Your progress0/6

Keep your progress

Enter your email and we’ll send you a one-click link. Save what you’ve done and pick it up on any device.

No password. If you’ve saved before, we’ll email your sign-in link right away.

Nice to meet you

Looks like you’re new. Add your name and we’ll create your account, then email your sign-in link.

We store your name and email to save progress and, per your choices above, send these emails. No spam, no sales calls.

Foundational· 8 min read· Lesson 2 of 5

Examples:

Turning clicks into customers: conversion funnel analysis

Make the lessons yours

Tell us where you sit and we'll run every example through a world like yours. One tap, and you can change it whenever you like.

What you'll be able to do

Break a customer journey into measurable, named steps
Calculate step conversion and end-to-end conversion, and see why they differ
Rank leaks by absolute people lost, not by ugliest percentage
Run a leak analysis on your own funnel numbers

Conversion funnel analysis means breaking a customer journey into named, measurable steps, calculating the conversion rate between each, and fixing the step that loses the most people. Done properly it replaces the most expensive sentence in marketing, "we need more traffic", with a cheaper one: "we need to stop losing people at step three".

Name the steps before you measure them

You cannot improve what you have not defined. Write the funnel as concrete, observable events, not vague stages: visited product page → added to basket → started checkout → entered payment → order confirmed. Each step must be a number you can actually pull from your analytics, unambiguously. "Considering us" is not a step; "viewed the pricing page" is. If two people on your team would count a step differently, the definition is not done.

Step rate vs end-to-end rate

Step conversion is the share of people moving from one step to the next. End-to-end conversion is the product of every step multiplied together, and it is always smaller than intuition expects. A journey of five healthy-looking 60% steps converts at 0.6 × 0.6 × 0.6 × 0.6 × 0.6 ≈ 7.8% end to end. Nothing is broken; that is just how multiplication punishes long journeys. Two practical consequences: shorter journeys convert better for purely arithmetic reasons, and a small improvement to any single step lifts the whole chain.

Find the biggest leak, with the maths shown

The classic mistake is chasing the ugliest percentage. Leaks are measured in people, not percentages: rank each step by volume × drop rate, the absolute number lost. A worked month for an online store:

Product page → basket: 40,000 people, 12% proceed. Lost: 40,000 × 88% = 35,200
Basket → checkout: 4,800 people, 55% proceed. Lost: 4,800 × 45% = 2,160
Checkout → payment: 2,640 people, 78% proceed. Lost: 2,640 × 22% = 581
Payment → confirmed: 2,059 people, 92% proceed. Lost: 2,059 × 8% = 165

The worst-looking rate is basket → checkout at 55%. The biggest leak, by thirty times, is the top step: 35,200 people a month leaving product pages. And context matters before panicking: browsing behaviour means product-page drop is always large, so compare each step against your own history and category norms, then ask where a realistic improvement buys the most. Moving product-page-to-basket from 12% to 13% adds 400 baskets a month; heroically lifting payment completion from 92% to 97% adds about 100 orders. Same effort, four times the return, from the step nobody was staring at.

Try it with your numbers

Pull last month's numbers for four steps of your own journey. For each step calculate: people entering × (1 − proceed rate) = people lost. Rank by people lost, not by rate. Then take the top of the ranking and estimate: if I improved this step's rate by one percentage point, how many extra orders is that? (People entering × 1%.) That number, against the same calculation for the step you were worrying about, tells you where the month's effort should go.

Two traps once you start

First, the same-people assumption. Step rates only chain together if the steps genuinely follow each other for the same people. If some buyers skip the basket via a buy-now button, your basket step undercounts and its "drop" is partly people taking a shortcut. Map real paths before trusting the chain. Second, the averaging trap from the previous lesson applies with force here: a step's blended rate can hide that mobile converts at half the desktop rate at exactly one step, which is a design problem with an address, not a general "mobile converts worse" shrug. Split the leakiest step by device, by traffic source, and by new versus returning, and the fix usually names itself.

Quick checkNo score: just to make it stick

A funnel shows: 20,000 visitors → 10% add to basket; 2,000 baskets → 50% reach checkout; 1,000 checkouts → 90% pay. Which step should get attention first?

Key takeaways

Define funnel steps as observable events anyone would count the same way, not vague stages.
End-to-end conversion is the product of every step, which is why five decent steps still make a small final number.
Rank leaks by absolute people lost (volume × drop), not by ugliest percentage.
Split the leakiest step by device, source and new vs returning: the blended rate usually hides the fix.

Common questions

As few as describe the real journey: four to six is typical for ecommerce. Every extra step is another multiplication, another report to maintain, and another place for definitions to blur. Add a step only when it isolates a decision you could act on.

Lesson check: three questions

Five steps each converting at 60% produce roughly what end-to-end rate?

Question 1 of 3

NextLesson 3Reading conversion rates without being fooled by volume