A/B Testing
Running two or more versions of a screen, creative or setting at the same time to see which performs better on a chosen metric.
Results only hold if the traffic split is random and the test runs long enough to be meaningful.
See also ,
Clear definitions of advertising, monetization, and user growth terms.
83 terms
Running two or more versions of a screen, creative or setting at the same time to see which performs better on a chosen metric.
Results only hold if the traffic split is random and the test runs long enough to be meaningful.
See also ,
A placement positioned in the part of a screen a user sees without scrolling.
Above the fold does not automatically mean viewable. Viewability depends on how long the ad is actually in view.
See also ,
A marketplace where publisher inventory is auctioned to many buyers in real time.
See also ,
Traffic or events created to claim ad revenue or attribution credit that was never earned, such as bots, device farms or click injection.
See also ,
The advertising space a publisher has to sell, counted as the impressions its placements can deliver over a period.
See also ,
A layer that decides which demand source gets each ad request, so a publisher can run several networks without calling each one in app code.
See also , ,
A company that buys inventory from publishers and sells it to advertisers, usually with its own demand and its own SDK.
Pricing models vary by network. Not every network sells on a fixed price.
See also ,
A call from an app or page asking for an ad to fill a placement.
See also ,
The configured slot that holds an ad, defined by its format, size and where it sits in the app.
See also ,
A full screen ad shown while an app is launching or returning to the foreground.
See also ,
Deciding which marketing touch gets credit for an install or an in-app event, using rules agreed in advance.
See also , ,
Revenue for a day divided by the number of daily active users that day.
ARPDAU = revenue for the day / daily active users for the same day
See also ,
Revenue over a period divided by the number of active users in that same period.
ARPU = revenue in the period / active users in the same period
State which user base is in the denominator, since monthly and daily bases give very different numbers.
See also ,
A rectangular ad anchored to an edge of the screen while the app stays in use.
Banners come in several sizes. The 320x50 unit is common but it is not the only one.
See also ,
The message a supply platform sends to buyers describing an available impression so they can bid on it.
See also ,
A buyer reply to a bid request, carrying a price and the creative to serve if the bid wins.
See also ,
Controls that keep ads away from content or apps an advertiser has decided not to appear next to.
See also ,
The share of users active at the start of a period who are no longer active at the end of it.
Churn rate = users lost in the period / users active at the start of the period
See also ,
Clicks divided by impressions for the same placement and period.
CTR = clicks / impressions
See also ,
A tool that collects and stores a user choice about data processing and passes that choice on to partners.
A consent tool supports compliance work. It does not by itself make a service legally compliant.
See also ,
A buying model where the advertiser pays for a defined action, such as a registration or a purchase.
CPA = spend / qualifying actions
See also ,
A buying model where the advertiser pays for each click.
CPC = spend / clicks
See also ,
A buying model where the advertiser pays for each attributed install.
CPI = spend / attributed installs
See also ,
The price for one thousand impressions.
CPM = (spend / impressions) x 1000
See also ,
The ad itself: the image, video, playable or native assets a buyer serves into a placement.
See also ,
The number of unique users who opened the app on a given day, counted by the publisher own definition of an active user.
See also ,
An identifier that links a buyer to agreed inventory and pricing inside a programmatic deal.
See also ,
The buying platform advertisers use to bid on inventory, set targeting and manage campaigns.
See also ,
Inventory sold on agreed terms between a publisher and a buyer, rather than in an open auction.
See also ,
Revenue earned per thousand impressions, used to compare demand sources whatever they are paid on.
eCPM = (revenue / impressions) x 1000
See also ,
The share of ad requests that were answered with an ad.
Fill rate = ads returned / ad requests
A returned ad is not the same as a shown ad. Compare fill rate with show rate.
See also ,
An auction where the winning buyer pays the price it bid.
See also ,
The lowest price a publisher will accept for an impression, set per placement, format or buyer.
See also ,
A limit on how many times one user can be shown the same ad or format in a period.
See also ,
European Union law governing how personal data is collected, used and shared, including for advertising.
See also ,
A resettable advertising identifier on Android, available when the user has not opted out of personalised ads.
See also ,
Running an auction among several demand partners before calling the primary ad server, so they compete on price.
See also ,
The advertising identifier on Apple devices, available only when the user has given app tracking permission.
See also ,
One counted delivery of an ad into a placement, measured by the rules of the counting party.
