Your brand was designed for paper. Now it lives where you never designed it.

The question almost always arrives with an apology. On paper, they tell us, the brand is fine, but online it looks amateurish, and never appears the same twice. Is it worth touching this, or is it vanity. Before answering, it is worth changing the question. What you describe is not a problem of taste or of finish. It is that your brand has stopped being an image and become a system, and a manual designed for paper has no answer for the places where it now lives: a favicon of a few pixels, an app icon, a dark background, a vertical video and, increasingly, a line of text inside the answer of an artificial intelligence assistant.

The brand has stopped being an image. It has become a system.

Let us start with the example no one will accuse of amateurism. In 2015, the most seen brand in the digital world rebuilt itself. Google stopped having a logo and came to have a three-state system: the logo written out, the coloured dots and a compact G. The dots are not decoration, they are the state that animates when an assistant is listening or answering. The G exists for the places where nothing else fits, such as the favicon of a browser tab. The reason is written by Google itself.

Users now engage with Google using a constellation of devices, and our brand should express the same simplicity and delight they expect from our homepage, while fully embracing the opportunities offered by each new device and surface. Google Design, Evolving the Google Identity, 2015

A constellation of devices. It is not a marketing phrase, it is an engineering description: the same brand has to work on a watch, on a voice speaker, on a television screen and on an icon the width of a fingernail. A single, fixed design, made for one format, does not survive them all. That is why Google did not design a better image. It designed a system with a recognizable centre and rules of adaptation.

The idea is not new, nor is it from technology. Casa da Música, in Porto, has had since 2007 an identity that does exactly this. Stefan Sagmeister reduced Rem Koolhaas's building to an angular shape and built a generator: six profiles of the building, seventeen planes, and colours recalculated from an image of each event, the colours of a portrait of Beethoven for a Beethoven concert. The shape stays, the expression changes from poster to poster. MoMA and the Cooper Hewitt keep the piece in their collections. What was the avant-garde of a cultural institution almost twenty years ago is today the minimum requirement of any brand that appears on a screen.

What builds recognition lives on repetition, and almost none of it is yours

Let us reframe the problem in terms a board recognizes. A brand is recognized by its distinctive elements: the colour, the typeface, the shape, the symbol. The problem is that most of those elements belong to no one in particular. In a study published in the Journal of Brand Management, two researchers tested 44 identity elements of the twelve largest car brands in the United Kingdom, measuring how many buyers associated each element with a single brand. Thirty-three of those forty-four elements failed the minimum threshold of unique ownership in buyers' memory. Only the logo proved consistently ownable.

Results of testing 44, in-market and de-branded assets show that logo is the most uniquely ownable asset type. Fonts, slogans, and colours all emerge as having Low Uniqueness Concentration. Fiocchi e Seyed Esfahani, Journal of Brand Management, 2024

Typefaces, colours and slogans that several brands share at the same time in the buyer's mind. And the authors say why: the elements that belong to no one come to belong to someone only through, in their words, consistent usage, repeated and disciplined use at every point of contact. Being recognizable is not having a pretty logo. It is appearing the same, many times, everywhere.

Add to this the speed at which judgment forms. Lindgaard and colleagues measured, in 2006, that the aesthetic appraisal of an interface forms in about 50 milliseconds, and that this judgment is stable: when repeated, it correlates strongly with itself. In other words, the eye decides early, and does not easily go back. Put the two things together. Recognition is built through consistent repetition, and the impression forms in an instant that does not repeat. A brand that appears degraded or different from itself in the contexts where that instant happens is not building recognition. It is spending it.

And your sector is not the average

It is worth saying what this data is and what it is not. The car study measures perceived ownership in the United Kingdom car sector, through an index of memory exclusivity, not revenue or balance-sheet value. Lindgaard measures the visual appeal of static images of web pages, not sales. Neither of the two proves that a coherent brand brings more money to your company. What they prove is a mechanism: recognition depends on consistency, and the first impression is fast and stable. The magnitude of the effect on your business depends on your market, and we will deal with that using your numbers, not the averages of others.

The number our own sector cites, and that we do not use

Here, by the logic of the sale, we should present you the statistic every agency has on a slide: consistent brand presentation increases revenue by 23%. We will not use it, and it is worth explaining why, because the explanation is the difference between a line of reasoning and a sales argument.

We went to the source of that number. It comes from a report by Lucidpress, now Marq, a company that sells brand management software, produced with Demand Metric from a survey of a few hundred marketing professionals. Even a brand agency that went to check the methodology, and that in fact argues that investing in brand pays off, describes it bluntly as a self-reported survey of marketers, not a controlled study. It is not measured revenue. It is the self-reported opinion of those who do marketing about the effect of their own marketing, published by those who sell the tool. The direction of the number is plausible, and fits what we have already said about the speed and stability of the first impression. The magnitude does not hold up. Selling you 23% as if it were a financial fact would be doing what this blog exists not to do.

Where the first impression happens now

That leaves the question that matters to whoever decides a budget: where is your brand actually seen. The answer has changed, and it changed fast.

First, the screen. Mobile is already the majority of web page views worldwide, at 51.51% against 47.12% for desktop, according to StatCounter in June 2026. Your brand is seen mostly on a small piece of glass, in the hand of someone who did not ask to see it.

Second, and more decisive, often the brand does not even reach a page. In the United States, which is the most measured market, 68.01% of Google searches ended without any click in the first four months of 2026, according to SparkToro's analysis of Similarweb data. And when Google shows an answer generated by artificial intelligence at the top, behaviour changes even more. The Pew Research Center followed 68,879 searches from 900 North American adults and measured it.

