What it does
It grades your value communication out of 100, from your website alone. Not your SEO. Not your traffic. How clearly a stranger can see what you are worth.
It is a diagnosis, not a pitch. We show you where the value leaks. Closing the gap is the work.
How it reads you
Five dimensions, scored the way a buyer weighs you, not the way a tool counts you.
It reads for two audiences now. The human who buys. And the AI engine that decides whether the human ever sees you.
It looks at your page, not just your words. The audit takes captures of your site and judges how well it is made: typography, colour, spacing and imagery. For design-led and premium businesses the craft is the proof. Ugly luxury is not luxury, and a page that says the right things badly still loses the sale.
Then it benchmarks you against every business we have put through it. You get your percentile, not just a number in a vacuum.
Why it is any good
It measures communication, not value. Plenty of strong businesses score badly. That is the point. The value is real. The website just does not say it.
It diagnoses before it prescribes. Ritson's rule. We name the problem before anyone sells you a fix, so the fix is aimed at the right thing.
The thinking under the hood
This is not a generic AI opinion. Step Change has worked with up to 100 challenger businesses a year for the past 20 years. The rubric encodes what we have watched separate the challengers that win from the ones that stall, and it draws on:
- Nail Your Value, our own method: claims audited Absolute, Relative or Parity; the Feature, Benefit, Implication ladder; the customer interrogation
- Ehrenberg-Bass and Byron Sharp on distinctiveness and mental availability
- Cialdini on proof, and why buried proof does no work
- Binet & Field on desire, the long and the short of it
- Rumelt on the difference between good strategy and bad
The rubric itself stays in the room. What you get is the verdict.
Your data, plainly
Your documents are never stored. Anything you attach is read once, in memory, to score your value proposition. It is never written to disk, never saved to a database, and never sent anywhere else. When the audit finishes, it is gone.
- Never used for training. Nothing you submit trains any model, ours or anyone else's. The audit runs on Anthropic's commercial API, whose terms prohibit training on business customers' inputs and outputs.
- Enterprise-grade AI. Powered by Claude on Anthropic's commercial platform, the same infrastructure and data protections their enterprise customers rely on. Your content is read, scored and returned. It is not mined, sold or shared.
- We keep the report, not your files. The report is our own work, and we keep it so you can return to it from your link, so the benchmark stays honest, and so we can keep sharpening the scoring. Your uploaded documents are no part of that.