Every piece of evidence on the page — the Stripe scan, the Rootly and Intercom leaderboard scores, the "stopped the busywork" frequency table — is the product demonstrating itself on other companies' pages. That's not the same as a customer saying it worked for them. The buyer who lands here is a marketing leader at a Series A–C SaaS company; that buyer has been burned by tools that diagnosed problems without fixing them, and they're looking for one signal that someone like them ran this and got a result. There is no such signal anywhere on the page. Add one named customer — a real company, a real person, a specific outcome ("We rescanned after shipping the fix; the score moved from 6.1 to 8.4 in three weeks") — and place it directly below the hero, before the feature explanation begins. Without it, the page is a very confident tool with no witnesses.

The hero positions Lytms as a diagnostic tool ("Score your marketing. See what works. Fix what isn't."). The body sells it as a category monitoring platform ("The category moves; the response is already drafted in your inbox"). The pricing section sells it as a copy generation suite ("Every finding ships with the rewrite drafted: page, blog, ad, social, in your voice"). The Accelerate section sells it as a managed service ("A senior marketing operator runs Lytms for you"). A buyer who lands on the hero thinking they're buying a page audit discovers mid-scroll that they're also buying a competitive intelligence feed, a content calendar tool, and an agency retainer. These are four distinct buying decisions with four distinct buyers. The page tries to convert all of them simultaneously and converts none cleanly. Pick the one buyer this page is built to close — the marketing operator who wants to know what's wrong with their homepage — and cut or subordinate everything that isn't that. The other use cases belong on the use-case pages already in the navigation.

The URL input bar sits in the hero; after that, the page runs through four feature explanations, a full pricing section, an FAQ, and a managed-service pitch before the final CTA ("Score your homepage →") reappears. A buyer who gets curious during the competitor scoring section — the moment they think "I wonder where I'd rank against Intercom" — has to scroll back to the top to act. That's the highest-intent moment on the page, and there's no action surface near it. Embed the URL input field directly below the competitor scoring section (section 03) and again at the bottom of the pricing section, before the FAQ. The scan is free; the friction of finding the input field is the only thing standing between curiosity and conversion.

The buyer's immediate skepticism is that any tool can generate a score; the question is whether the score is grounded in something real. The page gestures at this with the G2/Reddit corpus and the "same rubric" competitor comparison, but it never makes the methodology concrete enough to be credible. The Stripe scan snippet ("Stripe's specific proof — Hertz, URBN, £540 million — never reaches buyers who leave before scrolling past the generic claims that lead") is the closest the page gets to showing its work, and it's buried in a social proof stripe rather than used as the proof of methodology it actually is. Pull that Stripe example into the body of the page as a worked demonstration: here's what the scan found, here's the sentence it flagged, here's why. That single example does more to validate the scoring system than four sections of feature description.