Skip to content
Menu

Methodology

Last updated:

How every estimate on this site is produced, from the tier split to the quality checks on generated pages.

Every number on this site is an estimate built from publicly documented ranges and rounded conservatively. None of it is a quote from, or an endorsement by, any network. This page sets out every assumption so you can check it.

What the values represent

Mobile values are eCPMs: dollars earned per 1,000 ad impressions, after the network's revenue share. Website values are page RPMs: dollars earned per 1,000 pageviews at a reference layout of four ads per page and 60% viewability. Session RPM is page RPM multiplied by pages per session. The low and high figures shown beside each result are 0.7 times and 1.35 times the central estimate.

Tier system

Audience location is reduced to three tiers, and a single geography slider sets the mix between them.

  • Tier 1: United States, United Kingdom, Australia, Canada.
  • Tier 2: Western Europe and developed Asia-Pacific (Japan, South Korea, Singapore, New Zealand).
  • Tier 3: Rest of Asia, Latin America, Africa and Eastern Europe.

The slider anchors are 5/20/75 percent (Tier 1/2/3) at the Asia and Latin America end, 30/30/40 at the midpoint and 70/22/8 at the US, UK and Australia end, with straight-line interpolation between them.

Mobile eCPM calculation

monthly revenue (per format) =
  DAU × impressions per user per day × 30
  × (blended eCPM / 1000) × fill rate

blended eCPM =
  Σ (tier share × base eCPM[network][format][tier])
  × category multiplier × platform multiplier

Banner frequency is entered per session and multiplied by sessions per user per day. The iOS multiplier is 1.3 and Android is 1.0. Base eCPMs below are Android figures, in dollars per 1,000 impressions for Tier 1 / Tier 2 / Tier 3.

NetworkRewarded videoInterstitialBannerNativeApp open
Google AdMob12.0 / 5.0 / 1.48.0 / 3.2 / 0.900.90 / 0.40 / 0.103.5 / 1.5 / 0.405.5 / 2.2 / 0.60
Unity Ads13.0 / 5.2 / 1.39.0 / 3.4 / 0.900.80 / 0.35 / 0.09n/an/a
AppLovin MAX14.5 / 5.8 / 1.69.8 / 3.8 / 1.01.1 / 0.45 / 0.124.0 / 1.7 / 0.456.0 / 2.4 / 0.65
Meta Audience Network10.5 / 4.4 / 1.17.5 / 3.0 / 0.800.85 / 0.38 / 0.094.2 / 1.8 / 0.45n/a
ironSource / LevelPlay12.5 / 5.0 / 1.48.5 / 3.3 / 0.900.85 / 0.38 / 0.103.0 / 1.3 / 0.35n/a
Vungle / Liftoff12.0 / 4.8 / 1.28.2 / 3.2 / 0.800.70 / 0.30 / 0.083.2 / 1.4 / 0.35n/a
InMobi9.0 / 4.2 / 1.56.0 / 2.8 / 1.00.65 / 0.30 / 0.113.0 / 1.4 / 0.45n/a
Chartboost10.0 / 4.0 / 1.17.2 / 2.9 / 0.800.70 / 0.30 / 0.08n/an/a

Category and niche multipliers

CategoryMultiplierNicheMultiplier
Hypercasual Game0.80xTechnology1.15x
Casual Game0.95xFinance & Business1.90x
Mid-core / Strategy1.35xHealth & Wellness1.10x
Puzzle Game1.00xFood & Recipes1.05x
Utility App0.90xTravel0.95x
Social / Lifestyle1.00xGaming0.70x
Finance App1.60xLifestyle0.85x
Entertainment App0.95xNews & Entertainment0.65x
Education App0.85xEducation0.90x
General0.80x

Web RPM calculation

monthly revenue = (pageviews / 1000) × page RPM

page RPM =
  Σ (tier share × base RPM[network][tier])
  × niche multiplier × ads-per-page factor
  × viewability factor × format mix factor

ads-per-page factor = sqrt(ads / 4)
viewability factor  = viewability / 60

Pageviews are monthly sessions times pages per session. The format mix factor adds each enabled format's weight; the default display, in-article and sticky set equals 1. Base page RPMs in dollars for Tier 1 / Tier 2 / Tier 3:

NetworkPage RPM (T1 / T2 / T3)
Google AdSense7.00 / 3.00 / 0.80
Ezoic9.50 / 4.00 / 1.10
Media.net5.50 / 2.40 / 0.70
Monumetric10.50 / 4.40 / 1.20
Mediavine17.00 / 7.00 / 2.00
Raptive18.00 / 7.40 / 2.10

Eligibility flags use the entry bars the networks publish (sessions for Mediavine, pageviews for Raptive and Monumetric). Bars change, so treat them as a starting check.

