top of page
  • Writer's picturedKilo

CPA vs. CPM vs. CPI | Which is better for your DOOH campaign?

Could the same metrics used online apply when measuring digital out-of-home ads? What’s the equivalent of Dwell time or OTS in online ads? or what’s the equivalent of CTR or CPC in Digital Out of Home ads?

Usually, digital ads have been sold at an impression level, where advertisers paid for said impressions on a cost per thousand principle known as CPM (technically, cost per mille, which is Latin for “thousand”). During the early days of mobile marketing, many advertisers shifted to cost per install (CPI) The cost per install applies only to ads converting these impressions to app installs and is a term used by digital marketers online, however, as soon as realization dawned that there was more to growth than app store rankings and that an install was not a user, many opted to move towards the cost per action (CPA) which are great in theory, but attribution in the mobile space is neither perfect nor relevant in the DOOH space as the CPA methodology puts more risk on the publisher.

What are the different pricing models for advertising Campaigns?

The most common pricing model for programmatic DOOH advertising is CPM. The pricing model defines incentives across the process. The wrong pricing model sets the wrong incentives and therefore jeopardizes the campaign performance — regardless of how granular the targeting or how advanced the creative strategy may be. To maximize transparency and performance, marketers should pay CPM and optimize for CPA; this article explains why.

CPA vs CPM pricing

What is CPA? How is CPM calculated? What are the main differences between each pricing strategy? Below, you’ll find each model explained: definitions, pros, cons, and why you should only pay CPM for your marketing campaigns.

CPA — (Cost Per Action Campaigns)

CPA = Advertising Costs / Number of Actions

CPA became increasingly popular as many companies argued that it removed the risk from the marketer side: “you only pay for actual actions.”

The actions defined in a cost-per-action agreement relate directly to some type of conversion, with sales and registrations among the most common.

The cost-per-action (CPA) model is at the other end of the spectrum from the cost-per-impression model (CPM), with the cost-per-click (CPC)-Which is not applicable in the DOOH channels- model somewhere in the middle. In a CPA model, the publisher takes most of the advertising risk, as the campaign success depends on good conversion rates from the advertiser’s creatives and marketing strategy.

As a publisher, the notion of accepting CPA campaigns is dependant on inventory utilization, with priority given to CPM campaigns.

CPM — (Cost Per Thousand Impressions Campaigns)

CPM = (Advertising Cost / Total impressions) x 1000

For many years, the cost per thousand (CPM) pricing has been the main buying model for premium display advertising space. CPM allows publishers to generate revenue based on volume and allows advertisers to pay for impressions at a fixed price per thousand.

The CPM buying model works perfectly for premium inventory where the value proposition is more about branding than direct response, like buying a full-page ad in a glossy magazine. On the other hand, cost per action or acquisition (CPA) models have been utilized as an alternative to CPM by publishers looking for maximum fill rates to ensure an ad is served for every impression available.

Interestingly, CPA is often perceived to be more appropriate for performance-focused advertisers, as these methods allow cost management based on a predefined volume of users visiting a site or completing an action.

However, from a publisher’s point of view, only generating revenue when a user clicks or goes on to complete an action is a risk. It also assumes there is zero value in showing someone an ad they do not directly interact with.

Consequently, there are limitations in the types of the advertising inventory that can be purchased if buying on a CPA basis. While there are specialist suppliers who may have relationships with premium publishers allowing them to buy on a CPM basis and sell CPA, in reality, quality publishers know the value of their inventory. This means buying on a CPA model will limit reach across these premium channels.

What makes one impression better than another?

The answer to this question is simple — how well it performs. Measuring this means ensuring there are clear performance goals. These goals can still be based on effective CPA results worked back from the CPM, or could be about brand awareness and measured accordingly. Put simply, the best impressions will be the ones that most frequently deliver the required results.

Working out which impressions fall into this category involves periods of testing, learning, and optimizing. The testing phase for a campaign will depend on how quickly the system (and the resource managing it) can identify impressions that deliver the right result. Once these have been found, campaign management is a process of managing budgets and refining targeting.

