Insights · Web Design
Web Design · 6 min read
Five seconds is generous. The research says the aesthetic judgement is made in fifty milliseconds, and you have about ten seconds to justify it. Here is what has to happen in that window.
In 2006 a team led by Gitte Lindgaard at Carleton University showed people web pages for half a second and asked them to rate the visual appeal. Then they did it again at fifty milliseconds — a twentieth of a second. The ratings held. Their conclusion gave the paper its title: you have 50 milliseconds to make a good first impression.
Six years later a study involving Google's own UX researchers pushed it further. At exposures of 17 milliseconds — faster than a blink, which takes 100 to 400 — people were still forming consistent aesthetic judgements. Two things drove those judgements: visual complexity, and prototypicality. Simple, and familiar. Sites that were both scored highest.
So the five seconds in the title is a kindness. The look is judged before your visitor is conscious of having looked. What the following seconds decide is whether that judgement gets overturned — and whether they stay.
On that, the Nielsen Norman Group analysed page-visit data covering 205,873 pages and found that the odds of a visitor leaving are highest at the very start and fall steadily thereafter. Their instruction is precise: "To gain several minutes of user attention, you must clearly communicate your value proposition within 10 seconds."
Here is what has to be true in that window.
The single largest study of how people judge website credibility — Stanford's Fogg et al., with 2,684 participants — found that 46.1% of the comments people made about a site's trustworthiness concerned its visual design. Not its content. Not its claims. How it looked. The next largest category, information design and structure, accounted for 28.5%.
This is not vanity. Design quality is the proxy people use for competence when they have no other information about you. A dated template with three fonts and a stock photograph of a handshake is not a neutral choice. It is an active statement, and not the one you intended.
Nielsen Norman's guidance on homepages is direct: treat the homepage as an elevator pitch, "quickly and clearly conveying what your organization does and what users can accomplish on your site."
The test for your headline is brutal and takes ten seconds. Ask two questions of it: would it work just as well for your competitor? And would any company ever claim the opposite? If a rival could paste your headline onto their site without anyone noticing, and if nobody in your industry would ever advertise being un-integrated or un-reliable, then your headline conveys nothing at all. It is decoration.
Google and Deloitte studied 37 leading European and American brand sites across more than 30 million user sessions, tracking mobile load times hour by hour. A 0.1-second improvement produced measurable results: retail conversions up 8.4%, travel conversions up 10.1%, and for lead-generation sites — which most professional service businesses are — bounce rate on information pages improved by 8.3%.
A tenth of a second. Separately, Google's analysis of 11 million mobile landing pages found that as load time goes from one second to ten, the probability of a mobile visitor bouncing rises by 123%.
The thresholds to hit are published: LCP under 2.5 seconds, INP under 200 milliseconds, CLS under 0.1, measured at the 75th percentile of real visits. As of the most recent large-scale measurement, only 48% of mobile pages pass all three.
A note on a statistic you will see everywhere: the claim that "53% of visitors abandon a site taking over three seconds" does not appear on any current Google page. The verifiable figures are the 123% above, and Google's finding that 40% of consumers leave a page taking longer than three seconds. We would rather cite the real ones.
Nielsen Norman's work on B2B trust found that "seeing a physical location and humans behind the product or service bolsters their trust that the business is a legitimate one." Their four credibility factors are design quality, upfront disclosure, comprehensive and current content, and being connected to the rest of the web.
That means a real address, a real phone number, real faces, and real names — visible without hunting. A contact form and nothing else reads as a business that does not want to be contacted.
Nielsen Norman tracked over 130,000 eye fixations across 120 participants and found that 57% of page-viewing time happens above the fold, 74% within the first two screenfuls, and 81% within three. The top fifth of the page alone takes more than 42% of viewing time.
People do scroll — the old "nobody scrolls" claim is dead. But attention decays sharply, and content placed 100 pixels above the fold received 102% more views than the same content 100 pixels below it. Where you put something is a decision about how many people will see it.
Analysis of 45,237 real page views found that on an average page, users have time to read at most 28% of the words, and 20% is more likely. Nielsen Norman describe two scanning patterns worth knowing. The F-pattern — two horizontal sweeps then a vertical scan down the left edge — is what happens when a page gives the eye nothing to hold onto. It is a symptom, not a target.
What you want is the layer-cake pattern: fixations landing on headings and subheadings, with occasional deliberate dips into the body text. That only happens if your headings are descriptive and visually distinct. "Our Approach" is not a heading. "We build without templates, and here is why that matters" is.
Nielsen Norman are pointed about this: "Link labels and calls to action should be specific and have high information scent. Generic language like Click Here, Explore, or Learn More does not tell users what they will get."
One primary action, stated in the words your visitor would use. Book a free consultation beats Get Started. See what a site costs beats Pricing. Every additional competing button dilutes the one you actually want pressed.
You may have heard this justified with the famous jam experiment — 24 flavours produced 3% purchases, six flavours produced 30%. It is a wonderful study, but a meta-analysis of 50 experiments and 5,036 participants found the overall choice-overload effect to be effectively zero. Argue for a single clear call to action on the grounds of clarity, which is well evidenced, rather than choice overload, which is not.
Open it on a phone you have never used, on mobile data. Look for five seconds. Close it. Then answer, without looking again:
If you cannot answer all four, neither can the person deciding whether to call you. And unlike you, they will not give it a second try.
None of this requires a bigger site. Most of it requires a smaller one, more clearly written and considerably faster. Your vision, built to perform.
Lindgaard, Fernandes, Dudek & Brown (2006), Behaviour & Information Technology 25(2), DOI 10.1080/01449290500330448 · Tuch, Presslaber, Stoecklin, Opwis & Bargas-Avila (2012), International Journal of Human-Computer Studies 70(11) · Fogg et al. (2003), Stanford Persuasive Technology Lab, ACM DUX · Nielsen Norman Group: How Long Do Users Stay on Web Pages, Scrolling and Attention, The Fold Manifesto, F-Shaped Pattern, Layer-Cake Pattern, Homepage Design: 5 Fundamental Principles, Trustworthiness in Web Design, B2B Trust, How Little Do Users Read · Google/Deloitte & 55, Milliseconds Make Millions · Think with Google, mobile page speed benchmarks · web.dev, Web Vitals and CWV thresholds · HTTP Archive Web Almanac 2025 · Iyengar & Lepper (2000) and Scheibehenne, Greifeneder & Todd (2010).
All sources verified live on 7 August 2026.
← All Insights