Users who encountered an AI summary clicked on a traditional search result link in 8% of all visits. Those who did not encounter an AI summary clicked on a search result nearly twice as often (15% of visits). Pew Research Center, 2025

In the same study, users clicked a link inside the AI answer itself in only 1% of visits, and about one in five searches in March 2025 already produced one of those answers. Google, for its part, stated that the feature had 2 billion monthly users in July 2025. Translated into what matters: a growing share of the people who know your company will see it first not on your site, but inside an answer box you did not design, alongside competitors, reduced to a name, a line of text and, at most, a favicon. Google's own documentation confirms that a brand may be presented only by its favicon in the results, and that this favicon is not even guaranteed.

This is not a return calculation, and we will not present you one. It is a description of where you are seen. The first look, the one that forms in 50 milliseconds and then holds, happens increasingly on a favicon the size of a fingernail and on a line of text inside another company's screen. A logo designed for an A4 sheet has nothing to say in that place. And it is in that place that credibility is won or lost before any conversation.

What this evidence does not prove

This is the part that separates a line of reasoning from a brochure, and that is why we write it out in full.

None of these numbers measures your company's revenue. The car study measures perceived ownership in the United Kingdom, not money or your sector. The search data from SparkToro and Pew measure click behaviour, not brand recall or purchase, and it must be said that Google publicly disputed the Pew study, calling it flawed methodology. Different panels give different numbers: the same SparkToro, with another data provider, gave 58.5% of zero-click searches in 2024. The direction is robust, the decimals are not.

The cases we cited prove adoption, not return. That Google, Casa da Música or MIT have built adaptable brand systems shows the format was adopted by serious institutions. It does not show the format gave them profit, and we do not say so.

And what we said about AI is our reading, not a cited fact. Google's documentation confirms the favicon in search results, not in artificial intelligence answers. The idea that the brand loses control inside those answers is an argument we hold, presented as an argument. On dark mode, which also drives the design of identities, we give no number, because there is no general-population measurement we trust, only surveys of enthusiast audiences that overstate it.

The opposite temptation, and why it also fails

Faced with this, some conclude the opposite of the static: that everything should move, change shape, reinvent itself with each application. The sector itself feeds the idea. In a reference publication of the field, one reads that motion behaviour has become inseparable from brand identity and that brands now have to exist in a permanent state of motion. It is true of expression. It is not true of essence, and confusing the two is costly.

The clearest example is from the MIT Media Lab itself. In 2011, it adopted a generative identity with more than forty thousand possible variations, generated by algorithm, a different version for each person in the institution. Three years later, in 2014, Pentagram replaced that system with a single fixed monogram, ML. The same institution, in three years, retreated from tens of thousands of variations to a single symbol. The lesson is clean: a system that changes everything fixes nothing, and a brand no one can recognize twice is not flexible, it is invisible.

This is where the distinction we make runs, one a brand manual made for paper rarely makes: what stays the same, and what adapts. The core, the thing the eye locks in the first 50 milliseconds, is fixed and defensible. The expression, the way that core behaves on a favicon, on a dark background, in a vertical video or in an AI answer box, is what is designed to adapt without betraying itself. A dynamic identity is not artistic freedom. It is discipline: clear rules so that the brand keeps looking like itself in contexts no one showed it when it was designed.

What changes when you understand this

The practical conclusion is not to budget a new logo. It is to understand that your brand is no longer a file, it is a behaviour, and that the right question has stopped being whether it looks pretty and become whether it holds up in the places where it lives. We exist to convert perception into a financial asset, and perception, today, forms on small screens and machine surfaces your manual did not foresee. A brand recognized the same across all of them is an asset that defends itself. One that falls apart with each context is an expense that repeats.

Sources

Every number in this article was verified against the primary source. Where the source does not support the current reading, we say so in the body of the text.

1. Fiocchi, G. and Seyed Esfahani, M. (2024). Exploring the uniqueness of distinctive brand assets within the UK automotive industry. Journal of Brand Management, 31(1), 1-15. Source
2. Lindgaard, G., Fernandes, G., Dudek, C. e Brown, J. (2006). Attention web designers: You have 50 milliseconds to make a good first impression! Behaviour & Information Technology, 25(2), 115-126. Source
3. Google Design (2015). Evolving the Google Identity. Source
4. Cooper Hewitt, Smithsonian Design Museum. Graphic Identity: Casa da Música, Porto (Stefan Sagmeister, 2007). Source
5. TheGreenEyl / The, R. e Kang, E. R. (2011). MIT Media Lab Identity (generative system, more than 40,000 variations). Source
6. Pentagram / Bierut, M. e Fay, A. (2014). MIT Media Lab (redesign to a fixed monogram). Source
7. Fishkin, R. (2026). In 2026, Less than One Third of Google Searches Still Send a Click. SparkToro (Similarweb data). Source
8. Fishkin, R. (2024). 2024 Zero-Click Search Study. SparkToro (Datos data, a Semrush company). Source
9. Chapekis, A. e Lieb, A. (2025). Google users are less likely to click on links when an AI summary appears in the results. Pew Research Center. Source
10. TechCrunch (2025). Google's AI Overviews have 2B monthly users (Alphabet Q2 2025, Sundar Pichai). Source
11. StatCounter Global Stats (2026). Desktop vs Mobile vs Tablet Market Share Worldwide. Source
12. Google Search Central. Favicon in Google Search. Source
13. Obvious Agency (2025). The Numbers Behind Brand Investment (on the brand consistency survey by Lucidpress/Marq and Demand Metric). Source
14. Britton, J. (2025). How designers and brands are morphing the next era of motion design. It's Nice That. Source