You can see the bounce rate. You can see the total conversions. What you can't see — the part that actually matters — is the exact step where people quit.
9 out of 10 visitors leave the average landing page. Not 5. Not 7. Nine. And most teams have no idea which specific moment pushes them away, because the tools they use report outcomes, not the path to them.
Here's how to find your real drop-off points and fix them.
Why "where" matters more than "how many"
A 12% conversion rate tells you nothing about the funnel. It doesn't tell you if 80% of visitors left before they saw your form, or if 80% started the form and abandoned it halfway through. Those two problems have completely different fixes.
"Bounce rate" aggregates two very different failures — a visitor who took one look and left, and a visitor who scrolled, read everything, clicked a second link, and then left — into a single number that's unhelpful for either case.
The six places visitors actually drop off
Most landing page funnels leak in the same handful of places. Work through this list in order when you're diagnosing.
1. Before the page renders
53% of mobile visitors leave a page that takes longer than three seconds to load. The average mobile page takes around 8.6 seconds. If your page takes four seconds, a huge chunk of your "traffic" never actually arrived. Start with your real Time to First Contentful Paint on 4G, not what your desktop dev tools say.
2. The first viewport
If the headline doesn't match the ad or link that brought them, visitors leave in seconds. This is usually the single biggest leak on pages that "look fine" in a browser. Watch a few session recordings on your top landing page and count how many scroll past the hero. A low number means the headline isn't landing.
3. Scroll depth to the CTA
If your primary CTA sits below the first viewport, scroll depth becomes the gating metric. A heatmap will show you where the scroll dies. A common pattern: a wall of text at 40% depth that stops everyone before they reach the button.
4. The form itself
Once a visitor clicks the CTA or lands on the form, a new leak opens. Which field do they start on? Which do they abandon on? Form drop-off usually happens on one specific field — and it's rarely the one you'd guess. The fix patterns are in this form abandonment guide.
5. Validation errors
A visitor who hits "Submit" and gets a cryptic validation error is statistically gone. Every error shown is a drop-off point you should treat as a bug, not a UX nicety.
6. Post-submit
The submit happened, but no lead arrived in your CRM. That's a drop-off most teams don't know they have. Webhook failures, spam filters, and marketing automation queues quietly swallow a real percentage of submitted forms.
Why your analytics tool won't show you this
GA4 will tell you the conversion rate. It will not tell you that step 3 is where everyone quits — unless you instrumented every step as a custom event, before the drop-off started happening. That's the catch: to answer "where did they drop off?" with most analytics tools, you needed to predict the question months ago and ship tracking code for it.
Heatmaps get closer but stop at the click. They don't tell you what happened after someone clicked the CTA and disappeared.
Session replay shows you one visitor at a time, which is great for texture but useless for "what percent of visitors drop at this step?"
Define the funnel, let it backfill from existing sessions
The fastest way to answer "where do visitors drop off" on a live landing page is to skip custom event instrumentation and build the funnel from sessions that already happened.
With CloseTrace funnels, you define the steps — landing → scroll to CTA → click CTA → form started → form submitted — and the funnel is backfilled from your existing session data. You see the drop-off between each step without waiting weeks for new traffic, and you can click into a specific drop-off bucket to watch the sessions where it happened.
That last part is what closes the loop. The funnel says 62% leave at step 3; the replays tell you why. For a broader view of how this fits with other behavior signals, see this marketer's guide to user behavior analytics.
What to do once you find the leak
Fix in this order:
- The biggest drop step, not the most emotionally annoying one.
- The fix that ships today, not the redesign that ships next quarter.
- One change at a time, so you can measure what moved.
If your biggest drop is at the hero, rewriting the headline to match the ad copy usually moves the number more than any visual redesign. If your biggest drop is at a specific form field, consider removing it — ask whether your sales team actually uses the data.
Most landing page teams skip this diagnosis entirely and redesign the whole page. That's expensive, slow, and usually doesn't help because the leak stays where it was.
The practical takeaway
Before you touch copy, design, or targeting, find the step. A funnel built from real sessions — not from events you planned in advance — will tell you in minutes what guessing takes weeks. Once you know which step is leaking, the fix is almost always small.