How to Measure Each Stage of Your Course Funnel
Sep 24, 2026
By Barbara Amador, Course Conversion Strategist · Published
You measure an online course funnel one stage at a time, with one number per stage: opt-in rate, early email engagement, list engagement, click-through to the sales page, sales page conversion, checkout completion, and refund rate. Each number answers a different question about a different part of the business. Enrollments stall at the first stage that drops below its own baseline, read in the order the business is built rather than the order a visitor moves through it. That first stall is where the revenue leak starts. It is often a stage or two earlier than the place where the problem feels loudest.
Collecting the numbers is not the hard part. Reading them in the right order is. Here is what to measure, how to calculate each number, and how to trace a stall back to the layer that caused it.
What does it mean to measure a course funnel?
Measuring a course funnel means counting how many people reach each step between arriving and staying a customer, then dividing each count by the one before it. The result is a chain of stage rates. Each rate tells you what share of people made it through that step.
A single number at the end, like total sales, cannot tell you where people were lost. Ten sales from a launch could mean a strong page and a small list, or a big list and a weak offer, or a warm list that never got the email. Stage rates separate those stories. They show you which step lost the people, so you are not guessing from the result.
There is one catch. A funnel is laid out in the order a visitor experiences it. A course business is built in a different order: niche, lead magnet, email list, offer, sales page, launch, traffic. Two of those layers, niche and offer, are not steps a visitor passes through. They do not get their own number. They show up inside the other numbers, and part of reading a funnel well is knowing where to look for them.
The stages of a course funnel and the number for each
| Stage and number | How to calculate it | What it tells you | Layer it points to |
|---|---|---|---|
| Opt-in rateOpt-in | New subscribers from the page ÷ unique visitors to the opt-in page | Whether the free resource is worth trading an email for | Lead magnet, niche |
| Early engagementNew subscriber | Unique clicks in the welcome sequence ÷ new subscribers in that period | Whether the right people signed up | Lead magnet, niche |
| Active shareList | Subscribers who clicked any email in the last 90 days ÷ total list | How much of the list is actually warm | Email list |
| Click-through to sales pageEmail to page | Unique people who clicked to the sales page ÷ subscribers emailed in the sales window | Whether email is creating enough interest to look | Email list, launch |
| Page-to-checkout rateSales page | Checkout starts ÷ unique sales page visitors, split by source | Whether people who arrive want what is offered | Offer, sales page |
| Checkout completionCheckout | Purchases ÷ checkout starts | Whether something at the point of payment is stopping people | Offer (price), sales page |
| Refund rate and early startAfter purchase | Refunds ÷ purchases; buyers who begin the course in the first week ÷ purchases | Whether the offer delivered what the page promised | Offer |
Two summary numbers sit on top of the chain. Earnings per subscriber reads the whole list at once. List conversion rate, buyers from a launch divided by the list size when the launch started, tells you what share of your audience said yes. Both are useful. Neither can tell you which stage to fix.
Opt-in rate
Opt-in rate measures one thing: of the people who saw the offer of a free resource, how many took it. Count unique visitors to the opt-in page and new subscribers who came through that page, over the same period.
A stall here shows up as a steady stream of visitors and very few sign-ups. It usually means the freebie is not specific enough to feel worth an email address, or it is aimed at a problem the visitor does not have yet. What makes a lead magnet actually convert covers the fix. Opt-in rate alone is only half the picture, though, which is why the next number exists.
Early engagement
Early engagement measures whether new subscribers act like future buyers. Take everyone who joined in a given month and count how many clicked at least one link in the welcome sequence.
This is the number that catches a problem opt-in rate hides. A freebie can pull a high opt-in rate from people who will never buy. They take the download and go quiet. When sign-ups look healthy but new subscribers do not click, the lead magnet is attracting the wrong crowd, and that often traces further down to a niche that is too broad. Use clicks here, not opens. Some email apps load images automatically, which records an open whether or not a person read anything.
Active share of the list
Active share measures how much of your list is warm right now. Count subscribers who clicked any email in the last 90 days and divide by the total list.
List size is the number people report. Active share is the number that predicts a launch. A list can double in size while its active share falls, because the new names came in cold and the old names drifted. A launch sent to that list will underperform, and the launch will get the blame. Why a cold list sinks a good launch covers what that looks like from the inside.
Click-through to the sales page
Click-through measures whether your sales emails earn a look. During the sales window, count unique people who clicked from an email to the sales page and divide by the number of subscribers who were emailed.
Tag every email link to the sales page so your analytics can tell email visitors apart from everyone else. That one step makes the next number readable. A stall at click-through means people are not curious enough to open the page. That is an email and list problem. It is not a page problem, because the page has not been seen yet.
Sales page conversion, split by source
Page-to-checkout rate measures whether people who arrive want what is being offered. Divide checkout starts by unique sales page visitors. Then do it again separately for visitors from email and visitors from everywhere else.
The split matters more than the total. Warm visitors from your list and cold visitors from social or ads behave very differently on the same page. A blended rate mixes the two, and a jump in cold visitors can drag it down while nothing about the page changed. What counts as a good sales page conversion rate depends almost entirely on that mix, which is why there is no benchmark worth borrowing.
