Preview
What it is.
It runs Karl Blanks and Ben Jesson's 9-step process over your funnel: define the success metrics, map the funnel, research visitors in three dimensions, then design tests from what the research turned up.
Funnel mapping looks for two different problems, high-traffic paths that convert badly and stages that were never built, and the second never shows up in a page audit. It scores the funnel out of 10 and names the failing rows.
What you get.
- A 9-step process from success metrics through research to scaling wins
- A funnel map naming underperforming paths and the stages that are absent
- Visitor research in three dimensions: who they are, what blocks them, what stops them
- Evidence-based test design instead of copied best practice
- A score out of 10, at about 1.4 points for each of seven diagnostic rows
- A top band that requires one clear action per page and research behind it
How to set it up.
Install with npx skills add wondelai/skills/cro-methodology --global, or take the marketing-cro plugin.
Define success metrics tied to business goals before touching any page. That is step one, and skipping it is why programs drift.
Map the whole funnel next, looking for two things: high-traffic paths that convert badly, and stages that were never built.
Research visitors in three dimensions separately: who they are, the usability problems that block them, and the objections that stop them.
Turn each discovered reason into a test hypothesis, so every experiment traces to evidence.
Check the page has one clear action. The top scoring band requires it, and competing calls to action are the common failure.
Pricing Plans
Free. MIT license.
Checked against the repo in July 2026.
Use cases
Finding the missing stage
Conversions leak somewhere a page audit can't see. Map the funnel and the absent stages show up explicitly, alongside the high-traffic paths that convert badly.
Rebuilding the test backlog
The current test list is borrowed best practice with no evidence behind it. Run the visitor research steps and every new hypothesis traces to a reason a visitor gave.
Scoring the whole funnel
You want to know where the program stands before planning the quarter. The diagnostic scores the funnel out of ten and names the failing rows, so you work the weak rows, never the interesting ones.
One action per page
Competing calls to action are the common failure the top scoring band screens for. The check tells you which pages fail it before you design a single test.
Best for
A conversion program built on best practices
Copied tactics are what this framework argues against, and the research step is the replacement.
Funnels with a missing stage
Absent stages are invisible to a page-level audit, and the funnel map looks for them explicitly.
Choosing what to test next
Every hypothesis traces to a researched reason, which is a better backlog than an idea list.