--- title: Understanding how mobile app conversion tracking works description: Learn how mobile app conversion tracking actually works, why install numbers mislead, and how to audit your setup for misattribution and data corruption. image: https://weareaffective.com/hubfs/learning-centre-images/understanding-how-mobile-app-conversion-tracking-works.webp --- [Skip to content](https://weareaffective.com/learning-centre/understanding-how-mobile-app-conversion-tracking-works#main-content) [![we\_are\_affective\_logo\_200](https://weareaffective.com/hs-fs/hubfs/we_are_affective_logo_200.png?width=175&height=48&name=we_are_affective_logo_200.png "we_are_affective_logo_200")](https://weareaffective.com) - [Home](https://weareaffective.com) - About Us - [Our Story](https://weareaffective.com/about) - [How We Work](https://weareaffective.com/how-we-work) - Our Services - [App Planning & Strategy](https://weareaffective.com/app-planning-strategy) - [App Design](https://weareaffective.com/app-design-agency) - [App UX Design](https://weareaffective.com/app-ux-design) - [App UI Design](https://weareaffective.com/app-ui-design) - [App Technical Architecture](https://weareaffective.com/app-architecture) - [Existing App Audits](https://weareaffective.com/app-audit) - [Case Studies](https://weareaffective.com/case-studies) - [Pricing](https://weareaffective.com/pricing) - [Learning Centre](https://weareaffective.com/learning-centre) - [Get Started](https://weareaffective.com/get-started) Expert Guide Series # Understanding how mobile app conversion tracking works Table of Contents Conversion tracking sounds like a solved problem. You install an SDK, fire events, watch the numbers come in, and make decisions based on what you see. The data is right there. The dashboard is green. Downloads are climbing. So why do apps with growing acquisition numbers quietly bleed users and revenue at the same time? > The data is right there on the dashboard, but the question worth asking is whether it measures what actually matters. The answer is almost always in what the tracking is actually measuring, and whether the team understands the gap between the two. A conversion is not just an event that completed. A download is not a user. An install is not engagement. And a session that lasted four minutes is not evidence that anything worked. The numbers that look like progress can be, and often are, pointing in exactly the wrong direction. What follows is an attempt to describe how mobile app conversion tracking actually works, where the common misreadings happen, and what a more honest setup looks like. This draws on real work we have done across consumer apps in fitness, wellness, and social products, and on the patterns that keep appearing regardless of category or budget. ## What conversion tracking in mobile apps actually measures At a basic level, conversion tracking in mobile apps records when a user completes a defined action. That action could be installing the app, completing registration, finishing an onboarding flow, making a first purchase, or reaching a particular feature. The tracking system logs that the event occurred, ties it to a source where possible, and surfaces the data in a reporting layer. The technical machinery behind this involves an SDK embedded in the app, which sends event signals to an attribution platform. That platform then attempts to match each event to a traffic source, a campaign, or a specific ad. On Android, this process is relatively straightforward. On iOS, Apple's App Tracking Transparency framework and SKAdNetwork have made individual-level attribution much harder, replacing it with aggregated, delayed, and heavily constrained reporting. SKAdNetwork provides only 64 possible conversion values to represent all post-install events, which is a precise technical constraint rather than an interpretation, and it delivers that data with a delay of 24 to 72 hours rather than in real time, according to [Cometly](https://www.cometly.com/post/conversion-tracking-for-app-installs). What this means in practice is that even a technically correct tracking setup gives you a partial picture. Attribution data describes what happened at the point of acquisition. It says much less about what followed, and almost nothing about why. ## Where the misunderstanding begins: events, goals, and what gets labelled a conversion The word "conversion" gets applied to almost every event a user can complete, and that breadth is most of the problem. Teams label an install a conversion, then a registration, then a first session, sometimes all three in the same dashboard view. Each of those events is real. None of them is the same thing. An event is something that happened. A goal is a state the user reached that means something for the business. A conversion, properly defined, is the moment a user crosses from prospect to someone who has experienced real value. By that definition, most events labelled as conversions in a standard setup are merely milestones on the way to one. The confusion compounds because analytics platforms make it easy to mark any event as a goal, and this is exactly the kind of structural issue that surfaces clearly during an [App Experience Audit](https://weareaffective.com/app-audit). The technical capability to track something creates the assumption that it is worth tracking as a