It is one of the first questions we are asked, and it is fair: whoever runs a company has to budget, compare, decide. What follows is not a refusal to talk about money. It is the explanation of why a figure off the top of our head would be the least rigorous thing we could offer you.

What is being priced is not the number of pages

The price-per-page table assumes that the value of a website is proportional to the quantity of things it has. The evidence points in another direction, and sometimes in exactly the opposite direction.

Let us begin at the beginning. In 2006, Gitte Lindgaard and colleagues, of Carleton University, showed participants images of web pages for 50 and for 500 milliseconds, and asked them to rate the visual appeal. The ratings from the two conditions correlated strongly (r = 0.97). The aesthetic judgment of a page forms in about fifty milliseconds and barely changes with more exposure time.

It is worth saying what this study does not say, because it is the most misused statistic in the sector. Lindgaard measured the visual appeal of static images, with small samples (22, 31 and 40 participants). It did not measure trust, did not measure purchase, did not measure revenue. Whoever writes that you "have 50 milliseconds to win the client" is making it up. What the study authorises is narrower and, even so, uncomfortable: when your website opens, it has already been judged before the first line is read.

In 2012, Alexandre Tuch and colleagues, of the University of Basel, with Javier Bargas-Avila (Google), replicated the paradigm with 119 images and shortened the window to 17 milliseconds. The effect held.

What the instant impression uses

Knowing that the judgment is immediate does not say what it uses. For that there is the study by the Stanford Persuasive Technology Lab, led by B.J. Fogg. In 2002, 2,684 people evaluated the credibility of real websites and wrote free-form comments on what led them to trust or distrust. Of the 2,440 comments analysed, visual appearance (layout, typography, white space, images, colour) was the most mentioned theme of all, present in 46.1%. It was followed by information structure (28.5%) and information focus (25.1%).

The authors themselves anticipated the obvious objection, and answered it:

Are people really influenced that much by how a site looks, and not by more substantive issues? The answer seems to be yes, at least in this context.

It is a sentence from academia, not from an agency selling.

Even so, the caveats, because without them the number is worth nothing. It measures what people notice and comment on, not what determines the decision. The authors themselves warn that "design look" was the broadest category in the coding scheme, which inflates the percentage, and that participants had little motivation to scrutinise the sites in depth. With a more involved buyer, they say, the weight of appearance would fall, though it would not disappear. The authors even ask that the percentages be read as approximations. And the sites evaluated are now more than twenty years old. What is stable is the human bias the authors cite, documented since the 1940s, that looking good is interpreted as being good.

A useful contrast remains. It circulates everywhere, including in agency proposals, that "75% of people judge a company's credibility by the website's design", attributed to Stanford. We could not locate a primary publication with that number. What exists is 46.1%, refers to mentions in comments, and says something else. The distance between repeating blog statistics and reading the studies is what separates a proposal from a conclusion.

And your sector is not the average

The most ignored fact in the Stanford study is the variation by category. The percentage of comments on visual appearance was 54.6% in finance, where it weighed most, 50.5% in travel, 46.2% in e-commerce, 41.8% in health and 39.4% in non-profit organisations. There is no universal "weight of design". There is the weight it has in the judgment of those who buy what you sell.

The tension no one in our sector wants to admit

If NP holds that what is alike has no authority, there is a finding that forces us to be honest. Tuch and colleagues concluded that the pages perceived as most appealing were those of low visual complexity and high prototypicality, that is, those that resemble what is already expected of that type of site. And the Ehrenberg-Bass Institute, with Byron Sharp, holds the position opposite to ours: what makes brands grow would not be perceived differentiation, but distinctiveness added to physical and mental availability.

We do not hide this. We resolve it. The evidence rewards convention in structure, because the visitor has to recognise where they are and what is done there, and leaves differentiation to the expression of the brand. Even Sharp holds that a brand has to be immediately recognisable; what he contests is that the difference needs to be significant. Now, a ready-made template with the name swapped is, by definition, indistinct. NP builds from scratch not to be strange, but so that the result is recognisable as yours and reusable as an asset.

The economic reasoning: what perception does to margin

In this part, the one that matters to whoever signs the cheque, the management literature is clearer than the design literature.

In 2003, Raj Sethuraman econometrically decomposed the price premium that consumers say they are willing to pay for a manufacturer's brand over a private label. The average premium was around 37%. Of that premium, about 80% was attributable to brand equity, and about 85% of that equity came from brand image, not from perceived quality. In translation: most of what people are willing to pay extra is not bought with quality. It is bought with perception. And he found significant equity even in categories treated as undifferentiated, such as bleach and flour. This is stated willingness to pay, in groceries, not the actual purchase of services.

In the same year, Ailawadi, Lehmann and Neslin measured what happens when a brand raises its price. In brands with a high revenue premium, the elasticity was -0.183. In the weak ones, -0.921. This is the operational definition of competing on price: the indistinct brand is punished about five times more when it tries to raise prices. It does not compete on price because it chose to; it competes because it lost its room to manoeuvre.

The same authors compel a correction that serves us poorly and that we make all the same: the price premium, on its own, proves nothing. Only a third of the brands analysed had both a price premium and a volume premium; 55% charged more without selling more. There are strong low-price brands. NP does not sell high prices. It sells the capacity to sustain price, which is something else.