Typical ad setup by category

The estimate pages use a typical setup for each app category instead of one default for everything: how often each format is shown, how many sessions a day users open, how long they stay and how many installs are still active after one, seven and thirty days. Revenue per install multiplies revenue per daily user per day by the expected number of active days, using a curve drawn through those three retention points.

CategorySessions per user per dayTypical session lengthRetention day 1 / 7 / 30Rewarded videoInterstitialBannerNativeApp open
Hypercasual Game4.0430% / 8% / 2%1.5610-1
Casual Game3.5838% / 14% / 6%336--
Mid-core / Strategy3.01635% / 16% / 8%4.51---
Puzzle Game3.01140% / 18% / 8%335-0.5
Utility App2.0225% / 10% / 5%-1421
Social / Lifestyle5.0928% / 12% / 6%-0.5361
Finance App2.0430% / 15% / 9%-0.5321
Entertainment App3.01330% / 11% / 5%12431
Education App1.51027% / 11% / 5%1131-

Seasonality

Advertiser budgets follow the calendar: they peak in the fourth quarter and reset in January. Each website niche has its own monthly curve, normalised so the year averages 1.0. The estimate pages multiply the central figure by each month's index.

NicheJanFebMarAprMayJunJulAugSepOctNovDec
Technology0.800.860.950.930.950.930.900.971.021.081.321.29
Finance & Business1.051.001.021.000.950.920.880.900.971.031.121.16
Health & Wellness1.181.000.970.940.970.930.880.900.981.021.101.13
Food & Recipes0.820.840.900.950.940.900.860.880.971.101.381.46
Travel0.921.001.101.081.121.101.020.950.940.950.920.90
Gaming0.800.820.900.880.900.940.950.961.001.081.351.42
Lifestyle0.800.880.960.950.980.950.900.930.981.061.271.34
News & Entertainment0.900.920.980.970.980.960.940.951.001.061.151.19
Education0.950.961.000.980.920.820.801.081.181.081.101.13
General0.780.850.950.930.970.960.920.951.001.081.251.36

Country estimates

Country pages place each market inside its tier with a market multiplier, and blend Android and iOS using a rough iOS share for that country.

CountryRegion tierMarket multiplierAssumed iOS share
United StatesTier 11.25x55%
United KingdomTier 11.00x50%
CanadaTier 10.90x50%
AustraliaTier 10.95x55%
GermanyTier 21.15x35%
FranceTier 20.95x30%
SpainTier 20.75x20%
ItalyTier 20.80x25%
JapanTier 21.30x65%
South KoreaTier 21.10x25%
BrazilTier 31.25x15%
MexicoTier 31.10x20%
IndiaTier 30.55x5%
IndonesiaTier 30.60x10%
TurkeyTier 30.70x15%
VietnamTier 30.50x20%
PhilippinesTier 30.60x10%
PakistanTier 30.40x5%
NigeriaTier 30.45x5%
EgyptTier 30.45x10%
NetherlandsTier 21.05x35%
SwedenTier 21.10x45%
SwitzerlandTier 21.35x55%
NorwayTier 21.25x50%
DenmarkTier 21.15x50%
BelgiumTier 20.90x35%
AustriaTier 21.00x35%
IrelandTier 21.05x50%
New ZealandTier 21.00x45%
SingaporeTier 21.20x50%
United Arab EmiratesTier 20.85x45%
Saudi ArabiaTier 31.50x30%
PolandTier 30.95x10%
ThailandTier 30.70x25%
MalaysiaTier 30.75x20%
South AfricaTier 30.80x15%
ArgentinaTier 30.65x10%
ColombiaTier 30.72x12%
BangladeshTier 30.35x3%
KenyaTier 30.40x4%

Quality checks on generated pages

Comparison, estimate, country, audience-size, format and answer pages are generated from the tables above. Before any of them can be indexed by search engines it must pass three automatic checks at build time:

  • at least 350 words written for that page (editorial notes, data-driven sentences and answers, not shared boilerplate);
  • less than 60% overlap with the most similar page of the same type, measured on five-word sequences;
  • at least five figures that differ from that most similar page.

A page that fails any check is still available to visitors but is marked noindex and left out of the sitemaps until it is improved.

Translations

Pages in other languages are machine-assisted translations checked by back-translation: the translated text is translated back into English and compared with the source, and any change in meaning is corrected. Numbers are never translated by hand; they come from the same engine and are formatted for each language. A language is opened to search engines only once its top pages, the home page and the calculators, have been reviewed by a native speaker. Until then every page in that language is marked noindex.

Data sources

The ranges were sanity-checked against public documentation and reporting, then rounded down where sources disagreed:

Report an error

If a figure looks wrong for your market, tell us through the contact page. Corrections that come with a source get priority.