversion is worth its price. That is answerable: take the contribution a customer produces, multiply by the rate at which conversions become customers, and compare. If the marginal conversion is under that ceiling, the campaign is buying profitable volume and the rising average is irrelevant. If it is over, the campaign needs a target rather than a smaller budget, because a smaller budget cuts the good conversions along with the bad.
The distinction matters because the two remedies point in opposite directions, and the report showing a rising average supports neither on its own.
Relation to the value-based sibling
Where conversions are worth different amounts, the count-based strategy is the wrong instrument however it is configured, and no target fixes it.
A target constrains what a conversion may cost. It says nothing about which conversions to prefer, so the strategy continues to prefer the cheapest ones, now within a ceiling. An account selling both a A$40 item and a A$400 item under a single cost target will fill its budget with the cheaper one and hit the target comfortably.
The fix is to change what is being maximised rather than to bound it, which means sending a value with each conversion and moving to the value-based strategy.