The Conversion Optimisation stack · 51 cards
Conversion Optimisation: the science of the next tap
Conversion optimisation is removing everything that stands between a willing user and the thing they came to do.
Not tricking the unwilling. The dark-pattern version buys a metric and spends your reputation.
This stack is about the honest version: copy that speaks to the job, friction placed only where it helps, and flows that feel familiar enough to finish.
Why do users drop out of flows they started?
Because somewhere mid-flow, the cost of continuing exceeded the clarity of the payoff.
An unexpected field. A button that describes the system instead of the outcome. A step that looks different from every app they know.
Each one is small. Conversion dies by a thousand smalls.
The repair work is unglamorous: walk the flow as a first-timer, find every moment of hesitation, and either remove it or give it a reason to exist.
Common conversion optimisation problems, fixed
The same failures come up again and again when founders describe broken conversion optimisation. Each one below has its own guide: how to spot it in your funnel, why it happens, and the fix.
How do you write button copy that converts?
Say what the user gets, in the user's words.
Submit and Continue describe the system. Get my plan and Start listening describe the outcome. Every label is a tiny promise, and outcome promises outperform system descriptions in most tests.
The test worth running on every string: read it aloud after now you can. If the sentence collapses, rewrite it.
Is removing friction always good for conversion?
No, and this is where blanket advice fails.
Friction that raises confidence converts better than its absence: a confirmation before an expensive choice, a review step before payment, a deliberate pause that signals care.
The distinction is purpose. Friction with a payoff the user can feel builds trust. Friction from laziness just leaks people. Audit yours honestly and keep the first kind.
Should checkout flows be creative?
Almost never, and this is the one place the playbook is close to prescriptive.
High-stakes flows run on muscle memory. A checkout that behaves like every checkout the user has completed disappears from awareness, which is exactly what you want at the moment money moves.
Spend your creativity upstream, on the moments that build desire. At the decision itself, be boringly familiar.
The building blocks inside the Conversion Optimisation stack
- JTBD Copywriting — Every string speaks to the user's job, not the system's.
- Intent Mirroring — Hesitation gets an answer, not a louder button.
- Limited Offer — Urgency reserved for the genuinely undecided.
- Pattern Alignment — High-stakes flows behave the way users expect.
- Intentional Friction — A deliberate pause where confidence needs building.
There is no universal answer. Make it yours
None of the above is a rulebook, and the playbook does not pretend to be one.
Every building block ships with Make It Yours prompts: guided questions that shape the tactic around your product, your users, and your stage.
Work through them yourself, or hand them to your AI agent as context. The thinking should happen about your app, not the average app.
- Walk your main flow as a stranger: at which exact screen do you first hesitate, and why?
- Which of your buttons describe the system, and what outcome would each promise instead?
- Where in your flows would a deliberate pause raise confidence rather than cost you completions?
How the best apps do it
Airbnb: The booking flow surfaces total price, dates, and cancellation terms before commitment. Confidence-building friction, exactly where it belongs.
Uber: One screen, one decision, fare visible upfront. The flow disappears and the ride gets booked.
Pinterest: Signup copy sells the feed you are about to get, not the account you are about to create.
Usually works
- Write labels as outcomes in the user's words
- Show total costs before the final step
- Keep payment flows aggressively conventional
- Watch session recordings of real drop-offs
Usually backfires
- Optimise clicks at the expense of trust
- Remove the friction that was building confidence
- Invent novel interactions for high-stakes moments
- A/B test copy on a flow with a structural leak
Questions founders ask
What should you optimise first in a converting flow?
The structure, then the copy, then the polish.
A missing value story or a mistimed ask costs more than any button colour. Fix the order of value and asks first; words and details compound afterwards.
How do you find where users drop off?
Funnel analytics tell you where. Session recordings tell you why.
The pairing matters: a number shows the leaking step, but watching ten real users hesitate at it usually makes the cause obvious in an afternoon.
Do dark patterns work?
Short term, sometimes. They also train users to distrust every future screen you show them.
The playbook takes a side here: conversion is a reward for clarity and value, not a ransom extracted by confusion. The honest version compounds; the other kind churns.
Ask a question
Thinking…