ChatGPT
Claude
Perplexity
Gemini
GrokForget the goldfish, keep the scroll data
The claim that human attention span dropped to eight seconds, shorter than a goldfish, has circulated in marketing decks for a decade. It is not true. BBC journalist Simon Maybin traced the figure in a 2017 investigation to an unverifiable source called "Statistic Brain", not to Microsoft's actual 2015 Canada consumer report, and not to any peer-reviewed study. No credible research supports it, and repeating it on a school website misdirects the design brief toward a number that doesn't exist.
What does exist is scroll behaviour, and it has been measured properly. Nielsen Norman Group's eyetracking study of 120 participants and more than 130,000 fixations found that users spend around 57% of their viewing time above the fold, down from 80% in the same lab's 2010 version of the study, because modern users scroll more readily than they used to. Even so, 74% of total viewing time still happens within the first two screenfuls of a page, roughly the top 2160px of scroll on a typical desktop viewport (Nielsen Norman Group, "Scrolling and Attention", published 15 April 2018).
For a school's programme page, that is the practical brief: a prospect scrolling through a Bachelor's or MBA listing gives roughly three-quarters of their attention to what loads in the first two screens. Everything after that competes for the remaining quarter, against a browser back button that is one thumb-flick away.
Scanning, not reading, is what happens on a programme page
A prospect landing on a programme page is not reading it top to bottom; they are scanning it for a handful of facts. Nielsen Norman Group's eyetracking research on reading patterns found that 79% of users scan a page rather than reading word by word, and only 16% read close to word-for-word (Nielsen Norman Group, "F-Shaped Pattern For Reading Web Content"). Eyes move in short horizontal sweeps near the top of a block of text, then drop down the left edge, picking up headings, bold terms and numbers along the way rather than full sentences.
This is not laziness; it is how anyone reads a page they haven't committed to yet. A prospect comparing three or four Bachelor's programmes across different schools' websites in one sitting has no reason to read any single page in full before they have decided it is worth the time. Long paragraphs of brochure copy about "excellence" and "a unique pedagogical vision" simply get skipped, because there is nothing in them for a scanning eye to latch onto.
The practical consequence is that a programme page's most important facts cannot be buried inside prose. They need to sit where a scanning eye actually stops: headings, bold labels, short lists, and the top-left of any content block. A tuition fee mentioned only in paragraph four of a "Why choose us" section might as well not be on the page at all for 79% of visitors.
Our pillar guide on what Gen Z expects from a school website covers this scanning behaviour across the whole site, not just programme pages — worth reading alongside this one if the redesign in front of you touches navigation and homepage structure too.
What has to survive the first two screenfuls
Given that 74% of attention lands in the first two screenfuls, a programme page has to answer the questions that make or break a decision before a prospect ever reaches a scroll depth most of them won't hit. In practice, five elements compete for that space, and all five earn it.
| Element | Why it has to be above the second screenful |
|---|---|
| Tuition fees (with the academic year) | The single most-scanned figure on any programme page; burying it below the fold sends prospects to a competitor's page instead of your enquiry form |
| Format (full-time, part-time, apprenticeship, distance) | Filters out a mismatch in seconds — a working adult scanning for part-time options will bounce the moment format is unclear |
| Key outcomes (accreditation, QAA review status, employability, progression) | The credibility check a scanning eye looks for before trusting anything else on the page |
| Next intake date | Answers the "can I actually start this" question — and whether it follows the UCAS undergraduate cycle or runs as a rolling postgraduate or MBA intake, which a brochure paragraph rarely states plainly |
| One clear next step (apply, book a call, chat) | Converts attention into an action before the visitor scrolls away and never comes back |
None of these needs a full paragraph. A short label-and-value pair, a badge, or a two-line summary block does more for a scanning eye than three sentences of context ever will. If a school's current programme page needs a prospect to scroll past a hero video and a mission statement before reaching the tuition fee, that fee is effectively invisible to three-quarters of visitors.
The intake-date line matters more for UK schools than it looks. A prospect scanning an undergraduate page is usually working against the UCAS application calendar, with its own deadlines and Clearing window each August; a prospect scanning an MBA or postgraduate page is often checking whether a rolling intake starts in September or January. Stating which cycle applies, in the first two screenfuls, saves a scanning prospect from assuming the wrong one and closing the tab.
Book a demoWhere prospects actually give up
Attention above the fold is only half of the picture; the other half is how quickly people leave once they've arrived. Chartbeat's analysis of engaged time across roughly 100,000 page visits on its publisher network over one week found that around 34% of visitors leave a page within 15 seconds of active engagement — scrolling, clicking, moving a mouse or touching a screen, or pressing a key (Chartbeat, "Using Engaged Time to Understand Your Audience"). That figure counts genuine interaction time, not idle tabs left open in the background, which makes it a closer proxy for real interest than raw time-on-page.
Chartbeat's benchmarks also give a rough scale for what "enough time" looks like by content type: typical engaged time runs 5 to 30 seconds for landing pages, 60 to 90 seconds for news articles, and 3 or more minutes for long-form content. A programme page sits closer to the landing-page end of that scale than most admissions teams assume, which is exactly why the facts that decide a prospect's next move cannot depend on them reading past the first screenful.
| Page type | Typical engaged time |
|---|---|
| Landing page (comparable to a programme page) | 5-30 seconds |
| News article | 60-90 seconds |
| Long-form content | 3+ minutes |
Source: Chartbeat, "Using Engaged Time to Understand Your Audience", based on a sample of roughly 100,000 page visits across Chartbeat network sites over one week.
Read against the 34% abandonment figure, a 20-second first impression is not an edge case for a programme page — it is close to the norm. A page that also loads slowly compounds the problem: Google's own guidance on page experience and Core Web Vitals treats how quickly the largest visible element renders as a ranking and usability signal in its own right, which means a heavy hero image can eat into the 15-second window before a prospect has even had the chance to scan anything.
Why the chatbot matters after the scan, not instead of it
None of this means a programme page should try to say everything at once. A scanning prospect who found the tuition fee, the format and the next intake date in the first two screenfuls still has follow-up questions the page was never going to answer completely — a scholarship eligibility case, how credits transfer from a previous degree, whether a specific elective runs this year. Those questions are exactly where a chatbot earns its place on the page, sitting quietly in the corner rather than forcing the prospect to hunt for a phone number or wait for an email reply.
A chatbot doesn't compete with a well-structured programme page for the first two screenfuls; it picks up where the page's necessarily short scan-friendly answers stop being enough. Skolbot's widget sits on the programme page itself, ready the moment a prospect's specific question falls outside what a tuition table or an intake-date badge can cover, and routes anything that needs a human straight to the admissions team instead of leaving the prospect to guess. For a deeper look at the mobile side of this same problem, our mobile-first enrolment UX checklist covers what changes when most of that scanning happens on a phone screen rather than a desktop monitor, and our UX checklist for forms and open days covers what happens once a prospect decides to act on what they scanned.
