Conversion Funnel Diagnosing Where Visitors Drop Off

Most people hear "funnel" and think of TOFU, MOFU, BOFU — awareness, interest, desire, action. That's a real concept, and it's a useful one. It's also not what this page is about.

This page is about a different tool with the same name: the funnel report inside an analytics platform. It's how you find out exactly which step of your checkout, signup, or form is losing people — the kind of diagnosis that feeds directly into Conversion Rate Optimization, our complete guide to improving how many visitors convert once they're already on your site.

Conversion funnel vs marketing funnel: not the same thing

Two completely different jobs share one word — worth clearing up before we go further.

The marketing funnel is a strategy map. It sorts your channels and messaging by buyer stage — top of funnel (awareness), middle (interest, desire), bottom (action) — and helps you decide what to say to someone who's never heard of you versus someone who's about to buy.

The conversion funnel covered on this page is an analytics report. It doesn't care which channel brought the visitor in. It only tracks what happens once they land on a specific flow on your site — cart to checkout, signup step one to step two, trial to paid — and where they abandon it.

One is strategy. The other is diagnosis. You'll usually use both, but they answer different questions, and mixing them up wastes time. If you actually want the channel-and-buyer-stage version, read The Marketing Funnel: Stages Explained instead. If you want to know why 40% of people leave your checkout page, keep reading.

What a conversion funnel report actually shows

It's a sequence of steps you define, with a percentage attached to each one.

Say you run an online shop. You'd define a funnel like this: visits product page → adds to cart → starts checkout → enters payment details → reaches confirmation. Your analytics tool then shows how many visitors reach each step, and — this is the useful part — what percentage of the people who reached the previous step made it to the next one.

That step-to-step percentage is called the conversion rate for that step, and it's the number you're actually hunting for. A funnel report turns a vague worry ("checkout feels slow") into a specific, measurable fact ("68% of people who start checkout never reach payment").

The same idea applies well beyond e-commerce. A SaaS signup form, a lead-gen quote form, a free-trial-to-paid upgrade, even a multi-step content download — any flow with more than one step can be tracked this way.

How to read a funnel report and find the real leak

Look for the step with the biggest percentage drop — but don't stop there.

The instinct is right: the step that loses the largest share of visitors is usually where to focus first. If 90% of people move from product page to cart, but only 30% move from cart to checkout, that gap between cart and checkout is your leak.

Where people go wrong is treating that number as the whole story. Two questions to ask before you touch anything:

• Is this step actually necessary? A drop-off at "create an account" might mean the account creation is broken — or it might mean you're forcing an account when a guest checkout would do the job. Sometimes the fix is removing the step, not repairing it.
• Is some of this drop-off normal? A payment step will always lose some people — card declines, second thoughts, price comparison in another tab. Not every visitor who reaches checkout was ever going to buy. A raw percentage on its own doesn't tell you how much of the drop-off is fixable and how much is simply how people behave.

Once you've confirmed a step is genuinely leaking more than it should, that's where you dig — with session recordings, a heatmap tool, or direct user feedback — to find out *why* people are leaving, not just *that* they are.

This is also where a mid-body point about method matters: a funnel report tells you where to look. It rarely tells you why on its own. That's the gap between reading a report and running a proper diagnosis, and it's the reason CRO is a process rather than a one-off fix — something we cover end to end in our conversion rate optimisation service, which typically runs ₹40,000–₹1,50,000+ a month depending on how much of your funnel needs testing.

How to build a funnel report

The mechanics differ by tool, but the concept is always the same three things.

1. Pick one goal. A funnel tracks progress toward a single outcome — a completed purchase, a submitted form, an upgraded plan. Don't try to track five goals in one funnel; build a separate funnel for each.
2. Define the steps in order. List the pages or events a visitor should pass through on the way to that goal. Keep the sequence realistic — if visitors can legitimately skip a step (like jumping straight to checkout via a saved cart), your tool needs to allow for that or the numbers will look worse than reality.
3. Set the step boundaries as events or page views. Most analytics platforms let you define each step as a page URL, a button click, or a tracked event. The exact click-path for setting this up varies by tool and changes as platforms update, so we won't give you stale click-by-click instructions here — the concept above is what stays constant regardless of which tool you use.

Once it's built, check it against reality. If your funnel report says nobody ever reaches your confirmation page but you know orders are coming in, the tracking is broken, not your checkout.

Common funnel leak points worth checking

Four spots account for most of the drop-off we see across client accounts. Start here before building anything custom.

None of these leak points has a universal "normal" drop-off rate you can benchmark against — it depends heavily on your price point, your industry, and whether the traffic hitting that step was qualified in the first place. Treat the table as a checklist of where to look, not a set of targets to hit.

A quick example of why context matters more than the raw number: a high-consideration B2B trial with a long sales cycle will naturally lose more people between trial and paid than a low-cost consumer app people can buy on impulse. The same 70% drop-off can be a five-alarm problem in one business and completely expected in the other.