Enter the Conversion Per Impression (CPI)

This is a metric that shows the ratio between impressions and conversions. What CPI (Cost per Impression) does is show you which ad will receive the most conversions from the impression.

Every time your ad is displayed; you have a chance of a conversion. The user must first see your ad and then convert to receive the actual conversion, but measuring from the impression shows you the total conversions possible. Some of the oldest studies in marketing mention the rule of 7 (it takes an average of seven interactions with your brand before a purchase will take place), and the more famous rule of 27 in the sense that you truly see only one out of every three ads put in front of you. And you take action only after you’ve seen a brand’s marketing nine times. So, it takes at least 27 impressions of a brand before a prospect even responds.

Think about that for a second. Are your prospective customers getting all 27 hits? If not, you’re basically pouring your marketing budget down the drain. Let’s look at how CPI could give you better visibility on your ad spend.

CPI is calculated by dividing total impressions by the total conversions.

CPI = conversions / impressions

It is generally displayed as a percentage. Here are some examples:

The Advantage of using CPI as Your Testing Metric

The main reason to use CPI (Cost Per Impression) is when you want the most conversions possible. This metric is a simple metric that will show you which ad will lead to the most total conversions possible.

The methodology of measuring conversions vs. impressions in DOOH campaigns varies from one campaign to another and from one company to another. Some campaigns will utilize a QR code where their audience will scan to convert, while others will choose a certain promo code that’s used solely from their DOOH campaigns to track their conversion against said campaigns.

Working with CPI

There are times that when you examine the full metrics behind various ads, you might not always pick the highest CPI winners. This usually happens for a few reasons.

  • You are struggling with Quality Score and you want to pick a winning ad set if it leads to similar total conversions to raise QS.

  • You have a hard cap on how high your CPA can be and therefore, you have to pick an ad that is under your target CPA.

  • You only want a certain number of leads per month; in which you’ll pick the ad that can hit your total leads for the lowest cost.

  • In these cases, we would argue that CPI is still your best metric to use, you just need to readjust what a conversion is. For instance, someone calling your short number to enquire about your product could be a conversion, or even visiting your website could be a conversion. You decide; and therefore, CPI is still your best metric if you pick a “good” conversion metric.

Let’s take a look at the below example then examine how we’d pick the various winners.

If we just want the most conversions possible; then ad 7 is our clear winner. It receives 100 conversions for every 50K impressions. This is why CPI is such a great metric. Ads 5 and 6 expanded their reach on 100 cars because the strategy behind these campaigns was more focused on awareness however that didn't reflect optimum results when compared to the rest of the ad sets. In comparison to the other ads; ad 2 has the lowest number of cars and the most condensed ad slots, however, the CPA and CPI were definitely telling a different story which makes it a lousy campaign. Ads 3, 4, and 8 were competing closely However when you use both CPI and CPA to calculate the most conversions possible for an ad; then ad 7 is our clear winner.

If we are struggling with Quality Score; then we might pick ad 8 as it has a much higher CPI than the other ads relative to the number of cars, which is also why it has a lower CPA than the other ads, and its CPI is not too far behind our winner.

If our goal is 60 L.E. ($4) CPA and we want the most conversions under these 60 pounds; then we’d eliminate all the ads with a CPA higher than 60 and pick the highest CPI ad from the ones that are left as our winner.

When Not to use CPI

The one downfall of CPI is that it doesn’t care about generated revenue or total order amounts. If you have an online business, or an e-commerce website or app, and want to base your conversions on Return on Ad Spend (ROAS) then CPI is not a great number to use, especially in DOOH advertising because calculating CTR or CPC is very challenging.

CPI is great for lead generation, or the most conversions possible. However, since revenue and ROAS are not numbers used in its calculation, there are other metrics, such as RPI/PPI (revenue/profit per impression) metrics that are better to use for e-commerce companies. We’ll cover these additional metrics in the coming weeks.

dKilo pays people for being stuck in traffic by pushing digital ads to their cars, and sell the digital ooh ad space to companies and brands through an intelligent impressions' based online ads manager platform.


bottom of page