Ad exchange
Programmatic marketplaceA marketplace that connects supply and demand platforms and runs the auctions between them.
Explore the terminology, technologies, strategies and metrics shaping mobile growth, programmatic advertising and performance marketing, explained with formulas, examples and calculators.
A marketplace that connects supply and demand platforms and runs the auctions between them.
Revenue in a period divided by active users in that period.
Crediting a conversion (an install, sign-up or purchase) to the marketing touchpoint that caused it.
Dividing users into groups by behaviour or value so each can be targeted differently.
The message an exchange sends to bidders describing an available impression: app, placement, device and context.
Lowering a bid in a first-price auction toward the minimum needed to win, so you do not overpay.
The total cost to acquire one paying customer, including media and other acquisition costs.
Malware on a device fires a fake click just before an install completes, to steal credit for it.
Sending huge numbers of fake clicks so a fraudster wins credit for installs that would have happened anyway.
Grouping users by a shared start date to compare how their behaviour and value develop over time.
Targeting ads by the content and app environment rather than personal identifiers. Try it with Context Compass.
The share of users who complete a target action out of those who could, such as installs that become paid trials.
Recording the actions that matter, like registrations, trials and purchases, and sharing them with your MMP and ad partners.
Ad spend divided by completed actions such as a sign-up, trial or purchase. Formula: CPA = Spend ÷ Actions.
Ad spend divided by the number of clicks. Useful for comparing creative pull, not user quality.
Ad spend divided by the number of installs it generated. Formula: CPI = Spend ÷ Installs.
The cost of one thousand ad impressions. Formula: CPM = Spend ÷ Impressions × 1,000.
The time between an ad click and the first app open. Very short times point to click injection; a long, flat tail points to click spamming.
Software advertisers use to buy impressions automatically across many exchanges, bidding on each one in real time.
Rules and models that identify and reject fraudulent clicks, installs and events, before or after attribution.
Users rewarded to install or act. Intent is low, so many advertisers exclude it from billing.
Installs from device farms, emulators or bots that never become real users.
Any ad activity that does not come from a genuine human with real interest, from simple bots to sophisticated fraud.
A model that gives all the credit to the final ad click before the conversion. It is the default for most mobile install measurement.
Finding new users who resemble your best existing customers.
The maximum time between an ad click or view and the install for that touchpoint to receive credit.
The total revenue a user is expected to generate over their relationship with the app.
Algorithms that continually adjust bids, budgets and sources toward a KPI as new data arrives.
A third-party platform such as AppsFlyer, Adjust, Branch or Singular that attributes installs and in-app events to the ad source that drove them, giving every partner one source of truth.
Installs not attributed to any paid source, such as app store search or word of mouth.
Installs attributed to a paid campaign through your MMP.
An invitation-only auction that gives selected buyers access to premium inventory, often with agreed floor prices.
A server-to-server message an MMP sends to an ad partner to report an install or in-app event, so bidding can learn from it.
Forecasting outcomes such as LTV or churn from a user’s early behaviour.
Using models to predict each impression’s chance of converting, and its value, then bidding accordingly.
Buying media through automated systems and data, rather than manual insertion orders.
Bringing inactive users back into active use, measured by re-attribution in your MMP.
Showing ads to existing or lapsed users to bring them back to a specific action, such as a first deposit.
The share of users who return to the app on a given day after install, such as day 1, 7 or 30.
Revenue generated for every unit of ad spend. Formula: ROAS = Revenue ÷ Ad spend.
Profit relative to total investment. Formula: ROI = (Revenue − Cost) ÷ Cost.
An auction run in milliseconds for each individual impression as an app or page loads.
Apple’s privacy-preserving attribution framework for iOS. It reports campaign-level conversions without user identifiers; Apple’s AdAttributionKit builds on it.
Software publishers use to sell inventory to many buyers at once and maximise the yield of every impression.
Paid and owned activity that brings new users into an app, ideally optimized to valuable actions rather than installs.
Credits a conversion to an ad that was seen but not clicked, within a short lookback window, to capture the influence of video and display.
Our original deep dives, now split into individual guides.
See these metrics at work in real campaigns, or talk to the team that runs them.