The framework comes from David Aaker, who in 1991 defined brand equity as the set of assets and liabilities linked to the brand, to its name and symbol, that add to or subtract from value. The decisive word is the second: an indistinct brand is not worth zero. It subtracts. And in 2006, Madden, Fehle and Fournier showed that strong brands generated superior returns to shareholders against a relevant benchmark, with less risk, controlling for share and size.

From the intangible to the invoice

The link that ties this to the HTML is missing. In 2020, Deloitte Ireland, with data from Fifty-Five, analysed on Google's commission 37 brands and about 30 million mobile sessions. An improvement of 0.1 seconds in speed correlated, in retail, with +8.4% in conversions and +9.2% in average order value, which is, literally, how much each client spends.

It is correlation, not causation: the variations occurred naturally, without a controlled experiment. And it is 37 brands, not the internet. The report is honest to the point of publishing what contradicts it: in lead generation, the same improvement was associated with 8.3% less bounce on mobile and 21.6% more progression from the first form step to submission, but conversion on mobile fell 1.9%. The report does not explain the contradiction, and neither do we.

Notice what this does to the price table. If the most measurable factor of all, speed, produces opposite effects depending on the business model, selling websites by number of pages is a promise the data does not support. And Google, in a modelling study (neither published nor peer-reviewed, therefore to be cited with tongs), estimated that going from 400 to 6,000 elements on a page was associated with a 95% drop in the probability of conversion. More is not better.

Where the money is lost, concretely

Baymard maintains an average cart-abandonment rate of 70.22%, aggregated from 50 studies between 2006 and 2025. It is indicative, not a measurement. The decomposition is what matters: 43% of abandonments are from people who were just browsing. Of the rest, the top causes include extra costs (39%), being required to create an account (19%), a long or complicated checkout (18%) and site errors (15%). This is self-reported data, from the United States.

Look at the list. A mandatory account, a complicated checkout and errors are design and engineering decisions. They are not price, they are not product. Someone made them, or did not. That is the margin the website sustains, or does not.

And do those who treat design this way gain from it?

There are big numbers, and none proves what we would like. They come with the caveats, because without them they are worth nothing.

McKinsey tracked five years of design practices in 300 listed companies. Those in the top quartile of its index grew, over the whole period, 32 percentage points more in revenue and 56 percentage points more in total return to shareholders than the peers in the same industry. McKinsey itself calls it correlation, not cause: there is no control group, and it is just as plausible that those already growing have the slack to put design on the board's table. And McKinsey sells design consulting.

The Design Management Institute maintains, with Motiv Strategies, a portfolio of 16 listed companies. In the 2015 edition it returned 211% above the S&P 500 over ten years. It is not a study: it is a hand-picked portfolio, with no control group and no risk adjustment, and it includes Apple, which alone can explain much of the difference. They entered after already being recognised for design, which is selection on winners. The author, Jeneanne Rae, rejects the causal reading.

Forrester paired direct competitors between 2010 and 2014. In cable, AT&T U-verse grew 35% in compound annual rate against 6% for Comcast; in retail, Amazon grew in the United States sixteen times more than Wal-Mart, 31% against 2%. Forrester itself calls it correlation and delimits it: it exists only where customers can switch suppliers and the experiences are differentiated, and in health it found none. It is five pairs, anecdotal and not statistical, and they confuse experience with a growth phase. And they measure the complete experience, not the design of a brand.

The counter-example we provide ourselves. From 2012, IBM went from one designer for every 72 programmers to one for every eight, and entrusted about one hundred million dollars and more than 1,300 designers to Phil Gilbert. It documents investment, not return: over the same period IBM's revenue fell for more than twenty consecutive quarters.

The number everyone cites does not exist

One thing remains to be disarmed, the one that always appears at this point: fixing late costs a hundred times more, it is said, and from there descended the folklore that every euro spent on user experience returns a hundred. Hillel Wayne and Laurent Bossavit traced it back to a chart attributed to the "IBM Systems Sciences Institute".

There is a small catch with that IBM Systems Sciences Institute study: it does not exist.

It was an internal IBM training programme, not a research centre, and there is no data to support it. Being folklore does not make the direction wrong: Wayne notes that research points, tentatively, to fixing early being cheaper. What does not exist is the multiplier.

None of these numbers proves that design generates money. They prove that those who treat the brand as a measured discipline are systematically on the right side of the distribution. And they prove something else: the alternative, not measuring, does not even have numbers to show.

That is why we do not give you a figure off the top of our head

Because we do not yet know what we are pricing. We do not know what your brand can charge today, nor what current perception is subtracting from it, nor at what point in your commercial journey the experience produces value. And we have already seen that this point changes from sector to sector and from stage to stage. As everything we do is created from scratch, there is no off-the-shelf price or timeframe to present to you.

What we price is the perception your website produces in the first milliseconds, the structure that prevents the loss of orders to friction, and the traceability that allows each decision to be defended before whoever pays. That is the object of website creation at NP, and it is what the Method makes auditable rather than a matter of opinion.

The next step is not a quote. It is a Strategic Listening: we sit down with you, we listen to the business before proposing anything at all, and from there comes a real figure with the diagnosis that justifies it. Write to us here and let the Council show you the Method.

Sources

Every number in this article was verified in the original publication. Where the primary source was not accessible, we say so in the text. It is the minimum asked of those who claim that the brand can be measured.