EdTech Metrics That Matter

By Brian Tr · Updated Oct 7, 2026

An EdTech company has two scoreboards, and most teams only watch one. The first is the business scoreboard: sign-ups, revenue, churn, payback. The second is the learning scoreboard: did people start, did they keep going, did they finish, and did anything change afterward. Ignore the first and you run out of money. Ignore the second and you eventually lose customers who stop believing the product works.

This article is the umbrella map. It lists the metrics worth tracking at each stage of the learner and customer journey, defines each one in plain terms, and shows how the set changes depending on whether you sell to consumers, to companies, or as a software platform. Where a metric has its own deep-dive article in this library, we summarize it in a paragraph and link out.

Why EdTech Needs Two Sets of Metrics

In most software businesses, the user and the buyer are the same person, and usage is a decent proxy for value. In EdTech they often split. A company's HR director buys a course library, and 2,000 employees are supposed to use it. A parent pays for a tutoring app, and a child does the learning. Even in pure consumer courses, the person who paid in January may stop opening the app by February.

That split creates a specific failure mode. Revenue metrics look healthy while learning metrics decay, because annual contracts and prepaid courses keep cash coming in long after engagement has faded. The decay shows up later as a renewal that doesn't happen. So a good metric set links the two: each learning metric should have a revenue consequence you can name.

The structure below follows the journey from first visit to renewal.

Acquisition and Enrollment Metrics

Acquisition metrics tell you whether the top of the funnel is efficient. Enrollment metrics tell you whether the offer converts. The two are separate on purpose: a cheap visitor who never enrolls is not a cheap customer.

Customer acquisition cost (CAC). Total sales and marketing spend in a period divided by the number of new paying customers in that period. In EdTech, define "customer" carefully. For a B2C platform, it's a paying learner. For a corporate vendor, it's a signed company, which may represent hundreds of seats. Mixing the two produces numbers that look like a bargain and aren't. For a full treatment, see the library's customer acquisition cost entry.

Funnel conversion rates. Track each step separately: visit to sign-up, sign-up to first enrollment, enrollment to payment. A free-course funnel and a paid-bootcamp funnel have very different shapes, and blending them hides where people leave. The stages are mapped in the online learning enrollment funnel.

Registered versus paying learners. Large registered-learner counts are common headline numbers. Coursera, for example, reports approximately 197 million cumulative registered learners and counts a registration as a unique email, in its annual report for fiscal 2025. That's a useful brand-reach measure, but it isn't a revenue measure. Always pair a registration number with a paid number.

Activation: The Metric Teams Skip

Activation is the moment a signed-up learner does the first meaningful thing. The definition is yours to set, and it should be an action that predicts later behavior. Common choices:

• Finished the first lesson within seven days of enrolling
• Joined the first live session or submitted the first assignment
• For a B2B platform, a customer admin has invited a first group of learners and at least a set share of them have logged in

Activation matters because the biggest losses in EdTech happen early. A learner who never starts is invisible in a completion rate calculated after the course ends, but visible immediately in an activation rate. If you measure only completion, you learn about a problem months after you could have fixed it.

Engagement Metrics

Engagement metrics describe what learners do once they're in. The usual set includes active learners per week or month, sessions per learner, time spent, lessons or modules completed, and participation in discussions or live sessions.

Two cautions apply. First, time spent is easy to measure and easy to misread. A learner stuck on a confusing page logs a lot of minutes. A learner who skims a lesson they already know logs few. Time is a context metric, not a goal. Second, engagement should be defined relative to the intended pace of the course. A self-paced library has different healthy rhythms than a six-week cohort.

For definitions, benchmarks and the pitfalls of each measure, see learner engagement metrics. For a B2B vendor, the most useful engagement view is often the account-level one: what share of purchased seats are active, because unused seats are the clearest early predictor of a lost renewal.

Completion Metrics

Completion rate is the most quoted and most abused metric in the field. It's simple to compute: the number of learners who finished divided by the number who started or enrolled. The trouble is that the denominator changes the answer entirely.

Katy Jordan's study of MOOC enrollment and completion data found that the average MOOC enrolled around 43,000 students and about 6.5% of them completed (Jordan, IRRODL 2014). That one figure is a useful warning: a tiny rate against all enrollees says little until you know who was counted. Change the denominator from "enrolled" to "started" and the headline changes with it.

Practical rules for a clean completion metric:

1. Pick one denominator and publish it. "Of learners who started lesson one" is a common, defensible choice.
2. Measure by cohort. A blended rate across months hides whether last quarter's changes helped.
3. Track the drop-off point, not only the final rate. Knowing that most exits happen at module three tells you what to fix.
4. Separate mandatory from voluntary. Assigned compliance training and a paid-for-fun course live in different universes.

The full discussion, including what raises completion and what doesn't, is in course completion rates.

Learning Outcome Metrics

Completion tells you someone reached the end. It doesn't tell you they learned anything. Outcome metrics close that gap, and the most widely used scaffold for them is the Kirkpatrick Model.

