eCPM is probably the number you watch most closely. It shows up in every ad network dashboard, every mediation report, and every internal review deck. Yet publishers regularly misread it, misuse it, or compare it without the context needed to make it meaningful.
This guide covers what eCPM actually measures, the exact formula behind it, what good benchmarks look like for web and app publishers, and what genuinely moves the number.
What is eCPM?
eCPM stands for effective cost per mille, where “mille” is Latin for thousand. It tells you how much revenue you earn per 1,000 ad impressions, regardless of how the advertiser originally agreed to pay.
Advertisers buy inventory in different ways. Some pay per click (CPC). Some pay per install (CPI). Some pay per completed video view. eCPM normalizes all of those models into a single number so you can compare performance across campaigns, ad units, and networks on equal footing.
It’s a publisher-side metric. Advertisers think in CPM (what they pay). Publishers think in eCPM (what they earn). The two are related but not identical, because ad networks take a margin in between.
The eCPM formula
eCPM = (Total Revenue / Total Impressions) x 1,000
Divide your total ad revenue by the number of impressions served, then multiply by 1,000 to express it per thousand.
Example: You run a mobile game that served 500,000 impressions yesterday and earned $1,250 in revenue.
eCPM = (1,250 / 500,000) x 1,000 = $2.50
Your banner ad unit earned $2.50 for every 1,000 impressions it served.
For web publishers, the same formula applies. If your site had 2,000,000 pageviews, served 3 ad slots per page (6,000,000 impressions), and earned $9,000, your eCPM is $1.50.
Most ad networks calculate and display eCPM automatically in their dashboards, but knowing the formula matters when you’re aggregating data from multiple sources or validating numbers that don’t feel right.
eCPM vs CPM: what’s the difference?
CPM (cost per mille) is the advertiser’s bid, what they’ve agreed to pay. eCPM is the publisher’s yield, what actually lands in your account after the network takes its share.
An advertiser might bid $5 CPM on your inventory. After the network’s revenue share (typically 30-45%), your eCPM might come out to $2.75 to $3.50.
The more useful distinction for publishers: CPM applies to fixed-rate or direct campaigns; eCPM applies to everything, including CPC and CPA deals converted into impression-equivalent earnings. This is what makes eCPM a universal benchmark.
What’s a good eCPM benchmark?
There’s no universal “good” number. eCPM varies significantly by platform, geography, ad format, vertical, and time of year. Below are realistic benchmarks based on 2025 industry data for both app and web publishers.
App publishers (in-app advertising)
| Ad Format | Typical eCPM Range (US) | Tier |
|---|---|---|
| Offerwall | $50 – $200+ | Very High |
| Rewarded Video | $8 – $25+ | High |
| Interstitial (full-screen) | $4 – $12 | Medium |
| Native | $2 – $8 | Medium |
| MREC / Medium Rectangle | $1 – $4 | Low |
| Banner (320×50) | $0.30 – $1.50 | Low |
Web publishers (display advertising)
| Format / Placement | Typical eCPM Range (US) | Tier |
|---|---|---|
| Video (pre-roll / outstream) | $5 – $20 | High |
| Sticky / Anchor | $2 – $6 | Medium |
| Native Ad Unit | $1.50 – $7 | Medium |
| Leaderboard (728×90, above fold) | $1.50 – $5 | Medium |
| Medium Rectangle (300×250) | $1 – $4 | Low |
| Below Fold / Sidebar | $0.40 – $1.50 | Low |
Geography matters a lot. US, UK, Canada, and Australia consistently produce the highest eCPMs across both web and app inventory.
Tier 2 markets (Western Europe, Japan, South Korea) sit 30-60% lower. Tier 3 markets (Southeast Asia, South Asia, LATAM) can be 70-90% below US levels. Always segment your eCPM by geo before drawing conclusions.
Seasonality also has a real impact. Q4 (October through December) typically sees 30-60% higher eCPMs across all formats due to increased advertiser spend around the holiday shopping season. January is usually the lowest eCPM month of the year. Build this expectation into your planning.
What affects your eCPM
Ad format: Rewarded video and interstitials consistently outperform banners. Format choice is the single biggest lever most publishers control directly.
