Most funnel talk is about the two ends. Get more people in at the top. Get more revenue out at the bottom. That’s where the marketing spend lives and where the sales targets live, so that’s where the attention goes.
But I’ve lost more customers in the middle of a funnel than at either end. Someone lands on the site, fine. Someone buys, also fine. It’s the three or four steps in between where they quietly leave, and because nobody’s watching that section closely, it can go unnoticed for months.
This is about how to actually find that kind of leak and how to think about it once you’ve found it.
What “the middle” even means
A funnel is just a sequence of steps someone has to complete to become a customer, or to stay one. For a simple online store it might be: visit site, view a product, add to cart, start checkout, enter payment, complete order. For a subscription tool it might be: sign up for a trial, open the app, connect an account or upload data, use the core feature once, convert to paid.
The “top” is everything before the person has shown real intent. The “bottom” is the final one or two steps right before money changes hands. Everything else is the middle, and it’s usually the least defined part of the funnel because it’s the part that doesn’t map cleanly to a marketing channel or a revenue number. Nobody puts “connect an account” on a dashboard next to ad spend and MRR. It just sits there as an internal step, assumed to work.
That’s exactly why it’s worth checking. A step nobody is watching is a step that can quietly break.
Why top and bottom lie to you about the middle
Here’s the trap. If you only look at overall conversion rate, visitors to customers, a middle leak can hide behind a healthy-looking number for a while. Say your traffic quality improves and doubles the number of people entering the funnel. If your mid-funnel step is leaking 40% of people, that leak just doubles too, but your total customer count still goes up because you fed more people in at the top. The funnel looks like it’s working. It’s actually leaking twice as much as before, just diluted by more volume.
The reverse happens too. If your top-of-funnel traffic quality drops, so fewer people who enter are ever going to convert anyway, your overall conversion rate can fall for reasons that have nothing to do with a mid-funnel problem, and you can go looking for a leak that isn’t there.
The fix isn’t a smarter overall number. It’s breaking the funnel into steps and looking at the conversion rate from each step to the next one, not just start to finish.
Building the step by step view
Concretely, this means counting, at each step, how many people who reached that step went on to reach the next one. Not how many reached it out of everyone who started the whole funnel, how many out of everyone who reached that specific step.
Say you run a small online course. Your funnel is: visits sales page, clicks “start free preview,” watches at least one preview lesson, clicks “enroll,” completes payment. If 500 people visit the sales page in a month, 200 click start free preview, 150 watch a lesson, 40 click enroll, and 30 complete payment, the step to step rates look like this: 500 to 200 is 40%, 200 to 150 is 75%, 150 to 40 is 27%, 40 to 30 is 75%.
Read straight through, the number that jumps out is the 27% between “watched a lesson” and “clicked enroll.” That’s the middle leak. The step before it and after it both hold onto most of the people who reach them. This one loses nearly three out of four. That’s where I’d spend my time, not on the sales page and not on the payment form, even though those are the steps that feel more important because they’re the ones with a price tag or an ad campaign attached.
This is illustrative, not a claim about what your numbers will look like. Every business’s funnel has a different shape, and the only way to know where yours leaks is to actually build this table for your own steps.
What a leak like this can tell you, and what it can’t
The step to step rate tells you where people are stopping. It does not tell you why. That distinction matters more than it sounds like it should, because it’s tempting to look at a 27% conversion rate and immediately assume you know the story: price is too high, the offer isn’t compelling, whatever fits your existing theory of the business. The number doesn’t say any of that. It just says a lot of people who watched a lesson didn’t go on to click enroll.
It also can’t tell you whether fixing that step will actually grow revenue. Maybe some of the people dropping there were never going to buy regardless of how smooth that step was, they were curious and got their answer from the free lesson. Improving that step might raise the step’s own conversion rate without moving the number that actually matters to you, which is completed payments. A funnel step is a measurement of behavior, not a guarantee of what happens if you change something. You have to actually make the change and watch the downstream numbers to know if it mattered.
And a single month of data, especially at the volumes a solo business usually deals with, can be noisy. Thirty completed payments and forty enrollments is a small enough sample that a slow week or a batch of low-intent traffic can swing the step to step rate a fair amount without any real change happening. I’d want to see the pattern hold for a few periods before treating it as a real problem rather than a blip.
Common reasons a middle step leaks
In my own funnels, a mid-funnel drop usually traces back to one of a few things. A mismatch between what the earlier step promised and what this step delivers, someone expected the free preview to answer a specific question and it didn’t. A form of friction that has nothing to do with desire to buy, an account connection step that’s confusing, a page that loads slowly, a question the person doesn’t know how to answer yet. Or a timing gap, the step asks for a decision before the person has enough information or trust to make it, so they leave to go think, and a lot of them never come back to finish.
None of these show up in the funnel numbers themselves. The numbers tell you where to look. Finding out which of these it is takes something closer to the step: watching a handful of session recordings if you have them, reading support messages that reference that step, or just asking a few people directly why they didn’t continue. That’s slower and less satisfying than reading a dashboard, but it’s the part that actually explains the number.
Checking your own funnel
If you haven’t broken your funnel into steps before, the first pass is worth doing even before you suspect a leak. Write down the actual sequence someone goes through, not the idealized version, the real one including the steps you’d rather skip past. Count people at each step for a recent period and calculate the step to step percentage, not the percentage of the original total. Whichever step has the sharpest drop relative to the ones around it is where I’d start asking why, with the understanding that the number is a pointer, not an answer.
If you want more walkthroughs like this on funnels, cohorts, and the other numbers a one-person business actually needs to track, you can find the rest of the site here.
Get new guides and videos first — join the Telegram channel.