RESEARCH
All researchThe gap.
A product takes years to build. In a 2012 lab study, 17 milliseconds on screen was enough for a web page's design to shape how appealing people found it.
on screen was enough for a web page's design to shape how appealing people rated it.
TUCH ET AL., 2012 · STUDY 2, 82 PEOPLE, 80 SCREENSHOTSLIMIT · THIRTY PEOPLE SAW PAGES AT 17 MS. THEY RATED BEAUTY, NOT WHETHER TO BUY.The question.
Your product moves on every year. Your brand gets made once, for the company you were, and stays there. That distance is the gap. This piece asks what the gap costs a company whose buyers, investors and hires cannot check the technology for themselves.
How we know.
No single study answers this, so we read nine outside sources and say whose each one is. Two are lab experiments on first impressions of web pages: Tuch and colleagues in 2012, and Lindgaard and colleagues in 2006. Two are experiments on whether looks change how usable a machine seems, one that found it in 2000 and one that did not in 2022. The rest are a 2002 comparison of 2,684 consumers with 15 experts, a 2021 survey of deep tech ventures, two Gartner buyer surveys and a 2010 paper on distrust. The last source is our own audit of 45 funded deep tech websites in September 2026.
What it means.
If the people judging you cannot assess the technology, they assess what they can: the page, the words, whether anyone is named. The first part of that judgement, the look, forms within milliseconds in the lab. None of these studies measured enquiries or revenue, so read this as a mechanism, not a multiplier. The product matters most, and it is the one thing a stranger cannot check.
Who notices how it looks.
SHARE OF CREDIBILITY COMMENTS THAT MENTIONED THE DESIGN LOOKCONSUMER WEBWATCH AND STANFORD, 2002. THE PEOPLE WHO KNOW THE SUBJECT BARELY MENTION THE LOOK. THE EXPERT PANEL WAS 15 PEOPLE.
Three findings, three sources.
WHY THE LOOK CARRIES SO MUCH WEIGHTHow well the machine really worked had no significant effect on that answer, and a 2022 experiment did not find the effect of looks at all.
People who cannot check the science still have to decide.
The range is a Gartner analyst's description, not a survey result. That few of them can check your technology is our inference.
In depth.
THE LONG READSeventeen milliseconds on screen is enough.
In 2012 Alexandre Tuch and four colleagues published two experiments in the International Journal of Human-Computer Studies. In the first, 59 people, mainly psychology undergraduates at the University of Basel, saw 119 screenshots of company home pages for 50, 500 or 1,000 milliseconds and rated each from ugly to beautiful. In the second, 82 people saw 80 screenshots for 17, 33 or 50 milliseconds. Thirty of them were in the 17 millisecond group. At every exposure, pages that were visually simple and looked typical for their kind were rated the most appealing. Design already moved the ratings at 17 milliseconds. That is how long a page was on screen, not how long people took to answer. The work builds on Lindgaard and colleagues, who reported in 2006 that ratings of visual appeal after 50 milliseconds correlated with ratings after 500.
Looks were read as quality in one experiment, and not in another.
In 2000 Tractinsky, Katz and Ikar of Ben-Gurion University of the Negev reported an experiment in Interacting with Computers. It put 132 third-year industrial engineering students in front of a simulated cash machine. The researchers varied how good the screen layout looked, at three levels, and how well the machine actually worked, at two. Before use, how beautiful people found the interface correlated with how usable they judged it at 0.66. After use it was 0.71. The machine's real usability had no significant effect on what people said about its usability afterwards. Its looks did. That result is contested. In 2022 Sauer and Sonderegger had 110 people operate a fully automated coffee machine over seven weeks and found no effect of visual aesthetics on user experience, including perceived usability. They suggest product type and how strongly users identify with a product may explain it.
The expert is the last person to notice.
In 2002 Consumer WebWatch, the Stanford Persuasive Technology Lab and Sliced Bread Design ran two parallel studies. In one, 2,684 people each judged the credibility of two live websites and wrote down why. In the other, 15 experts, eight in health and seven in finance, judged the ten health and ten finance sites those consumers had seen. On finance sites, 54.6 percent of consumer comments mentioned how the site looked. For the finance experts it was 16.4 percent. On health sites the figures were 41.8 and 7.6 percent. The experts wrote instead about who was behind a site, the quality of its information and the motive of the company. A founder is the expert on their own site. Most of the people arriving at it are not. The limits are plain: fifteen experts, consumer sites, the year 2002, and only part of the 2,684 saw those two categories.
Most of the people judging cannot check the science.
In March 2021 BCG and Hello Tomorrow surveyed 116 deep tech ventures and investors. Of the ventures, 81 percent said investors on average lack the scientific or engineering expertise to assess deep tech potential. A Gartner analyst, in the release of a survey of 632 business buyers run in August and September 2024, describes buying groups of five to 16 people across as many as four functions. A year later, from 646 buyers, it reported that 67 percent prefer a rep-free experience. If buyers would rather not talk to a salesperson, what the company has published has to do the talking. In September 2026 we read 45 funded deep tech websites. On 43 we found no working email for a named person. Six of those could not be confirmed because the page never loaded. Gartner's surveys cover business buyers in general, not deep tech.