primary metric. So teams end up optimising toward events that are easy to measure and easy to move, rather than toward the outcome that actually tells them the product is working. A fitness app that counts "profile created" as its primary conversion metric has no idea whether users ever opened a workout. A supplement brand counting "checkout started" has no idea whether purchases completed. The event fired. The conversion, in any meaningful sense, may not have happened. ## UX/UI design built around *real* psychology We design app interfaces around how people actually think and behave. User research, psychology-driven UX/UI design and technical specs delivered as one complete package. [See how we work](https://weareaffective.com/how-we-work) [Get started](https://weareaffective.com/get-started) No commitment ## The setup decisions that corrupt your data from day one Most tracking problems are setup decisions that seemed reasonable at the time and quietly distort everything that follows. The most common is choosing the wrong primary conversion event. If a team sets "app install" as its optimisation signal, paid channels will find users who install apps, and that is not the same as finding users who stay. The algorithm optimises toward what it is told to value, and if install is the goal, install is what you get, regardless of what those users do next. > The algorithm finds exactly what you tell it to value, so optimising toward installs gets you users who install apps, not users who stay. A second common problem is event duplication, where the same action fires more than once because the trigger logic was not tested thoroughly. A purchase event that fires on both the confirmation screen and the thank-you redirect inflates revenue data and makes channel attribution look stronger than it is. A third is inconsistent event naming across platforms. If iOS calls an event "purchase\_complete" and Android calls it "order\_finished", every cross-platform report is comparing different things without knowing it. The numbers add up. The insight does not. Before you run any acquisition spend, audit every primary conversion event end-to-end in a test environment. Confirm the event fires once, at the right moment, with the right parameters attached. A corrupted signal from day one will shape every decision that follows. ## Why download and install numbers are the wrong place to start On average, around 77 per cent of apps lose their daily active users within the first three days of download. Even well-built products with clear value propositions typically see a 40 to 50 per cent retention drop over the same period. The gap between those two figures, 77 per cent against 40 to 50 per cent, is where most of the real work happens, and it is a gap that download metrics cannot see. We have seen this play out directly in our own work. On the water tracking app we developed, we were initially reading user silence as approval. Complaints were low. Support tickets were sparse. Downloads were steady. But when we looked at retention data, we saw a drop-off pattern that matched what we later found in published figures for personal finance apps, where approximately 71 per cent of daily active users are lost between day one and day thirty. Download numbers look like growth. In some cases they are. But a download is [an expression of curiosity, not commitment](https://weareaffective.com/learning-centre/what-curiosity-looks-like-in-a-first-session-and-why-most-products-design-past-i). A user who installed the app and opened it once, then deleted it three days later, counts the same as a user who has logged in daily for six months. Both are a single data point in the install column. The metric that actually describes product health is retention at day three, day seven, and day thirty. If those numbers are not in the room when acquisition decisions are made, the team is flying without most of the instruments. ## The retention gap that reveals what acquisition metrics hide About 25 per cent of mobile apps are abandoned after a single use. That is a high number, and the reason it stays high is that teams rarely see it happening in time to respond. A user who opens the app once and never returns generates no complaint, no churn event in most tracking setups, and no signal of any kind. They simply go quiet. This is the retention gap. Acquisition metrics tell you that someone arrived. Retention metrics tell you that they stayed. The space between those two data points is where most of the real story lives, and standard conversion tracking ignores it almost entirely. The gap matters because it reveals something about channel quality that cost-per-install figures cannot. A campaign that delivers 10,000 installs at a low CPI might be generating users with a day-one retention rate of 15 per cent. A smaller campaign with a higher CPI might be generating users who are still active at day seven. The second campaign is more valuable by any honest measure. But if retention data is not stitched into the attribution view, the first campaign looks like the winner and keeps getting budget. Segment your retention cohorts by acquisition channel from the start. A channel that looks expensive on CPI might be delivering users who actually stay. One that looks cheap might be filling your funnel with users who leave before they see your