One Decision Per Step

---
name: one-decision-funnel
description: Audits any funnel or landing page for the newbie mistake that quietly kills conversion — asking a visitor to make more than one decision per step — then rewrites the page structure so each step has exactly one job. Trigger on "audit my funnel", "why isn't my landing page converting", "check my funnel for leaks", "is my website a real funnel".
---

# One Decision Per Step

Here's the tell that separates a real funnel from a website pretending to be one: on a
real funnel, a visitor never has to figure out what to do next. There's exactly one
door. Everything else, no matter how nicely designed, is asking a stranger to do the
marketer's job — decide what matters — and most of them won't.

This skill audits any page or sequence and finds where more than one decision is being
asked at once.

## Before you start

You need the actual URL(s) in the funnel, loaded and read as a visitor would see them
— not described from memory. If ads are already running, note which ad sends traffic
to which page; a mismatch there is itself a finding.

## Step 1: Walk it as a stranger

Load the entry page cold, as if you'd never heard of the business. Before reading
anything else, answer: **in the first five seconds, is it obvious what this page wants
me to do?** If you have to scroll or think, that's the first finding.

## Step 2: Count the decisions on the entry page

List every clickable option: nav links, multiple CTAs, a phone number AND a form,
links to other services, a blog link, social icons above the fold. Each one is a fork
where a visitor can leave the intended path. A landing page built FOR an ad should have
close to one: the single next step the ad promised.

Common newbie pattern: driving paid traffic straight to the homepage instead of a page
built for the specific promise in the ad. The homepage has to serve everyone, so it
serves no one particularly well — a dozen decisions instead of one.

## Step 3: Check each subsequent step the same way

A funnel is a chain of pages, and the rule doesn't relax after step one. A booking
calendar with an extra unrelated question, a form with fields that don't matter yet, a
confirmation page with three more links — each one is a place someone can wander off.
The fewer fields and choices before the ONE next action, the more people complete it.

## Step 4: Match the promise across every step

The ad said one thing. Does the landing page say the same thing, in the same words,
above the fold? Does the booking step still reflect that same promise, or has the
language drifted into generic service copy? A visitor who feels the story change loses
trust exactly there — note precisely where the language stops matching.

## Step 5: Rank the leaks

For every extra decision found, rank by how early it appears (earlier leaks cost more
— fewer people survive to see later problems) and how far it is from the intended path
(an unrelated nav link is worse than a second legitimate CTA).

## Step 6: The rewrite

For the top 3 leaks, propose the fix in one line each: "Remove the services dropdown
from the nav on this page — it's not a browsing page, it's a conversion page." "Move
the phone number below the form so the form is the one obvious action." Never propose
a full redesign — propose the smallest change that removes the decision.

## The report

1. **The verdict:** real funnel or website-pretending-to-be-one, one sentence.
2. **Decision count per step,** with the worst offender named first.
3. **The promise-match check** — does the story hold across every step.
4. **Top 3 fixes**, each a single sentence, ranked by how early they occur.

## Rules

- Load every page. Never assess a funnel from a description of it.
- "It looks nice" is not a finding. Count actual decisions.
- A funnel that only has one step (an ad linking straight to a form) still gets audited
  — the single step needs to be the RIGHT one decision, matched to the ad's promise.