What none of this shows.
No study here measured enquiries, deals or revenue. Tuch measured beauty ratings of screenshots. Tractinsky measured the perceived usability of a cash machine. The Stanford work counted comments about consumer health and finance sites. Putting them together to say that appearance decides who calls is our argument, not a result anyone measured. The gap can also be closed in the wrong direction. Darke, Ashworth and Main reported in the Journal of the Academy of Marketing Science in 2010 that when a product fails to live up to what was claimed for it, the resulting distrust can spread to unrelated products and companies, even when people can examine the next product themselves. That is consumer research, not deep tech. So the aim is not to look better than the product. It is to look as good as it is.
Three pieces close the gap.
None of them is a rebrand. Say it: one sentence anyone can repeat, because a group of five to 16 people has to pass it on without you in the room. Show it: a brand that looks as good as the work, because the look is judged within milliseconds. Prove it: real people and real results, easy to find, starting with a named person and a direct email. Make one claim checkable in under a minute by someone who is not a specialist. Then test it on someone outside your field, because you are the expert. Show them the home page for a few seconds and ask what the company does and who they would contact. The product matters most. It is also the one thing a stranger cannot assess.
What to do with it.
THREE PIECES CLOSE THE GAPSay it.
One sentence anyone can repeat. A buying group of five to 16 people has to pass it on without you in the room.
Show it.
A brand that looks as good as the work. In the lab, the look is judged within milliseconds.
Prove it.
Real people and real results, easy to find. Start with a named person and a direct email.
ONE QUESTION
How wide is your gap: close, a bit, or a long way?
Pick one. If it is a long way, start with whichever piece is missing: the sentence, the look or the proof.
Asked about this research.
NOT ANSWERED HERE?
tony@appart.giEVERY FIGURE HAS A SOURCE
Sources and limits.
- TUCH ET AL., 2012Tuch, A.N., Presslaber, E.E., Stöcklin, M., Opwis, K. and Bargas-Avila, J.A. The role of visual complexity and prototypicality regarding first impression of websites: Working towards understanding aesthetic judgments. International Journal of Human-Computer Studies 70(11), 794 to 811. Two lab studies, 59 and 82 people.
- LINDGAARD ET AL., 2006Lindgaard, G., Fernandes, G., Dudek, C. and Brown, J. Attention web designers: You have 50 milliseconds to make a good first impression! Behaviour & Information Technology 25(2), 115 to 126. Three studies. Read at abstract level.
- TRACTINSKY ET AL., 2000Tractinsky, N., Katz, A.S. and Ikar, D. What is beautiful is usable. Interacting with Computers 13(2), 127 to 145. One experiment, 132 students, 124 analysed.
- SAUER AND SONDEREGGER, 2022Sauer, J. and Sonderegger, A. Visual aesthetics and user experience: A multiple-session experiment. International Journal of Human-Computer Studies 165, article 102837. 110 people over seven weeks. Read at abstract level.
- CONSUMER WEBWATCH AND STANFORD, 2002Stanford, J., Tauber, E.R., Fogg, B.J. and Marable, L. Experts vs. Online Consumers: A Comparative Credibility Study of Health and Finance Web Sites. Consumer WebWatch, October 2002. 15 experts, compared with the 2,684 consumers of the parallel Stanford Persuasive Technology Lab study, How Do People Evaluate a Web Site's Credibility?
- BCG AND HELLO TOMORROW, 2021The Deep Tech Investment Paradox: a call to redesign the investor model. Boston Consulting Group and Hello Tomorrow, 2021. Survey of 116 ventures and investors, March 2021.
- GARTNER, 2025Press release, 7 May 2025. Gartner Sales Survey Finds 74% of B2B Buyer Teams Demonstrate Unhealthy Conflict During The Decision Process. 632 B2B buyers, August to September 2024. The five to 16 figure is an analyst's statement in the release.
- GARTNER, 2026Press release, 9 March 2026. Gartner Sales Survey Finds 67% of B2B Buyers Prefer a Rep-Free Experience. 646 B2B buyers, August to September 2025.
- DARKE ET AL., 2010Darke, P.R., Ashworth, L. and Main, K.J. Great expectations and broken promises: misleading claims, product failure, expectancy disconfirmation and consumer distrust. Journal of the Academy of Marketing Science 38(3), 347 to 362. Read at abstract level.
- OUR AUDIT, 2026APP/ART audit of funded deep tech company websites, read one by one on 14 and 15 September 2026 and recomputed on 18 September. 52 sites read, 45 kept with the full audit.
- LIMITSNo source here measured enquiries, deals or revenue. The lab studies rated screenshots and a simulated cash machine. The expert comparison rests on 15 experts and consumer sites in 2002. The Gartner surveys are not about deep tech, and the BCG survey reports what ventures say. One 2022 experiment found no effect of looks. In our audit, six contact routes could not be confirmed. They are unknown, not failures.
Send your site. We name the first fix.
One line about where you are going is enough. The reply comes back in writing. No call needed.
Or book a 30 minute call