core product. ## How onboarding friction distorts conversion data before users reach your core product Most conversion tracking setups measure what happens after onboarding, because that is where the events the team cares about live. But onboarding is where a significant proportion of users leave, and their departure is frequently misread as a product problem rather than a flow problem. Forcing registration early in the app experience produces a drop-off rate of around 15 to 20 per cent in uninstalls at that point, typically within the first 60 to 120 seconds of use. That means a meaningful share of the users a brand paid to acquire never reach the product at all. They encounter a gate, decide the exchange is not worth it, and leave. That exit registers in the data as low conversion on the registration step, which is accurate, but the reason is not visible without knowing where in the flow it occurred and what the user was asked to give up. Onboarding friction compounds because each additional step has its own drop-off rate, and those rates multiply. A four-step onboarding flow where each step loses 10 per cent of users delivers only 65 per cent of those who started it to the first real product moment. If the conversion tracking starts after that point, the team is analysing the behaviour of users who were already filtered by friction, and drawing conclusions that do not apply to the whole audience. #### What onboarding data should capture The tracking setup needs to log each individual step in the onboarding sequence as a distinct event, with the time elapsed between steps. That gives you both a completion rate and a sense of where users are pausing or abandoning under pressure rather than by choice. ## Trust-sequencing and why premature data requests register as drop-off, not hesitation When a user abandons an app at the point of a permissions request or a data form, the tracking system records a drop-off. That is technically correct. What it does not record is that the user was willing to continue, but not on those terms, at that moment. The exit was a response to timing, not a rejection of the product. We addressed this directly on a fitness social network project we worked on. The product needed users to share location data to connect them with potential running partners nearby. Asking for precise location early in the experience was causing significant drop-off. The solution we built was to replace precise location disclosure with an approximate proximity indicator, showing users that a potential running partner was within a 200 to 500 metre radius without revealing exact whereabouts. Users could then view a profile and message the other person through the app's built-in chat before choosing to share their precise location and arrange a meet-up. The sequencing of that trust exchange changed the conversion rate on the location permission step. But in a standard tracking setup, none of that would be visible. The old setup would show drop-off at the permission request. The new one would show conversion at the same step. The reason for the change, the trust-sequencing decision, lives entirely outside the tracking data. Where your app needs a high-stakes data request, such as location, camera access, or payment details, [map the trust journey that leads to it](https://weareaffective.com/learning-centre/how-do-i-build-trust-with-users-when-theyre-making-such-big-decisions). [The permission request should come after the user has experienced enough value](https://weareaffective.com/learning-centre/what-makes-users-trust-a-product-enough-to-enter-their-card-details) to understand why sharing makes sense for them, not at the point when it is convenient for you. ## The difference between a completed event and a meaningful conversion An event completing and a user converting are two different things, and confusing them produces reporting that looks healthy while the product is struggling. Consider a wellness app that tracks "meditation session started" as a conversion event. The event fires reliably. The numbers are encouraging. But if users start sessions and abandon them at the two-minute mark, or if they complete one session and never return, the event is not measuring what the team thinks it is measuring. The conversion happened technically. The value exchange did not. A meaningful conversion is one where the user reached a state that predicts future behaviour. In a subscription app, that is usually the moment a user first experiences the core value the subscription is built around. In a social app, it might be the moment they make a genuine connection. In a fitness app, it could be completing a full session and logging a result. These moments are harder to define, and harder to track, than a button tap or a screen load. But they are the ones that actually predict whether a user will still be there in thirty days. #### Defining your true conversion moment The exercise worth doing is to look at the [users who are still active at day thirty](https://weareaffective.com/learning-centre/why-the-first-thirty-days-decide-the-next-twelve-months-of-retention-in-beauty-a) and trace back to the earliest action they had in common. That shared early behaviour is almost always a better primary conversion event than whatever the tracking setup was built with on day one. ## What granular