See also ,
Monetizing an app by selling advertising space inside it.
See also ,
Letting demand partners bid on each impression in real time inside the app, instead of calling them in a fixed order.
See also ,
Revenue a user pays directly inside the app, such as a subscription or an item.
See also ,
Ads that ask for an action from the user, such as a tap, swipe or short playable step.
See also ,
A full screen ad shown at a natural break, closed by the user.
See also ,
Ad requests, impressions or clicks that do not come from genuine user interest, whether from bots or from misconfiguration.
See also ,
The metric a team has agreed to judge a campaign or a placement on.
See also ,
The delay between a request and its response, which in advertising affects fill and user experience.
See also ,
The revenue expected from a user over the whole time they stay with the app.
LTV is a projection. Say which period and which revenue types it covers.
See also ,
In TapMind, the demand side of the platform: programmatic, direct, partner and brand budgets that publishers can reach.
See also ,
Calling demand sources one after another in a set order until one fills the request.
See also ,
A deeper rectangular ad placed inside the content, with more creative area than a banner.
See also ,
A TapMind layer added beside an existing mediation setup, used to connect and test additional demand without replacing that setup.
See also ,
An independent provider that records installs and in-app events and applies agreed attribution rules across sources.
See also ,
Unique users who opened the app at least once in a calendar month.
See also ,
An ad built from the same parts as the surrounding content, with its advertising disclosure kept visible.
See also ,
Revenue after agreed deductions such as platform fees and revenue share, which is what actually reaches the publisher.
See also ,
A screen where users choose from a list of advertised actions in exchange for in-app currency or a benefit.
See also ,
The IAB Tech Lab specification that standardises the messages used in real-time bidding.
See also ,
A user choice to stop personalised advertising or a specific type of data processing.
See also ,
A specific position in the app where an ad can appear, defined by screen, format and trigger.
See also ,
A short interactive ad that lets the user try part of an experience before installing.
See also ,
A server message confirming that an install or in-app event happened, sent to the partner that needs it.
See also ,
An invitation-only auction where selected buyers bid on agreed inventory.
See also ,
Buying and selling ad inventory through automated systems and auctions rather than manual orders.
See also ,
An auction run for each impression while it is being requested, decided in milliseconds.
See also ,
The share of a cohort still active a set number of days after install.
Day 7 retention = users from the cohort active on day 7 / users in the cohort
Always state the day and the cohort, since day 1 and day 30 retention are not comparable.
See also ,
Revenue attributed to a campaign divided by what was spent on it, measured over a stated window.
ROAS = attributed revenue in the window / ad spend in the same window
See also ,
A split of gross advertising revenue between parties, agreed as a percentage.
Revenue share is not profit share. It is taken from revenue, before the costs each side carries.
See also ,
A video the user chooses to watch in exchange for something in the app, with the reward given on completion.
See also ,
One continuous period of app use, ended by a timeout the publisher defines.
See also ,
The share of returned ads that were actually shown to a user.
Show rate = ads shown / ads returned
See also ,
Apple framework that reports campaign performance without device level identifiers, using limited and delayed postbacks.
See also ,
A package a developer adds to an app to use another company services, such as serving ads or reporting events.
See also ,
The platform publishers use to sell inventory, connect demand and manage pricing rules.
See also ,
Choosing which users, contexts or inventory a campaign should reach, within the rules of the market it runs in.
See also ,
The time a system waits for a response before moving on, used to keep ad loading from slowing the app.
See also ,
Deciding what share of ad requests a demand partner receives, used to test a partner or scale one back.
See also ,
Running paid campaigns to bring new users to an app, measured against install or post-install outcomes.
See also ,
An IAB Tech Lab specification that tells a video player which ad to fetch and how to report on it.
See also ,
The share of video ad starts that ran to the end.
VCR = completed views / video starts
See also ,
Whether an ad met an agreed standard for how much of it was in view and for how long.
See also ,
See Mediation Waterfall: demand sources called in a fixed order until one fills the request.
See also ,
The share of a buyer bids that won the auction.
Win rate = auctions won / bids submitted
See also ,
Here is what a publisher told us after their first quarter with TapMind.
We kept the mediation setup we already trusted and still opened up demand we could not reach before.
10+ years · Apps · Games · OTT · Web
The TapMind team will review your setup and reply within two working days.