Kirkpatrick Partners describes four levels (Kirkpatrick Partners):

It advises starting with Level 4 and working backward, so you define the result you want before designing the measurement. And it notes that Level 1 alone isn't enough to tell you whether learning happened, which is why a five-star rating shouldn't be your outcome metric.

For a consumer platform, Levels 1 and 2 are usually practical: ratings and assessment gain, plus longer-term indicators such as whether the learner earned a credential or got the job they wanted. For a corporate vendor, Levels 3 and 4 decide renewals, because the buyer has to show leadership a business result. Methods for that are in Kirkpatrick model and proving training ROI.

Revenue and Retention Metrics

This is the business scoreboard. Most of these metrics come from SaaS, and they apply to subscription EdTech with a few tweaks.

MRR and ARR. Monthly recurring revenue is the predictable subscription revenue you expect each month. Annual recurring revenue is that figure annualized. Coursera's filing describes calculating ARR by annualizing each customer's monthly recurring revenue for the most recent month at period end (Coursera 10-K). Don't count one-time course sales, implementation fees or services as recurring.

Churn. The share of customers (logo churn) or revenue (revenue churn) lost in a period. Track both for B2B. A company that cancels a ten-seat plan and one that cancels a ten-thousand-seat plan are the same logo and very different revenue.

Net revenue retention (NRR). Revenue today from the customers you had a year ago, divided by what they paid then. Coursera's filing describes it exactly that way for its Paid Enterprise Customers: current-period ARR from the same customers, excluding new customers, divided by prior-period ARR. The number was 93% for 2025, up from 87% in 2024 (Coursera 10-K). Udemy reported a Udemy Business net dollar retention rate of 93% at the end of 2025 and 97% for customers with at least 1,000 employees, and said churn outpaced expansion across existing customers (Udemy 10-K).

Read those two examples for what they teach, not for a benchmark. Both are public companies with their own definitions, and a figure below 100% means customers shrank or left faster than others expanded. In learning, that often traces back to unused seats. For the mechanics, see net revenue retention.

Lifetime value (LTV). The revenue, ideally gross profit, you expect from a customer over their relationship with you. EdTech LTV is tricky because learners often leave by design: a person who finishes a certification and moves on is a success, not a failure. For consumer products, the better lever is often a learning path that gives people a reason to come back.

CAC payback. Months of gross-margin revenue needed to recover what you spent to acquire a customer. It combines the acquisition and revenue sides into one number and is more practical than a long-horizon LTV for most early-stage companies.

How the Metric Set Changes by Segment

The same seven stages apply everywhere, but the weights differ. Use this as a starting point, not a rule.

A few segment notes. In B2C, the learner is both user and buyer, so engagement and revenue are tightly linked, and cohort curves tell you most of what you need. In B2B, one customer holds many learners, so you need two layers of every metric: account level and learner level. And for an LMS or other learning software, "completion" is partly a product metric (do your customers' courses get finished on your platform?) and partly a customer-success metric. Which of these numbers you can even collect depends on the learning tech stack and the tracking standards it supports, such as SCORM and xAPI.

Building a Metrics Dashboard That People Use

You don't need thirty metrics. Most teams do better with one headline metric per stage and a small set of supporting ones.

A workable starting set:

• Acquisition: CAC and visitor-to-sign-up rate
• Activation: Share of enrollees who hit the activation event within seven days
• Engagement: Weekly active learners, and for B2B, active seats as a share of purchased seats
• Completion: Cohort completion rate with a stated denominator
• Outcomes: One pre/post learning measure per course, plus one business indicator for enterprise customers
• Revenue: MRR/ARR, logo churn, NRR (B2B) and CAC payback

Then follow four habits:

1. Segment everything by cohort and channel. An average across all learners hides the difference between paid-search signups and referrals.
2. Write a definition for each metric and keep it fixed. Changing the definition of "active" mid-year makes every trend line meaningless.
3. Review leading indicators weekly and lagging ones monthly. Activation and engagement move first. Churn and NRR confirm later.
4. Tie each learning metric to a revenue effect. If you can't say what a drop in completion would do to renewals, the metric isn't earning its place.

Common Mistakes

Treating registrations as customers. Big registered-learner counts are brand-reach numbers. They aren't revenue and they aren't engagement.

Using completion as the only learning metric. Completion can be high for a course that teaches nothing, and low for a course that people use only for one module. Pair it with an outcome measure.

Key Facts

• EdTech needs a business scoreboard and a learning scoreboard, and each learning metric should have a named revenue consequence.
• Completion depends on the denominator: the average MOOC in Jordan's study enrolled about 43,000 students and about 6.5% completed (Jordan, 2014).
• Kirkpatrick Partners defines four evaluation levels (Reaction, Learning, Behavior, Results) and advises starting with Level 4 and working backward (Kirkpatrick Partners).
• Coursera's Net Retention Rate for Paid Enterprise Customers was 93% for 2025, up from 87% in 2024 (Coursera 10-K).
• Udemy Business net dollar retention was 93% at the end of 2025, and 97% for customers with at least 1,000 employees (Udemy 10-K).
• Coursera reports approximately 197 million cumulative registered learners, counted by unique email, which is a reach measure and not a paying-customer count (Coursera 10-K).