When warm visitors who clearly understand the offer still do not start checkout, the stall is usually below the page, in the offer itself. Is it a sales page problem or an offer problem? has the test for telling them apart.
Checkout completion
Checkout completion measures what happens after someone decides to buy. Divide purchases by checkout starts.
A stall here is specific and worth finding early. People who reach checkout have already said yes once. If many of them stop, look at what they meet at that step: a price that was not clear on the page, a payment plan that was promised and is missing, a surprise fee, a broken form on mobile. Test the checkout yourself on a phone before assuming anything else.
Refund rate and early start
Refund rate measures whether the offer held up after the sale. Early start, the share of buyers who begin the course in their first week, measures whether they believed it enough to use it.
These are the numbers conversion work tends to skip, and they are where offer problems hide. A new page or a bigger bonus stack can lift sales and push refunds up at the same time. If you only watched the front of the funnel, that looks like a win. Read the reasons people give when they ask for a refund. They are some of the most honest offer feedback you will ever get.
Why the order you read the numbers matters
A visitor moves through the funnel from traffic to purchase. The business underneath is built the other way. The niche shapes who the lead magnet attracts. The lead magnet shapes who is on the list. The list shapes who sees the offer. Each layer inherits the health of the one before it.
That means a weak number late in the funnel is often a strong signal about something earlier. A low sales page rate can be the page. It can also be a cold list sending lukewarm visitors, or a lead magnet that filled the list with people who were never going to buy this course. The number shows up on the page. The cause lives somewhere else. The symptom is rarely the problem is the longer version of that idea.
So read the stage rates in the order the business is built. Start with opt-in and early engagement. Then the list. Then click-through. Then the page, checkout, and refunds. Stop at the first one that is clearly off. That is where the leak begins, even if a later number looks worse. How to diagnose why your course business isn't selling lays out all seven layers and how each one feeds the next.
How to find where enrollments stall
This is a structured analysis you can run on your own numbers in an afternoon. It is the numbers side of finding the bottleneck in your course business, which tests each layer with a question instead of a rate. The two work best together.
- Pick one window. Use a single launch, or one full month if you sell evergreen. Every count comes from the same window.
- Pull the raw counts. Opt-in page visitors, new subscribers, welcome sequence clickers, active subscribers, subscribers emailed, sales page clicks from email, sales page visitors by source, checkout starts, purchases, refunds.
- Calculate each stage rate. Divide each count by the one before it, using the formulas in the table above.
- Split warm from cold. Wherever you can, calculate the sales page and checkout rates separately for list visitors and for everyone else.
- Put it next to your own baseline. Compare each rate to the same stage in your last launch or a comparable month. Your own history holds your audience, price, and offer steady. Someone else's numbers do not.
- Find the first stall. Read the rates in build order and stop at the first stage that dropped against your baseline, or the first place the count falls off sharply.
- Confirm it with people. Numbers show where people left. They do not say why. Talk to a few people who stopped at that stage, or read their replies and refund reasons, before you change anything.
Step six is the discipline. You will likely see more than one number you do not like. The earliest one gets fixed first, because it may be the reason the later ones look weak. What should I fix first in my funnel? covers what to do when two stalls are both real.
A worked example
The numbers below are made up and rounded to keep the logic visible. They are an illustration, not targets.
A creator runs two launches of the same course at the same price. After the second one, sales fell and the sales page rate dropped from 10 percent to about 3 percent. The obvious conclusion is that the page stopped working.
| Stage | Launch 1 | Launch 2 |
|---|---|---|
| List size at launch | 1,500 | 2,000 |
| Click-through from email | 150 (10%) | 80 (4%) |
| Cold visitors from social | 0 | 220 |
| Purchases from email visitors | 15 (10% of 150) | 8 (10% of 80) |
| Purchases from cold visitors | 0 | 2 |
| Blended sales page rate | 10% | about 3% |
Split by source, the page did not change at all. Warm visitors bought at the same rate both times. What changed was click-through, which fell from 10 percent of the list to 4. The blended page rate dropped because cold visitors were mixed in, and cold visitors buy at a much lower rate on any page.
One stage further back, the reason shows up. The 500 newest subscribers came in through a new freebie, and almost none of them clicked anything in the welcome sequence. The list got bigger and colder at the same time. Rewriting the sales page would have changed nothing about that.
That is what a revenue leak looks like in practice. If click-through had held at 10 percent on the larger list, 200 warm visitors would have reached the page. At the same page rate, that is about 20 sales instead of 8. The missing sales were lost at the list, not at the page.
How a structured analysis uncovers revenue leaks
A revenue leak is the gap between what a stage produced and what it produced before, carried forward through every stage after it. Leaks compound. A drop early in the funnel shrinks the number of people who reach every later step, so the loss shows up again at the page, again at checkout, and again in total sales.
That is why the order of analysis matters as much as the numbers. A review that starts at the sales page will find something to improve there. It almost always can. But if the leak sits two stages earlier, the improvement lands on a smaller and colder group of visitors than it should, and the result is hard to read. The work was real. It was aimed at the wrong stage.
