By Ryan Richardson · Published 8 October 2026
Across a fortnight of 5,965 recorded events on one funnel, 66% of visitors never scrolled a quarter of the way down the page. A thirty-day cut of the same funnel put a number on what that costs: of 2,119 sessions, 55% engaged at all, 23.8% reached the offer, and 2.9% bought. More than three in four visitors never saw the offer, not because the offer was weak, but because the page never got there.
Writing builds its case and lands the argument near the end. Nobody reads a page they didn't choose to be on that way. Put the promise, the price and the button inside the first two screenfuls, written last, once the rest of the page has worked out what the offer actually is.
A landing page view rate above roughly 70% is the working benchmark for a page that renders and loads fine. Well below that points at page speed or mobile rendering, not the advertisement sending traffic to it; mobile converts below desktop in every published panel this library draws on, so a page that reads fine on a laptop can still be failing on the device most visitors are actually holding.
Give every proof point its own box above the point where visitors stop scrolling, instead of burying it in a paragraph. A claim written into prose only reaches the visitor who reads carefully. A box reading a single number with one line under it reaches everyone scanning past it too, and it costs nothing to build; it's the cheapest fix most pages never make.
This isn't the same finding as a page that reads better selling worse (see the related failure below): formatting proof as scannable units is about whether a claim is seen at all before the scroll stops, not about whether a rewritten page holds attention longer.
A hero image pushing the price below the fold on a phone. A headline written to be admired rather than understood. A desktop-only review that never meets the handset the page actually arrives on. Commission the build and the mobile work in that order, hero first, then check it on somebody else's phone, out loud, before launch.
| Claim | Value | Source |
|---|---|---|
| Visitors who never scrolled a quarter down the page, one funnel, over a fortnight | 66% of 5,965 recorded events | The Sixty Steps manuscript |
| Thirty-day cut of the same funnel: engaged / reached offer / bought | of 2,119 sessions, 55% engaged, 23.8% reached the offer, 2.9% bought | The Sixty Steps manuscript |
| Working benchmark for landing page view rate | above roughly 70%; well below points at page speed or mobile rendering | The Sixty Steps manuscript |
| Matrix step: build the page assuming the hero is all most people see | Moderate difficulty, Standard practice, Automated tooling | The Sixty Steps matrix |
| Matrix step: format proof as scannable units, not prose | Easy difficulty, Often skipped, Automated tooling | The Sixty Steps matrix |