CPM: Cost Per Mille Formula, Calculator, and Meaning

The standard formula is:

Google Ads expresses the same relationship as cost divided by impressions in thousands in its video advertising metric documentation. Mailchimp also uses total campaign cost divided by impressions, multiplied by 1,000, in its CPM overview. The arithmetic is stable across platforms. The definition of a qualifying impression can differ.

Example 1: calculate CPM from cost and impressions

A campaign spends $2,500 and delivers 400,000 impressions. Divide $2,500 by 400,000, then multiply by 1,000. The result is a $6.25 CPM. This means the campaign spent $6.25 for every 1,000 impressions counted by the report. It does not mean each 1,000 impressions came from 1,000 unique people.

Example 2: calculate budget from CPM and impressions

Rearrange the formula to plan cost: campaign cost = CPM x impressions / 1,000. If a proposal quotes a $9 CPM for 750,000 impressions, the planned media cost is $6,750. Keep fees and taxes separate unless the quoted rate explicitly includes them.

Example 3: estimate impressions from budget and CPM

Rearrange it again: impressions = budget / CPM x 1,000. A $12,000 budget at an estimated $8 CPM buys about 1.5 million impressions. This is an estimate, not a delivery guarantee. Auctions, pacing, eligibility, viewability rules, and the final placement mix can change the realized CPM.

An impression records an ad delivery event under a platform's rules. Reach estimates unique people or accounts exposed to the campaign. Frequency is impressions divided by reach. One person who receives five impressions adds five to the CPM denominator but roughly one to reach. Read the fuller ad impression definition before treating impressions as people.

Served does not necessarily mean viewable. Google's Active View documentation says a display impression is viewable when at least 50 percent of the ad area is visible for at least one second. For video, at least 50 percent must be visible while the video plays for at least two seconds. Large display units use a different pixel threshold. These rules follow industry viewability standards, and Google is explicit that viewability cannot prove a person looked at the screen. See Google's viewability definitions and the MRC viewable impression guideline.

Different denominators make many CPM comparisons unreliable. If Campaign A reports served CPM and Campaign B reports viewable CPM, the cheaper number does not establish cheaper viewable exposure. Check measurable rate, viewable rate, format, placement, device, and the exact metric label first.

CPM, eCPM, and vCPM all use a per-thousand scale, but they do not always describe the same business side or denominator. Treat the letters as instructions to inspect the report, not as interchangeable labels.

Google Ad Manager documents several eCPM variants based on ad requests, matched responses, or impressions. The number rises as the denominator narrows, even when revenue stays the same. Its eCPM denominator guide is a useful reminder that a metric name is incomplete without its counting point. For advertisers, Google's viewable CPM bidding charges against impressions measured as viewable, as described in its vCPM bidding documentation.

CPM answers a delivery-cost question. The metrics that follow it in the customer journey answer different questions. A campaign can have a high CPM and an efficient CPC if the ad earns a strong click-through rate. It can have a low CPC and a poor CPA if the clicks do not convert. It can show an attractive attributed ROAS while still failing an incrementality test.

When the same spend, impressions, and clicks are used, CPC can be derived from CPM and CTR. At a $10 CPM and a 1 percent CTR, 1,000 impressions generate about 10 clicks, so the effective CPC is about $1. At the same CPM and a 0.25 percent CTR, the effective CPC is about $4. This relationship helps explain a result, but it does not make CTR the only creative objective. Curiosity clicks can raise CTR and still weaken conversion quality.

Use CPC for traffic cost, CPA for a defined action, and ROAS or profit measures for financial outcomes. Then state the attribution modelthat assigned the conversion or revenue credit. CPM sits early in that chain. It is often diagnostic, but rarely sufficient as the campaign's success criterion.

CPM changes when the price or mix of eligible impressions changes. Because most digital inventory is selected through auctions or automated delivery systems, one blended number can hide several causes.

Segment before explaining. Break CPM down by country, platform, placement, device, format, audience, campaign, and time period. SocialPeta's market intelligence taxonomy places advertising cost research in the market insight workflow. Use external cost signals to form a hypothesis about competition or placement mix, then confirm the cause with first-party campaign reporting.

There is no universal good CPM. A useful benchmark must match the platform, country, objective, audience, format, placement, season, bid strategy, and impression definition. A broad display campaign and a narrow business-to-business video campaign are buying different inventory. Comparing their CPMs without context produces a precise-looking answer to the wrong question.

Start with your own like-for-like history. Compare a complete current period with a comparable prior period, then use a rolling median or weighted average for the same segment. External benchmarks can help flag an outlier, but they often combine accounts with different goals and measurement rules. Investopedia's CPM reference describes the metric as a way to assess relative campaign cost effectiveness, not as a universal performance grade.

A rising CPM can be acceptable if the campaign reaches a more valuable audience and downstream economics improve. A falling CPM can be harmful if delivery shifts toward low-viewability inventory or low-intent users. The Sprout Social CPM glossary and AdRoll's CPM guide both frame the metric around awareness, but fixed frequency rules or generic network benchmarks should not replace campaign-specific evidence.

Diagnose CPM in layers. First verify the calculation. Then locate the segment that changed. Only after that should you change bids, audiences, placements, or creative. This order reduces the chance of reacting to a reporting artifact or fixing the wrong part of the campaign.

1. Confirm scope and maturity. Match currency, time zone, date range, spend treatment, impression event, campaign status, and reporting delay.
2. Decompose the blended CPM. Segment by market, platform, placement, device, audience, creative format, and campaign. Calculate totals from summed cost and impressions instead of averaging row-level CPMs.
3. Inspect delivery conditions. Check eligible reach, frequency, bid and budget changes, auction diagnostics, inventory mix, measurable rate, and viewability.
4. Follow the funnel. Read CPM with attention, CTR, CPC, landing-page behavior, conversion rate, CPA, and qualified value.
5. Change one identifiable lever. Record the hypothesis and the metric expected to move. Avoid changing audience, bid, creative, and landing page together if you need to learn what caused the result.

To combine campaigns, sum all eligible cost and all matching impressions, then calculate CPM from those totals. Do not take a simple average of campaign CPMs. That would give a small campaign the same weight as one with millions of impressions and could distort the result. Practitioners ask about this often because dashboards can expose row metrics without making the aggregation rule obvious.

For a broader sequence that connects delivery, attention, conversion, and value, use SocialPeta's advertising analytics framework or the ad performance diagnosis workflow.