Geography: Where your users are located has more impact on eCPM than almost any other variable. US traffic is worth significantly more than global average traffic.
Ad placement and viewability: Above-the-fold placements command higher bids. Viewability rates below 50% will suppress demand. On the web, placements with 70%+ viewability see noticeably better fill.
Demand competition: Using header bidding (web) or in-app bidding (mobile) puts multiple buyers in competition for each impression. More competition means higher effective yields.
App category / content vertical: Finance, health, and lifestyle content attracts premium CPMs. Gaming apps generally command higher eCPMs than utility apps due to engagement levels and audience profiles.
User engagement: Time in session, session depth, and return visit rate all feed signals to advertisers and algorithms. Deeper engagement generally means higher eCPMs over time.
Device and OS: iOS eCPMs are consistently 20-40% higher than Android for comparable inventory, largely driven by advertiser attribution capabilities and audience spending power signals.
Mediation stack quality: A poorly configured mediation stack with outdated floor prices and missing demand partners can silently suppress eCPM by 20-40% compared to a well-managed setup.
How to improve your eCPM
For app publishers
Switching to in-app bidding from a traditional waterfall is typically the highest-impact change you can make. Waterfalls with manually set floors miss real-time demand. Bidding gives every ad network an equal shot at each impression, and prices clear at the true market value.
Also read: In-App Bidding vs. Waterfall: A Guide for App Publishers
But if you want to stick to waterfall and don’t have to make huge changes to your ad stack, we suggest you work with us, UndrAds. It’s impossible for a human to optimize so many bids coming in, hence, we have trained AI agents who do this for you. All bids, optimized.
On format mix: just running banners is not a good idea. Adding rewarded video, even at lower frequency, tends to lift overall eCPM meaningfully because the format commands premium rates and doesn’t necessarily hurt retention when placed at natural session breaks.
Floor prices matter, but the direction most publishers get wrong is setting them too high and causing fill rate to collapse. A low fill rate means more unsold impressions, which hurts overall revenue even if the ads that do serve are at a better rate. Test floors carefully with enough data before committing.
Also check: Top 25 Mobile Gaming Publishers and Developers by Revenue
For web publishers
If you’re not running header bidding, that’s the first place to look. Platforms like Google Ad Manager with Open Bidding, or a standalone header bidding setup, generally outperform a single-demand-source setup because competition drives prices up.
Viewability optimization directly affects eCPM on web. Lazy loading ads so they only render when a user is likely to see them, reducing page load time, and keeping ad density reasonable all improve viewability scores, which in turn make your inventory more attractive to programmatic buyers.
Audience data activation can move eCPMs significantly for content publishers. If you have logged-in users or can build first-party audience segments (interests, behavior, demographics), those signals let advertisers target more precisely and justify higher bids.
Improving eCPM isn’t always about chasing the highest rate. A slightly lower eCPM with 95% fill is often better total revenue than a higher eCPM with 65% fill. The number that actually matters is total revenue = impressions x (eCPM / 1,000).
Common mistakes when reading eCPM
Comparing eCPMs across ad networks without controlling for format, geography, or time period is one of the most common misreads in ad ops.
Network A might show a $6 eCPM and Network B might show $3, but if Network A is serving rewarded video to US users and Network B is serving banners to global traffic, those numbers say nothing meaningful about which network performs better on equivalent inventory.
A second trap: treating eCPM as a quality metric in isolation. High eCPM with low fill rate means many impressions aren’t earning anything. If your overall effective yield (eCPM multiplied by fill rate) is declining, a rising eCPM headline number can actively mislead.
Third: ignoring day-of-week and seasonality patterns. eCPM on weekdays typically runs higher than weekends for B2B-adjacent content. Q1 runs lower than Q4 almost universally. Short-term eCPM drops can cause unnecessary alarm if you don’t have context for the baseline.
Finally, watch for the difference between gross eCPM and net eCPM. Some networks report what advertisers paid (gross); others report what you receive after their take. If you’re comparing networks, make sure you’re comparing the same measure.
Want higher eCPMs without the manual ops overhead? UndrAds automates your mediation, floor pricing, and demand partner management so your stack is always working at full capacity.
