What is CPC?

A campaign spends A$4,500 in a month and the platform reports 3,000 clicks. CPC is 4,500 ÷ 3,000 = A$1.50. If only 1,800 of those were outbound link clicks and the rest were reactions, comments and taps on the page name, the cost per link click is 4,500 ÷ 1,800 = A$2.50 — the same spend, the same campaign, and a figure two-thirds higher. Nothing about the auction changed; only the definition did.

Cost per click divides what a campaign spent by the number of clicks attributed to it. As arithmetic it is the simplest metric in advertising. As a number to compare, it is among the least stable, because every platform counts a different set of events as a click and most of them report more than one.

Meta reports an all-clicks figure that includes reactions, comments, shares, taps on the page name and expansions of the image, alongside a link-clicks figure that counts only the taps that sent somebody to a destination. Both are divided into spend and both come back labelled CPC. Google Ads counts a click on the headline of a search ad, and reports interactions separately for the formats where a click is not the action being bought. Video placements generally count a click on the call-to-action rather than a tap on the video itself.

The practical consequence is that two people comparing click prices are usually comparing two different denominators. Before any CPC figure is worth an argument, settle which clicks were in it. A campaign whose all-clicks price looks healthy and whose link-clicks price does not is telling you that the ad earned engagement rather than traffic, and that is a creative finding rather than an auction one.

Cost per click is not set directly, even when a bid is typed into a box. It falls out of two other numbers:

Cost per click = (CPM ÷ 1,000) ÷ click-through rate

An advertiser paying a A$12 CPM at a 1% click-through rate is paying A$1.20 a click. The same A$12 CPM at 2% is paying A$0.60. Nothing about the auction, the bid or the budget has to change for the click price to halve; the creative can do it on its own.

That identity is the most useful thing about the metric, because it says where to look when the number moves. A CPC that rose while CPM held flat is a click-through-rate problem: the creative has fatigued, the audience has shifted, or the hook has stopped working on the people being reached. A CPC that rose while click-through rate held is a delivery problem: more competition in the auction, a broader or more expensive audience, a seasonal bid-up, or a drift in placement mix toward pricier inventory.

Reading the two inputs together takes a minute and settles which half of the account to open. Reading the click price alone produces a fortnight of bid adjustments aimed at a creative problem.

Search and social auctions are second-price or a variant of it: the winner pays what it needed to beat the next bidder, not what it was willing to pay. A manual CPC bid is therefore a ceiling rather than a price there, and an account whose average click price sits well under its bid cap is not underperforming — it is winning auctions cheaply. Programmatic display largely moved to first-price, where the winner pays their bid and no such gap appears.

Auctions also rank on expected value rather than on money alone. Google's ad rank folds a quality estimate into the bid; Meta's total value estimates an action rate and a quality signal alongside it. The advertiser whose ad is more likely to earn the click can win the same position for less. This is why better creative lowers the click price twice: once by producing more clicks from the same impressions, and again by making the platform willing to serve those impressions at a lower price.

It is also why bidding harder is the weakest available response to an expensive click. Raising the bid raises the ceiling. It does not make the ad more likely to be chosen.

Inside a single account the click price varies by more than most advertisers expect, and nearly always for reasons about who was reached rather than how much was bid.

Placement is the largest of them. Feed inventory, stories, in-stream video, search partners and audience networks are separate markets with separate demand, and a campaign left to distribute itself across all of them reports one blended figure that describes none of them. Splitting a report by placement usually shows a spread wide enough to change a decision.

Audience is the second. Retargeting a warm list competes for a small, heavily bid-on set of people; broad prospecting competes for a large, cheap one. The retargeting click almost always costs more and almost always converts better, which means the campaign with the lowest click price in an account is frequently the one producing the least revenue.

Geography, device, language, day of week and hour of day all move the figure for the same reason. Each is a separate auction with its own bidders, and an average across all of them is a number no single auction ever charged.

It says nothing about what happened after the click. An account can halve its click price by chasing cheap attention — a broad audience, cheap placements, a headline that provokes rather than qualifies — and end up paying more per customer than it did before. Cheap clicks that do not convert are the most common way an account gets worse while one of its headline numbers improves.

It says nothing about volume either. The cheapest clicks in any account are bought first, so the average price rises as budget rises. A click price held rigid while spend scales is a cap on the account wearing the costume of a target, and it shows up later as delivery that will not grow.

And it is unstable in small numbers. A campaign with forty clicks in a week has a figure that moves on the next two, which is why week-on-week click-price panics are usually noise, and why automated bid strategies struggle on accounts this thin.

Work the click-through rate before the bid. The creative, the first seconds of the video, the offer stated in the headline and the match between the ad and the page it promises are all yours, while the auction has other bidders in it. Halving the CPM and doubling the click-through rate move the click price by exactly the same amount, and only one of those is under your control.

Then narrow what is being bought. Excluding placements, audiences, geographies and search terms that have spent without ever converting raises the average quality of the traffic and lowers the price of the traffic that remains.

Then check the denominator. A reporting view set to all clicks flatters every figure in the account; switching it to outbound clicks is often the single change that makes the rest of the analysis honest.

Then, and only then, adjust the bid — and adjust it against what a click is worth to the business, which is the conversion rate multiplied by the contribution margin of whatever the conversion produces. A click that earns A$4 of contribution can be bought at A$2 all day. A click that earns A$0.40 cannot, at any bid.

• Read the click price beside CPM and click-through rate, never on its own. The three are one identity, and whichever of the other two moved tells you whether to open the creative or the targeting.
• Fix the denominator before comparing anything. An all-clicks figure and an outbound-clicks figure from the same campaign can differ by a third or more, and that gap is a reporting setting rather than a performance change.
• Test creative before touching bids. A better hook lowers the click price twice — more clicks from the same impressions, and cheaper impressions, because the auction rewards the ad it expects to be clicked.
• Exclude what has never converted. Placements, audiences, geographies and search terms carrying spend and no result raise the average price of every click around them.
• Judge a click price against the cost per conversion it produced, not against a cheaper click somewhere else. The cheapest clicks in an account are usually the least valuable ones.