Users quit at step 3 — the form asks why at step 6

You've optimised your landing page, your ad copy is sharp, and your analytics show people are clicking through. Then the drop-off starts. Step one is fine. Step two is fine. Step three, and suddenly you're watching people bail like it's the last train out of Central. Meanwhile, the feedback form — the one that asks why they left — doesn't appear until step six. By then, everyone who could have told you something useful is already gone.

The feedback paradox: asking the right people at the wrong time

Here's the thing about exit surveys and feedback forms: they're built on a reasonable-sounding premise. Ask people what went wrong, fix it, move on. The problem is placement. You're asking the people who stayed what it was like for the people who left. The survivors of your form aren't a representative sample — they're the ones who found it tolerable enough to finish. Their feedback is real, but it's filtered through a very specific lens: "I made it through, here's what I noticed."

The people who quit at step three had a different experience entirely. Maybe the form asked for their phone number and they weren't ready to give it. Maybe a field label confused them. Maybe they were on a phone, on the train, and the keyboard kept covering the next input. Whatever it was, it was decisive enough to stop them cold — and you'll never hear about it if your feedback mechanism sits three steps downstream.

This is a classic case of what Kahneman and Tversky would call an availability problem. You're making decisions based on the feedback that's available to you, not the feedback that's representative. The people who complete your form are the ones whose experience was good enough to continue. Their input skews toward minor gripes and polish requests. The catastrophic friction — the stuff that actually kills conversions — lives in the silence.

Why step three is where the story gets interesting

Let's talk about what's actually happening at step three. In most multi-step forms, step one is identity (name, email), step two is context (what are you here for?), and step three is where you start asking for something that feels like a commitment. A phone number. A company name. A budget range. A specific date.

Behavioural research on commitment and consistency tells us that people are more likely to continue once they've invested something. But there's a threshold effect. Ask for too much too soon, and the cost-benefit calculation flips. The user isn't thinking "this is a reasonable form." They're thinking "I've given you my name and email, and now you want my mobile number before you've told me anything useful. What's the payoff here?"

There's also a well-documented phenomenon called decision fatigue — the more choices and inputs you stack up, the more each additional one costs the user in cognitive effort. By step three, they've already made several micro-decisions. If that step introduces a new category of demand (personal contact details, financial information, a written explanation), you're asking them to cross a psychological line. Some will. Many won't. And the ones who won't are the ones you most need to understand.

The mismatch between where people leave and where you ask

Think about the structure of a typical lead form or onboarding flow. The feedback mechanism — a "how are we doing?" prompt, an exit-intent survey, a "tell us why you're leaving" field — is almost always placed at the end. It's a logical place if you think of the form as a linear process. You finish, you reflect, you share.

But user behaviour isn't linear. People don't quit at the end. They quit at the point of maximum friction. And that point is almost never the final step. It's the step where the ask outpaces the perceived value. It's the step where the form introduces a new requirement that wasn't implied by the previous steps. It's step three, four, or five — rarely step six.

A concrete example: a Sydney-based professional services firm I know of had a six-step enquiry form for prospective clients. Step three asked for "approximate project budget." Step six asked, "Was there anything that made this form difficult to complete?" They were getting maybe forty completions a month and a handful of feedback responses. Almost all the feedback was about minor things — the budget field was awkward, the date picker was fiddly. Useful, but not transformative.

Then they added a lightweight exit-intent prompt that triggered if someone was inactive on step three for more than twenty seconds. The prompt was simple: "Stuck on this step? Tell us what's tripping you up." Within a month, the feedback was completely different. The budget field wasn't just awkward — it was a dealbreaker. People didn't know what number to put, didn't want to anchor themselves, and didn't see why it was needed before they'd even spoken to a human. The firm moved the budget question to a post-enquiry phone call. Completion rates jumped by roughly a third.

The lesson isn't "remove the budget question." It's "ask about the friction where the friction happens."

Designing for the moment of doubt, not the moment of completion

So how do you actually do this without turning your form into a pop-up circus? A few practical principles.

Instrument the steps, not just the endpoints. Track where people drop off, how long they linger on each step, and whether they interact with a field before abandoning. If step three has a high abandonment rate and a long dwell time, that's your signal. You don't need a feedback form to tell you something's wrong — the behaviour already did.

Offer a lightweight escape hatch at the friction point. Not a full survey. A single optional field, a "having trouble?" link, a one-question micro-prompt. The goal is to capture the reason at the moment it becomes relevant. If someone's stuck on a phone number field, ask "prefer not to give a number yet?" and offer a skip. You'll learn more from that interaction than from a hundred end-of-form responses.

Separate diagnostic feedback from satisfaction feedback. End-of-form surveys measure satisfaction. They're useful for that. But diagnostic feedback — why did you stop — needs to happen in situ. Treat them as different instruments for different jobs.

Be willing to restructure the form, not just annotate it. The most common outcome of good friction data is that you realise a question doesn't belong where you put it. Maybe it belongs later. Maybe it belongs in a follow-up email. Maybe it doesn't belong at all. The feedback isn't there to help you explain the form to users. It's there to help you redesign the form.

Where this goes next

The forward-looking move is to stop thinking of feedback as a destination and start thinking of it as a distributed signal. Every step in your form is a potential feedback point. Every abandonment is a data point. Every micro-interaction — a field focus, a pause, a backspace — is a clue.

The businesses that get this right won't be the ones with the prettiest exit surveys. They'll be the ones who treat the form itself as a conversation, and who build in ways to listen at every turn. Step three isn't a problem to be solved with a better step six. It's a question being asked. The trick is being there to hear the answer.