A structured analysis does three things a general review does not. It reads the stages in build order. It stops at the first real stall instead of cataloging everything. And it separates warm from cold before judging anything, so a change in who arrived is not mistaken for a change in how well the page works. How to tell if someone can actually fix your course funnel goes further into what that standard looks like when you are paying for the work.
What to set up before your next launch
- Tag your sales page links. Every link to the sales page from email gets a tag that identifies it as email. Every link from social gets a different one. Without this, you cannot split warm from cold.
- Build a one-page tracking sheet. One row per stage, one column per launch or month. Fill in the counts and let the sheet calculate the rates.
- Record the baseline first. Write down this launch's numbers before you change anything for the next one.
- Change one thing at a time. If the freebie, the emails, and the page all change in the same launch, the numbers cannot tell you which change did what.
- Keep the words, not just the numbers. Save replies to sales emails, questions asked during the launch, and every refund reason, in the person's own words.
When the numbers are too small to trust
Stage rates get unreliable when the counts behind them are small. If 40 people reached your sales page, three more or three fewer sales moves the rate by several points. That is noise, not a signal.
At low volume, look at counts as well as rates, use longer windows, and lean harder on conversations. Five honest conversations with people who read the sales page and did not buy will tell you more than a rate built on a few dozen visitors. The numbers still help. They point you at a stage. The people who stopped there tell you why.
It also changes what to do with traffic. More visitors will make your numbers more stable, but they will not fix a stall. Sending more people into a funnel with an early leak just makes the leak bigger. Do I have a traffic problem or a conversion problem? sorts out which one you are facing, and when you are ready to pay for traffic covers the point where more visitors start to pay off.
Common measurement mistakes
- Measuring only the end. Total sales tell you the result, not where it came from.
- Reading the funnel in visitor order. The first bad number a visitor meets is not always the first bad number in the business.
- Blending warm and cold traffic. A blended sales page rate can fall while the page is working exactly as before.
- Counting opens as engagement. Opens can be recorded automatically. Clicks and replies are closer to real attention.
- Tracking list size instead of active share. A bigger list can be a colder list.
- Comparing against someone else's benchmark. Your last launch is the only comparison that holds your context steady.
- Stopping at the purchase. Refunds and early start are where offer problems show up after a lift.
- Fixing every weak number at once. Fix the earliest stall, then measure again. Some of the later ones will move on their own.
The short version
Give every stage of your course funnel one number: opt-in rate, early engagement, active share, click-through, sales page conversion split by source, checkout completion, and refund rate. Compare each one to your own last launch, not to a benchmark. Read them in the order the business is built and stop at the first stall. That is where the revenue leak starts, and it is usually earlier than the place where the problem feels loudest. Confirm it with the people who stopped there before you change anything.
If you want a faster read on which layer is holding your enrollments back, the free Course Business Diagnostic tests all seven layers in order, stops at the first one that is not holding, and tells you what that means for what to fix first.
Start the free Course Business Diagnostic
FAQ
What metrics should I track for an online course funnel? Track one number per stage: opt-in rate, early engagement of new subscribers, active share of the list, click-through from email to the sales page, sales page conversion split by warm and cold visitors, checkout completion, and refund rate. Add earnings per subscriber and list conversion rate as summary numbers. The per-stage numbers tell you where people were lost. The summary numbers tell you how the whole system is doing.
How do I know which stage of my funnel is the problem? Compare each stage rate to your own previous launch or a comparable month, then read the stages in the order the business is built, starting with opt-in and ending with refunds. The first stage that dropped is where to look first. Confirm it by talking to a few people who stopped at that stage before you change anything.
What is a good conversion rate for each stage? There is no universal one worth using. Every stage rate depends on how warm your audience is, what the course costs, and how well the people arriving match the offer. Your own history is the useful comparison, because it holds those things steady. A stage that dropped against your last launch is a better signal than one that falls below a number you found online.
Why did my sales page conversion rate drop when I didn't change the page? The mix of visitors probably changed. If more cold visitors reached the page, from social or ads, the blended rate falls even when warm visitors buy at the same rate as before. Split the rate by source. If warm visitors are converting the same as last time, the page is fine and the change is further back in the funnel.
Can I measure niche and offer? Not directly, because neither one is a step a visitor passes through. They show up inside other numbers. A niche problem tends to appear as healthy opt-ins with weak early engagement. An offer problem tends to appear as warm, engaged visitors reaching the page and not buying, or as refunds rising after a sales lift.
How much traffic do I need before funnel metrics mean anything? Enough that a handful of people either way does not swing the rate. At low volume, use longer windows, look at raw counts alongside the rates, and rely more on conversations with people who considered buying. The numbers still point you at the right stage even when they cannot prove a small difference.
What is a revenue leak in a course funnel? It is a stage that is losing more people than it used to, carried forward through every stage after it. Because each stage feeds the next, a drop early in the funnel shows up again at the sales page, again at checkout, and again in total sales. The leak is the first stage where the loss began, not the last place you noticed it.