behavioural data can tell you that funnel metrics cannot Funnel metrics describe the shape of a journey. Behavioural data describes what happened inside it. Both matter, but they answer different questions, and treating funnel data as the complete picture leaves out most of what is useful. Standard engagement figures, session length, daily active users, monthly active users, do not tell you whether users are finding value or just staying for other reasons. A long session might mean deep engagement with a product that is working. It might equally mean a user who is confused and cannot find what they are looking for, or a product that has been designed to retain users through patterns that have nothing to do with value. These are very different situations that look identical in a session length report. Granular behavioural data, things like which features are revisited, how long passes between sessions, what path a user takes before they reach a core action, and where they pause before completing a step, gives you the texture that funnel metrics flatten out. On the water tracking app we developed, it was this kind of behavioural data that revealed the retention problem the download figures were hiding. Users were installing and opening the app, but the patterns of when they returned, and when they stopped, showed us where the experience was failing them. The session-length figure alone would have told us nothing useful. It was the sequence and timing of behaviour within and between sessions that showed where the real problem was. ## How to audit an existing tracking setup for misattribution Misattribution in a mobile tracking setup compounds over time. A team that has been optimising toward a corrupted signal for six months is actively making decisions based on bad data that looks plausible. The audit is not optional once this is suspected. The process has a logical order. 1. Map every event currently firing against the outcome it is supposed to represent. Ask whether the event actually measures that outcome or just correlates with it loosely. 2. Check for event duplication by running a test session and counting how many times each event fires per completed action. One action should produce one event. 3. Confirm event naming consistency across iOS and Android. Every event that exists on one platform should have an identical name on the other. 4. Compare attributed installs against organic installs in your source-of-truth analytics, usually Firebase or a comparable platform. A large gap between what the attribution platform claims and what the analytics platform records is a sign of misattribution or double-counting. 5. Look at ATT opt-in rates. Industry averages sit around 25 to 30 per cent globally according to [Cometly](https://www.cometly.com/post/conversion-tracking-for-app-installs), which means the majority of iOS attribution data is modelled rather than observed. Build that limitation into how you read any iOS channel performance figure. 6. Cross-reference your top-performing acquisition channels against day-seven retention by cohort. If your cheapest install source also has the lowest retention, the CPI figure is misleading you about the channel's actual value. This audit will not produce a perfect picture. Mobile attribution on iOS in particular carries structural uncertainty that no setup can eliminate. But it will usually reveal which parts of the current data can be trusted and which cannot, and that is enough to start making better decisions. ## Conclusion Conversion tracking is not broken. The frameworks exist, the tools exist, and the data is genuinely there to be read. The problem is that the default setup measures what is easy to measure, and what is easy to measure is almost never the thing that predicts whether a product is working. Downloads count curious people. Installs count people who had a few seconds to spare. Sessions count moments of access. None of those things, on their own, tell you whether a user found value, will return, or will pay. The conversion that matters is the one where a user experienced enough of your product to understand why it exists. Everything before that is acquisition. Everything after it is retention. And the tracking setup that does not distinguish between the two will keep telling a flattering story until the business stops believing it. The work we have described here, from [re-sequencing trust flows on the fitness social network](https://weareaffective.com/learning-centre/what-a-fitness-social-app-changed-when-it-realised-precise-location-was-the-wron) to reading retention data on the water app rather than relying on silence as a proxy for approval, is the kind of work that changes what the numbers mean. Getting the tracking right does not just produce better reports. It produces better decisions, which produce better products, which produce users who actually stay. If your current setup is telling you a story you are not sure you believe, [let's talk about your tracking setup](https://weareaffective.com/get-started). ## Frequently Asked Questions What does conversion tracking in a mobile app actually measure? Conversion tracking records when a user completes a defined action, such as installing the app, registering, or making a purchase. The tracking system logs the event, attempts to tie it to a traffic source or campaign, and