CPM Calculator

How to Use the CPM Calculator

Cost per thousand impressions is how display, video and most social advertising is priced. It is also the metric most easily misread, because a cheap impression price attached to a poor click-through rate is an expensive click price wearing a disguise.

1. Enter the campaign budget and impressions. Either the ones a platform delivered, or the ones a media plan promises. The calculator works equally well as a post-campaign measurement and as a pre-campaign estimate.
2. Enter the click-through rate. This is what bridges the impression price to the click price. Display often runs between 0.05% and 0.5%; social sits higher; video pre-roll lower still.
3. Enter the conversion rate for the traffic this campaign sends. Awareness-led traffic converts a good deal worse than search traffic, and using a site-wide average here flatters every display campaign you will ever run.
4. Enter the average frequency — how many times the typical person sees the ad. Impressions divided by frequency gives reach, which is the figure that actually matters in an awareness campaign.
5. Compare the effective cost per click against what you pay in search. That single comparison settles most arguments about where a display budget should go.

If the effective cost per click comes out higher than your search cost per click, the campaign is buying expensive traffic in a cheap-looking wrapper.

CPM Formula

One division, and then two bridges into the metrics that matter.

Frequency deserves more attention than it usually gets. The same budget here buys 581,250 people seeing an ad 3.2 times or 1,860,000 seeing it once, and those are different campaigns with different purposes. Below roughly three exposures most brand messages do not register; above eight or nine the extra impressions are mostly irritating people who have already decided.

Example

$12,400 buying 1,860,000 impressions

1. CPM: 12,400 ÷ 1,860,000 × 1,000 = $6.67.
2. Clicks at a 0.85% rate: 1,860,000 × 0.0085 = 15,810.
3. Effective cost per click: 12,400 ÷ 15,810 = $0.78.
4. Conversions at 2.2%: 15,810 × 0.022 = 347.8.
5. Cost per conversion: 12,400 ÷ 347.82 = $35.65.
6. Reach at a frequency of 3.2: 1,860,000 ÷ 3.2 = 581,250 people.

Cheap impressions, cheap clicks — this time

A 78-cent click is genuinely good compared with most paid search, and that is what makes this campaign worth running. But the whole result rests on the 0.85% click-through rate. Halve it to 0.425% and the effective cost per click doubles to $1.57 while the CPM does not move at all. The media price is not the variable that decides whether display is cheap; the creative is.

Reach or frequency?

The same $12,400 could be planned as 581,250 people seeing the ad 3.2 times, or 1,860,000 people seeing it once. For a brand launch that needs recognition, the first is far more effective and the second is close to wasted. For a reminder to an audience that already knows you, the reverse can be true. Deciding which you want before buying is the difference between a media plan and a media purchase.

Why a Cheap CPM Is Not a Cheap Campaign

What $12,400 buys at different impression prices, holding the click-through rate constant.

The relationship is exactly proportional: double the CPM and you halve the impressions, the clicks and everything downstream. That linearity makes CPM easy to negotiate and easy to misjudge, because a 20% better media price and a 20% better creative produce identical arithmetic — and only one of them is available on demand.

Impression prices vary by an order of magnitude across placements. Programmatic display can clear a couple of dollars per thousand; social feeds typically run between $5 and $15 depending on audience; connected television and premium video routinely exceed $25. A high CPM is not automatically bad — a $30 impression seen by exactly the right person can be cheaper per outcome than a $2 impression served to nobody in particular.

That is why the cost per conversion is the row to end on. It absorbs the media price, the creative quality and the audience quality into one number, and it is directly comparable with what you pay for a conversion in search. The CPC calculator gives you the search side of that comparison, and the ROAS calculator tells you whether either is affordable at your margin.

Five Things a CPM Does Not Tell You

Five things a CPM figure does not tell you.

Whether the ad was seen. An impression is served, not necessarily viewed. Viewability standards typically require half the pixels in view for a second or two, and a campaign at 45% viewability is paying full price for impressions that half of nobody saw. Ask for the viewability figure, not just the impression count.

Whether a human saw it. Invalid traffic exists on every open exchange. Reputable inventory and a fraud-detection layer cost more per thousand and usually cost less per outcome.

Where it appeared. The same CPM buys a premium publisher's article page or an auto-refreshing slot at the bottom of a content farm. Placement reports are tedious to read and are where most display waste is actually found.

How concentrated the frequency was. An average of 3.2 can mean everyone saw it three times or that a tenth of the audience saw it thirty times. Frequency distribution, where a platform reports it, is far more useful than the average.

What it did that clicks cannot show. Display and video mostly work by being remembered rather than clicked, so judging an awareness campaign purely on its click-through rate measures the wrong thing. That is a real limitation of this page and of the metric — the arithmetic here is sound and the model of how advertising works behind it is deliberately simple.

The most useful discipline with impression buying is to decide the objective before opening the media plan, and to hold the reporting to it. A campaign bought for reach should be judged on reach, frequency and any brand measurement you can afford; a campaign bought for response should be judged on cost per conversion. The failure mode is buying for one and reporting on the other, which happens constantly and produces the reliable conclusion that display does not work — a conclusion drawn from measuring a reach campaign with a response metric.

For campaigns bought on clicks rather than impressions, the CTR calculator and the CPC calculator cover the same ground from the other direction, and the engagement rate calculator is the better fit for social placements where interaction rather than a click is the goal.