EdTech app metrics
The ten metrics that tell you whether a learning app is working, how each is calculated, and how they connect from the first ad to a paying learner.
The key metrics for an EdTech app are cost per install and per sign-up, activation (first lesson started), time to first value, learning streaks, course completion, D7 and D30 retention, free-to-paid conversion, LTV and payback. Read them as one chain: each step multiplies the cost of the next.
What are the key metrics for an EdTech app?
The key metrics for an EdTech app follow the learner from ad to renewal: what a learner costs to acquire, whether they start learning, whether they keep learning, and whether they pay enough, for long enough, to cover that cost.
- Acquisition: cost per install and cost per action, where the action is a sign-up or trial.
- Activation: activation rate on a first lesson started, and time to first value.
- Engagement: learning streak rate and course completion rate.
- Retention: D7 and D30 retention and churn for subscriptions.
- Monetisation: subscription conversion, LTV and payback period.
Learner: a user who has started at least one lesson. Counting learners rather than sign-ups keeps every rate below anchored to real learning.
How do you measure EdTech success?
Measure EdTech success with one north star that reflects learning (for example, weekly learners who complete at least one lesson) and a short chain of funnel metrics underneath it, each with a fixed definition and time window. The tree below shows the chain; every tile links to the formula and a calculator.
- AcquisitionWhat a learner costs to bring in
- ActivationDid they start a first lesson, and how fast?
- EngagementAre they learning, and finishing?
- RetentionAre they still here at day 7, day 30 and renewal?
- MonetisationDo they pay enough, for long enough?
Metric tree: a diagram that breaks one outcome into the metrics that feed it, so a change at the top can be traced to the step that caused it.
How the EdTech metrics connect
Every step in the tree divides the users who reach the next one, so costs multiply as you go down the funnel. That is why a cheap install can still produce an expensive paying learner.
- Cost per activated learner = CPI ÷ activation rate.
- Cost per trial = cost per activated learner ÷ the share of activated learners who start a trial.
- Cost per paying learner = cost per trial ÷ trial-to-paid rate. This is the media part of CAC.
- LTV to CAC = learner LTV ÷ cost per paying learner. On gross margin, it needs to stay above 1 once payback time is counted.
The calculator runs this chain with example inputs. Change any step to see which one moves the cost of a paying learner most; in most funnels it is the earliest step with the lowest rate.
Example inputs: change them to your own numbers. The examples only make the calculator show a worked answer; they are not benchmarks or targets. Runs in your browser; nothing you type is sent anywhere.
- Cost per activated learner
- $10
- Cost per trial
- $50
- Cost per paying learner (media CAC)
- $125
- LTV to media CAC
- 1.2x
What to watch in each EdTech metric
- Activation on the first lesson, not the sign-up. Sign-up is easy to optimise for and often loosely linked to learning. A first lesson started is a stronger activation event and a better event to send back to ad platforms once volumes allow.
- Time to first value in minutes. Placement tests, long goal questionnaires and paywalls before any content all push time to first value up. Report the median and the share of new users who finish a lesson on day 0.
- Streaks that mean learning. A streak kept alive by one-tap visits is a habit without learning. Count only days with a completed lesson, and say whether streak freezes are included.
- Completion on a deadline. Compare course completion within a fixed number of days from starting, otherwise recent cohorts always look worse.
- Seasonal retention. Test-prep and school apps see retention and churn move with exam dates. Compare a cohort with the same cohort last year, not with last month.
- LTV with a horizon. Learners renew, upgrade and buy the next course, so LTV depends heavily on the horizon. Quote D90 or D365 cohort LTV and say whether it is revenue or margin.
Where EdTech numbers go wrong
- Blended funnels. Learners from creator campaigns, school partnerships and paid social behave very differently. Blended activation and retention hide the channel that is dragging the average down, so split each rate by source.
- Trials counted too early. Trial-to-paid measured while trials are still running understates conversion. Wait until the whole cohort's trials have ended.
- Benchmarks used as targets. Published figures help with orientation, not goals. For example, the median EdTech app in the State of Mobile App Growth: India Startups 2026 report turned about 29% of installs into sign-ups, with a wide spread between apps; AttributionBenchmarks' EdTech page shows the range. Your own trend matters more than the median.
Further reading
Linkrunner's Metrics that Matter: EdTech edition looks at which of these metrics predict course completion, and its guide to the events every EdTech app should track covers the event set you need before any of the formulas here can be calculated.
Measure the EdTech tree by channel
Questions
What is the north star metric for an EdTech app?
A common choice is weekly learners who complete at least one lesson, because it combines reach and real learning. Course platforms sometimes use weekly course progress instead.
Which EdTech metric should I fix first?
Start with the earliest step that has a low rate, usually activation or time to first value. Every later metric is calculated on the users who get past it, so early gains carry through the whole tree.
How do you calculate the cost of a paying learner?
Divide CPI by each conversion rate in turn: activation, trial start and trial-to-paid. At a $3 CPI with 30%, 20% and 40% conversion, a paying learner costs $125 in media.
Is course completion a retention metric?
It is closer to an engagement outcome. Retention counts learners who come back on a given day; completion counts learners who finish what they started, which is why the two are tracked separately.
Are the calculator's example inputs benchmarks?
No. The example inputs only make the calculator show a worked answer when the page loads; they are not typical, good or target values. Replace them with your own numbers.
Does the calculator send my numbers anywhere?
No. It runs entirely in your browser, and nothing you type is stored or sent anywhere.
More verticals
- Fintech app metricsThe metrics that show whether a payments, lending, investing or savings app turns installs into funded, active customers, and how each one is calculated.
- OTT and streaming app metricsThe metrics that show whether a streaming app turns installs into viewers and viewers into subscribers who stay, with the formula behind each.
- Mobile game metricsThe metrics game teams use to judge a game and its traffic: from what an install costs, through progression and retention, to daily revenue and ROAS.