Measurement

Experiment Tracker

Write the hypothesis down first, or you will remember whichever result you preferred.

Included on the Free plan4 MCP tools

The problem this solves

You changed the headline and signups went up. Or the week was different. You have no way to tell, because the hypothesis was never written down and there was nothing to compare against — so the lesson does not survive to the next test.

What it does

Each test records what you believed, what you changed, what you measured and what happened. The value is not the individual result — it is that six months in you have a record of what works for your product rather than a feeling.

What you end up with

The output, not the description. This is what exists in your account after you use it.

A written hypothesis per test, recorded before the result is known.
The measured outcome next to the prediction, including the tests that failed.
A learnings log that survives longer than your memory of why you changed something.

How to use it

1
State what you expect
Before the change. A hypothesis written afterwards is a description.
2
Change one thing
Two changes and a result you cannot attribute is a week spent learning nothing.
3
Record the failures too
A log of only the wins is a highlight reel, and it will send you round the same dead end twice.

Where to find it

In the app

Included on the Free plan.

From your AI tool

Connect the MCP server and drive it from Claude, Cursor or ChatGPT:

create_experiment
list_experiments
log_experiment_result
update_experiment

Works with

🏗️

Sales Funnel Mapper

Five stages, a number on each, and the one that leaks.

Industry Benchmarks

Whether your 5% reply rate is good depends entirely on what you sell.

🎬

Video & Image Creatives

Ad creative from your own brand record, rendered on your machine.

Start with the experiment tracker

It is one part of a guided 8-step route from a product nobody knows about to paying customers. See how the journey works, or compare what each plan includes.

Start free