surfaces that data in a reporting layer. However, even a technically correct setup gives you only a partial picture of what happened after acquisition. Why is an app install not the same as a conversion? An install simply means the app was downloaded, not that the user experienced any real value. A true conversion is the moment a user crosses from prospect to someone who has genuinely engaged with what the product offers. Labelling an install as a conversion flatters the numbers without telling you anything meaningful about retention or revenue. What is the difference between an event, a goal, and a conversion? An event is something that happened, such as a tap or a screen view. A goal is a state the user reached that carries business meaning, and a conversion is the point at which a user has experienced real value. Most events labelled as conversions in a standard setup are actually just milestones on the way to a genuine conversion. How has Apple's App Tracking Transparency affected conversion tracking on iOS? Apple's App Tracking Transparency framework and SKAdNetwork have made individual-level attribution significantly harder on iOS. SKAdNetwork limits reporting to just 64 possible conversion values and delivers data with a delay of 24 to 72 hours rather than in real time. This means iOS attribution is aggregated and constrained in ways that Android tracking is not. Why can growing download numbers coincide with falling revenue? Download volume and revenue are measuring different things, and strong acquisition numbers can mask poor engagement or onboarding further down the funnel. If the events being tracked as conversions are early milestones rather than moments of genuine value, the data can appear healthy while users are quietly churning. The gap between what is being measured and what actually matters is where revenue tends to disappear. What is the risk of marking too many events as goals in an analytics platform? Analytics platforms make it technically easy to mark any event as a goal, which can create a dashboard full of activity that looks like progress but lacks strategic meaning. When every milestone is treated as a conversion, it becomes harder to identify which moments actually drive business outcomes. This is a structural problem that can mislead decisions about product investment and campaign spending. Does conversion tracking tell you why users behave the way they do? Attribution data can describe what happened at the point of acquisition, but it says very little about what followed and almost nothing about why. Knowing that a user dropped off after onboarding, for example, does not explain whether the experience was confusing, the value proposition was unclear, or something else went wrong. Conversion tracking gives you signals, not explanations. What does a more honest conversion tracking setup look like? A more honest setup starts with a clear definition of what a genuine conversion means for the specific product, rather than borrowing a generic definition from the analytics platform. It distinguishes between milestones and moments of real value, and it connects acquisition data to downstream behaviour rather than stopping at the install. Regular auditing of what is being tracked and why is also an important part of keeping the setup accurate over time. ## Related Articles [![We Are Affective](https://weareaffective.com/hubfs/we_are_affective_logo_mark.svg)](https://weareaffective.com) 20-22 Wenlock Road London, N1 7GU United Kingdom +44 20 4572 8062 [hello@weareaffective.com](mailto:hello@weareaffective.com) Services [App planning & strategy](https://weareaffective.com/app-planning-strategy) [App design](https://weareaffective.com/app-design-agency) [App UX design](https://weareaffective.com/app-ux-design) [App UI design](https://weareaffective.com/app-ui-design) [App technical architecture](https://weareaffective.com/app-architecture) [Existing app audits](https://weareaffective.com/app-audit) Legal [Privacy policy](https://app.termly.io/policy-viewer/policy.html?policyUUID=b8fa9921-7518-4fb5-8ddd-9dc7f5977ed2) [Terms](https://app.termly.io/policy-viewer/policy.html?policyUUID=8b6a6ad5-91bd-4176-a5f7-6d36b0398f70) Case studies [TravAI](https://weareaffective.com/case-studies/travai) [Meditech](https://weareaffective.com/case-studies/harley) [WorkingWeight](https://weareaffective.com/case-studies/workingweight) [SkinSync](https://weareaffective.com/case-studies/skinsync) [Three Lochs](https://weareaffective.com/case-studies/three-lochs) [Drift](https://weareaffective.com/case-studies/drift) About us [Our Story](https://weareaffective.com/about) [How We Work](https://weareaffective.com/how-we-work) Guides [Creating an app](https://weareaffective.com/how-to-create-an-app) [Building an MVP](https://weareaffective.com/building-an-mvp) [Cost and budgeting](https://weareaffective.com/app-development-cost) [App technology](https://weareaffective.com/app-development) [Planning and strategy](https://weareaffective.com/app-planning-strategy) [User research](https://weareaffective.com/app-user-research) [Onboarding design](https://weareaffective.com/app-onboarding-design) [User psychology](https://weareaffective.com/user-psychology-app-design) [Launch and growth](https://weareaffective.com/app-launch-growth) Copyright © 2026, weareaffective.com. All rights reserved.