The CTR Formula, and the CPC It Commits You To

Your click through rate went from 1.2 percent to 0.9 percent and the room agrees the creative is tired. Maybe it is. But before anybody briefs a new one, look at what that drop has already done to the price you are paying: at a $12.00 CPM your cost per click went from $1.00 to $1.33. A quarter came off your CTR and a third went onto every click you buy for the rest of the month.

That is why the CTR formula is worth more than thirty seconds. The arithmetic is trivial and almost every guide stops there. What they leave out is that CTR is not a report card on your creative. It is the exchange rate between the two prices an auction will sell you, and it is the one term in that exchange you are allowed to move.

The formula, and the thing it is denominated in

Nobody disputes this part:

CTR = clicks / impressions x 100

Nine hundred clicks on 75,000 impressions is 1.2 percent. The argument is never about the division. It is about what counted as a click and what counted as an impression, and the two platforms you spend most of your money on answer both questions differently.

On Meta, the column labelled CTR (all) counts every click the creative received, including expands, likes, profile taps and comments. The column labelled CTR (link click) counts only the people who left for your site. The first is always the larger of the two and can be several times the second, and it is the one that tends to sit in the default view. If you have ever been handed a 3 percent CTR alongside traffic that behaves like 1 percent, that gap is not a tracking bug. It is two different numerators wearing the same name.

The denominator splits the same way. Search impressions count auctions your ad appeared in, on queries somebody typed deliberately. Display and social impressions count renders, including the ones that scrolled past below the fold, which is why viewable impressions exist as a separate metric and why a viewable CTR is always the flattering version. Email adds a third split: divide by delivered and you get CTR, divide by opened and you get click to open rate, and the two answer different questions about the same send.

One rule keeps you out of trouble here. Your CTR has to be denominated in the same unit your bill is. You are charged per impression served, not per impression seen, so the CTR you use for money decisions is built on impressions served. Keep the viewable version for the separate question of whether the placement is worth buying at all.

CTR is the exchange rate between CPM and CPC

Here is the identity that makes the number useful. A thousand impressions cost you the CPM and produce CTR x 1000 clicks, so:

CPC = CPM / (CTR x 1000)

Run a $12.00 CPM through it. At a 0.6 percent CTR a thousand impressions buy 6 clicks and each one costs $2.00. At 1.2 percent they buy 12 and the click costs $1.00. At 2.4 percent, 24 clicks at $0.50. You are not choosing between a CPM buy and a CPC buy. You are buying the same inventory at the same price and letting your CTR set the denominator on your own invoice.

Read that column from the bottom up and the asymmetry shows. Going from 1.2 to 2.4 percent, a doubling, takes $0.50 off the click. Going from 1.2 down to 0.6 percent, the same doubling in the other direction, adds a full $1.00. CTR moves your CPC along a curve rather than a line, so an identical relative slip costs you more the lower you already sit. That is why a weak CTR on an expensive placement is quietly the most costly thing in most accounts, and why it is worth running the figures through the CTR calculator and the CPC calculator before the creative argument starts.

The CTR your margin requires

None of that tells you which CTR you need. For that you have to come at it backwards, from the only ceiling that actually applies to you, which is what an order is worth once the cost of filling it is out. Across this site the worked example runs on an $86.00 average order at a 42 percent contribution margin, so $36.12 is what a customer has to pay for.

If two clicks in a hundred become an order, a click can be worth at most $36.12 x 0.02, which is $0.72. Put that back through the identity at a $12.00 CPM and the CTR you need is 12 / (1000 x 0.7224), or 1.66 percent. Not a good CTR. The break-even one. Below it the placement is selling you clicks you cannot afford, however handsome the creative is.

That table says something uncomfortable about where CTR work belongs. Moving your conversion rate from 2 to 3 percent drops the CTR you need by a third, from 1.66 to 1.11 percent, and nobody touched an ad to do it. A landing page fix and a creative fix are competing for the same job, and the page is usually the cheaper of the two to attempt. Your own figures go into the conversion rate calculator and the contribution margin calculator, and the media half of it sits in the CPM calculator.

When a better CTR costs you money

CTR is also the easiest metric in the account to improve dishonestly, which is exactly why it should never carry a target of its own. Broaden the hook, promise a little more than the page delivers, put a face and a big number in the thumbnail, and the rate goes up. So does the share of people arriving with the wrong expectation.

Work the trade through. You lift CTR from 1.2 to 1.8 percent, a 50 percent gain, and your CPC falls from $1.00 to $0.67. Every figure on the ads dashboard improved. But the looser hook drags your conversion rate from 2.0 percent down to 1.2, and your cost per acquisition, which was $1.00 / 0.02, or $50.00, is now $0.67 / 0.012, or $55.56. You made clicks a third cheaper and customers 11 percent dearer.

This is the trade that CTR optimisation makes by default and rarely declares. The only safe way to hold it is to judge a creative test on cost per acquisition, or on contribution per thousand impressions, and to treat CTR as the diagnostic that tells you which half of the funnel moved. A CTR that climbs while CPA climbs with it is not a win that needs more time to prove itself. It is a gap between what the ad promised and what the page does.

Why CTR benchmarks travel badly

A 0.9 percent CTR on brand search terms would be a catastrophe. The same figure on a cold prospecting audience would be respectable. Search CTR is measured against an auction you had already qualified for, on a query somebody typed on purpose. Social CTR is measured against a feed nobody asked you to appear in. Same division, two different populations, and the results do not convert in either direction.

Placement, format, position, device and season move it as well, which is why an industry average CTR mostly describes that industry's traffic mix and says very little about your account. Use published figures, the ones on this site included, as a check that you have not misplaced a decimal point. The number that constrains you is the break-even CTR you worked out two sections ago, and it belongs to your margin and your conversion rate rather than to your sector. The same caution applies to CPM benchmarks and to every other cost figure somebody hands you.

Do one thing this week. Take your three largest line items, pull the CPM and the CTR each one actually delivered, and compute the break-even CTR that its conversion rate and your contribution margin allow. Anything running below its own line is not a creative problem waiting for a brief. It is a placement you are currently paying to lose money on, and the quickest fix is usually to stop buying it.

Related reading: what counts as a good CTR and CPM vs CPC.