How Nudgent measures conversion health

For people

1. 01Cognitive ClarityCan I process this without strain?
2. 02Decision ClarityIs the next step obvious?
3. 03Trust SignalDoes this feel safe?
4. 04Motivation StrengthIs the value clear before the ask?
5. 05Comfort LevelAre my anxieties addressed?
6. 06Flow CoherenceDo the steps connect logically?
7. 07Identity MatchCan I see myself here?

For AI agents

1. 01ExtractabilityCan I read this without fighting the page?
2. 02Passage RetrievabilityCan an engine lift a self-contained answer out of this page?
3. 03Claim VerifiabilityAre these claims attributed and checkable?
4. 04AnswerabilityDoes the page carry the facts an agent would cite?
5. 05FreshnessAre the dates and recency signals current?
6. 06Markup HygieneIs the structured data clean enough to parse?
7. 07Entity & AuthorityIs the entity named here and corroborated elsewhere?

For people

10 is excellent and 0 is broken: 8 to 10 is strong, 5 to 7 needs work and 0 to 4 is critical. The overall score is the sum of the seven, scaled to 100.

For AI agents

The overall score weights the seven by how much each matters on that kind of screen; by default, whether the page carries the facts an agent would cite and whether a passage makes sense lifted on its own count for the most.

It then falls into one of three bands: below 45, agents are unlikely to use the page; 45 to 64, they can find and use it; 65 and above, it is well prepared for them. A page that blocks bots lands in the lowest band whatever it scores, and one that partly blocks them cannot rise above the middle.

For people

Nearly a hundred named principles from psychology, behavioural economics and decision research: Cialdini on social proof and commitment, Kahneman on loss aversion, Fogg on what it takes for someone to act, and Sweller on cognitive load, among others.

For AI agents

Close to forty more, drawn from the standards a parser actually reads, such as Schema.org, the WCAG accessibility guidelines and Open Graph, and from information-retrieval research into how an engine chooses a passage to quote.

For your kind of business

Before the analysis starts, the page is classified by industry and business model. Where we have written a brief for that kind of business, the analysis reads the page against it.

Recorded today

Each audit records what it predicted: the scores, the problems it found and the order it ranked the fixes in. When a page is audited again after changes, the score changes are recorded against the fixes that were made, and benchmarks for each industry and type of page are recomputed every night.

Not there yet

The test that matters most is whether a fix moved the conversion rate, and that needs teams to tell us their results. Too few have so far for us to tune the scores against conversion outcomes. Until that changes, the scores are a structured reading of the page, not a forecast backed by results, and this page will say so.

Live traffic or sessions

An audit reads the page itself, from its URL or from a screenshot you upload, which is also how it reaches screens behind a login.

Brand strategy

Positioning is judged only as far as it shows on the page.

What to charge

We score whether a pricing page reads clearly, not whether the prices are right.

Search, ads and attribution

Ranking, ad performance and attribution are left to the tools built for them.

Visual taste

Looks count only where they change how clear or trustworthy a page feels.

Bot protection

The agent questions note when a CAPTCHA or a rate limit keeps agents out, and never recommend removing a defence you rely on.