Your funnel is a story you're telling wrong

A funnel is usually drawn as a set of declining percentages — visitors, signups, activations, purchases — each stage a leakier bucket than the last. We stare at the biggest drop and throw tactics at it: a new headline, a shorter form, a discount. Sometimes it moves the number. Often it doesn't, and we're baffled, because the analytics said that's where the problem was.

The analytics were right about where and silent about why. A funnel isn't a series of buckets. It's a story your customer is being told, one promise at a time. Every stage makes a promise; the next stage either keeps it or breaks it. People don't drop out where the form is long. They drop out where a promise stops being believed.

Conversion isn't a mechanical problem. It's a question of whether each step keeps the promise the last step made.

Read the funnel as promises, not percentages

Take any funnel and rewrite each stage as the promise it implies:

• The ad promises: this solves a problem you have.
• The landing page promises: and here's specifically how, for someone like you.
• The signup promises: this will be worth the friction of starting.
• The onboarding promises: you'll get the thing you came for, quickly.
• The purchase promises: paying is the obvious next step, not a leap.

Now the drop-offs read differently. A collapse between ad and landing page isn't a design problem — it's a broken promise. The ad sold one thing; the page delivered another; the customer felt the bait-and-switch and left. No amount of button-colour testing fixes a promise that was never kept.

The drop that matters isn't the biggest one

The instinct is to fix the largest percentage drop. But the largest drop is often natural and healthy — plenty of curious visitors were never going to buy, and losing them is qualification, not failure. The drop that matters is where you lose people who had already believed you. Someone who completed onboarding and then didn't purchase was sold. You had them. Losing them is a broken promise late in the story, and it's far more expensive than a stranger bouncing off the homepage.

So before optimising the biggest leak, find the most committed leak — the latest stage where believers are still walking away. That's where the narrative is failing the people who matter most.

A stranger leaving your homepage is qualification. A believer leaving at checkout is a broken promise.

How I actually move these numbers

When I led journey analytics and experimentation work that delivered +3.5% conversion and +8% funnel efficiency, almost none of the wins came from cosmetic A/B tests. They came from finding the exact moment a promise broke and repairing the narrative there.

1. Watch the believers, not the averages

Segment the funnel by intent, not just by stage. The behaviour of high-intent users reveals where genuine promises break; the average blurs it. A retention strategy I designed improved repeat behaviour precisely because we stopped optimising for the median user and started watching where committed users hit friction.

2. Map the emotion, not just the click

Every stage has a feeling attached: hope, doubt, relief, hesitation. Journey mapping that captures the emotional arc tells you where confidence drains — which is always upstream of where the click is lost. The drop-off is the symptom; the doubt one step earlier is the cause.

3. Fix the promise, then test the execution

Get the narrative right first — does each stage deliver what the last one sold? — and then run experiments to optimise how that promise is kept. Testing tactics on top of a broken promise is rearranging furniture in a house with no foundation.

The reframe

Stop asking "where are we losing people?" and start asking "where are we losing people who already believed us, and what did we promise them that we failed to keep?" The first question gets you a heatmap. The second gets you a customer back. Your funnel was always a story. The only question is whether you're telling it on purpose.