What is Bid cap?
A bid cap tells the platform never to bid above a stated figure in any single auction. Meta and TikTok offer it under that name; Google Ads reaches it through a maximum cost-per-click. It binds every auction rather than an average.
The strictest of the controls
A bid cap tells the platform that it may never bid above a stated figure in any individual auction. Not on average, not across the campaign — in each auction, one at a time. Meta and TikTok both offer it under that name; on Google Ads the equivalent is the maximum cost-per-click on a manual or portfolio strategy.
That makes it the hardest of the bidding controls, and the difference from the others is worth stating precisely because the three are frequently discussed as though they were interchangeable settings of different strengths.
A bid cap binds every auction. A cost cap binds an average, leaving the bidder free to pay more here and less there. A spend cap binds a total, and says nothing about the price of anything.
They are three different promises, and the first one is the only one that is true of every single transaction.
What it protects and what it costs
The protection is real: with a bid cap in place, no single auction can cost more than the number. For a business with a genuine hard ceiling — a contractual maximum, a commercial line it will not cross, a deliberate decision to buy only where the market is cheap — that is exactly the right instrument, and an averaged control does not provide it.
The cost is delivery. Auctions clear where they clear, and a cap below the clearing price means the campaign declines them. It does not negotiate; it simply does not participate.
The resulting pattern is distinctive and widely misread. Spend is low. Cost per result looks excellent. Volume is a fraction of what the budget could support. An account reading only the cost column concludes the campaign is efficient, when what has actually happened is that it opted out of most of the market and bought the small share that happened to be cheap.
The diagnostic is to look at spend against budget, not at cost against target. A capped campaign spending a third of its budget is telling you where the cap sits relative to the market.
The cap is not a cost per result
The most common confusion is between the price of an auction and the price of an outcome, and a bid cap only addresses the first.
Capping the bid at A$2 means no impression or click costs more than A$2. It says nothing about how many of them convert. If the traffic bought at that price converts poorly, the cost per conversion will be many multiples of the cap, and the cap will have done nothing to prevent it.
This is why a bid cap is a poor instrument for controlling acquisition cost even though it looks like one. The control that speaks the language of outcomes is a target or cost cap on the result; the bid cap speaks the language of inputs.
Where an account wants both — a ceiling on the result and a ceiling on any individual auction — it needs both settings, and should expect the combination to be restrictive.
Setting one
The number should come from the same place every price ceiling in an account comes from: what the thing being bought is worth.
Work back from the outcome. A conversion worth A$120 of contribution, arriving from clicks that convert at 4%, makes a click worth A$4.80 at break-even. A bid cap belongs under that, by however much of the contribution the business needs to keep.
What it should not come from is the current average cost. Setting a cap at the campaign's average guarantees that roughly half the auctions it used to win are now declined, which is a substantial reduction in delivery achieved by accident.
A workable practice is to set the cap at a multiple of the current average rather than at it — high enough that only genuinely expensive auctions are refused — and to lower it deliberately while watching spend, stopping at the point where volume starts falling faster than price.
Where it earns its place
Hard commercial ceilings. Where somebody has agreed a maximum and the agreement is real, an averaged control does not implement it.
Deliberate cherry-picking. A campaign whose job is to take only the cheap end of a market — filling residual capacity, opportunistic buying — is well served by a cap that refuses everything else.
Controlled tests. Knowing the maximum price of every unit makes a test result easier to reason about, because price variation is removed as an explanation for what happened.
Protecting against a mistake. A cap is a backstop against a misconfigured target or a runaway audience, and as a safety net set well above normal operation it costs nothing and occasionally saves something.
Where it does not
As the primary lever on acquisition cost. It controls the wrong quantity.
On a campaign that needs volume. Any campaign whose problem is not enough results is poorly served by a setting whose main effect is to decline auctions.
Alongside an automated strategy with a target. The strategy is already declining auctions it does not expect to clear its target; adding a per-auction cap on top means two constraints binding at once and a campaign that is very hard to diagnose. Where both are present and delivery is poor, removing the cap first is usually the faster path to understanding which one was binding.
Set once and forgotten. Auction prices move with competition and season. A cap that was generous in March can be binding in November without anybody changing it, and the campaign will quietly stop delivering.