Every few months a founder tells a designer, “We need a logo that captures who we are.” It sounds sensible. It is also the opening line of a long, expensive misunderstanding. A logo cannot capture who you are. It can only announce that you have arrived.

Think of it as a doorbell. Nobody falls in love with a house because of the doorbell. But once you love the people inside, that two-note chime starts to mean dinner and the friend who always brings dessert. The bell did none of that work.

Logo design is the craft of creating a compact visual mark, whether a symbol, a wordmark or both, that identifies an organisation at a glance and, over time, triggers everything people already know about it.

Notice what that definition leaves out. Explaining. Selling. The logo does neither, and asking it to is how good money gets spent on bad marks.

The $35 doorbell

In 1971 a Portland State University graphic design student named Carolyn Davidson billed Phil Knight $35 for a mark he wanted to suggest speed. His verdict has become folklore: “Well, I don’t love it, but maybe it will grow on me.”

He was right, though not for the reason most case studies imply. The swoosh did not grow on anyone by itself. It grew because it appeared on the Nike Cortez in 1972, and then on half a century of shoes and athletes that gave it something to stand for.

The mark was a hook. Nike spent the next fifty years hanging meaning on it.

The story usually gets told as proof that a great logo can be cheap. The more useful lesson: meaning flows from the company to the logo, almost never in reverse. A mediocre company drags even a beautiful mark down to its level; a brilliant one makes a plain shape look inevitable.

Davidson was thanked properly in 1983 with a diamond-set gold swoosh ring and 500 Nike shares. Lovely. Those shares grew because of the shoes, not the curve.

What should a logo actually say about my business?

This is the question small-business owners ask most, usually smuggled into a brief that wants the mark to show the product, the founder’s values, the heritage, the city, the future and ideally a cheerful animal. Designers swap war stories about clients who spend an hour listing everything the logo must contain, then balk at the one simple idea that might work.

The honest answer is: almost nothing.

Consider Mitsubishi. Its three diamonds echo the three-leaf crest of the Tosa clan, founder Yataro Iwasaki’s first employer, and the stacked rhombuses of the Iwasaki family crest. The name joins mitsu (three) with hishi, the word for water chestnut that Japanese has long used for a diamond shape.

Nothing in the mark mentions cars, banks, ships or air conditioners. It doesn’t need to. The symbol identifies; the businesses explain.

Or take Shell, an energy giant whose emblem is a seashell. The company swapped an earlier mussel for a scallop in 1904, and the version Raymond Loewy redrew in 1971 is still in use. Nobody pulls into a forecourt believing petrol comes from molluscs. The shape works because it is simple and old, and you cannot buy old on launch day.

A logo that tries to tell your whole story ends up telling none of it, at roughly the size of a fingernail on a phone screen. Pick one idea, or no literal idea at all, and let the business do the talking.

What a logo cannot do, however much you pay

It cannot fix your product. It cannot rescue a muddled strategy, and it will not persuade anyone to trust a company that keeps letting them down.

Picture a café in Joo Chiat with patchy coffee and slow service. The owner, convinced the problem is “branding”, commissions a gorgeous new mark with hand-drawn lettering and a tiny shophouse.

Six months later the coffee is still patchy, and the new logo sits on cups that remind regulars of everything they dislike. The doorbell got an upgrade. The house still smells of burnt milk.

The cautionary tale every designer knows is Pepsi in 2008. The Arnell Group reworked the familiar red, white and blue globe, re-angling the white band through its middle, for a fee widely reported at $1 million. Then a 27-page document attributed to the agency surfaced online, tying the new mark to the golden ratio, the Mona Lisa and something it called the gravitational pull of Pepsi. Genuine or not, it became one of the most ridiculed documents in design history.

The circle itself was competent. The problem was the belief that a soft drink logo could carry cosmic significance.

Overclaiming is logo design’s original sin, and clients encourage it, because it is comforting to think one image can fix a business. It can’t.

Why do people get so angry about logo redesigns?

Because, for your most committed customers, the logo stopped being your property years ago. It became theirs.

That loyalty cuts in a surprising direction. People who feel strongly attached to a brand tend to dislike a redesign, and their feelings about the whole company can sour with it. Casual customers, meanwhile, often welcome the refresh. The very people you most want to keep are the ones most likely to be upset.

Gap learned this in October 2010. It swapped its navy box for a plain wordmark with a small blue square perched above the “p”. Customers erupted on Facebook and Twitter.

The company first defended the design, then invited the public to suggest improvements, which is roughly like asking a mob to help redecorate. About a week after the new logo appeared, the blue box was back, along with an admission from Gap’s North America president that the process had been handled badly.

Cracker Barrel replayed the lesson in August 2025, at higher volume. The restaurant chain dropped its “Old Timer”, a man seated in a chair with one arm on a barrel, in favour of plainer lettering. After days of fury that reached as far as the White House, it announced that the old timer would stay.

Is the outrage ever right?

Sometimes. More often it is just loud.

When London unveiled its 2012 Olympic logo in June 2007, a jagged stack of numerals by Wolff Olins with a reported £400,000 price tag, the press savaged it and more than 28,000 people signed a petition against it. Sebastian Coe asked for patience, insisting it was a brand rather than a logo. Five years later it carried the Games without being changed, and some of its harshest critics had come round to the idea, if not to the shapes.

Airbnb’s Bélo followed the same arc in July 2014. Within a day the internet had found rude anatomy in it, a bear’s nose, and an awkward resemblance to the mark of a software firm, Automation Anywhere; the two companies called it a coincidence. Today most people simply read it as Airbnb.

So the curve runs like this. Outrage peaks in week one and fades once the mark is attached to enough real experiences. The new doorbell sounds strange for a while. Then it sounds like home.

The exceptions are changes made carelessly, like Gap’s, or changes that seem to tell loyal customers they no longer matter. Then the logo becomes a lightning rod for anxieties that were already there.

When should you change the doorbell?

Rarely. And usually by taking something away.

In 2011, its 40th anniversary year, Starbucks dropped the words “Starbucks Coffee” and the ring that framed its siren, leaving the green siren on her own. Nike had made a similar move in 1995, retiring the capitalised NIKE wordmark from its official logo.

Both changes were earned. After decades of exposure, the symbol could stand alone. Subtraction is a privilege of the famous; a start-up that removes its name from its logo isn’t being confident, it’s being anonymous.

There are good reasons to redesign: a mark that breaks at small sizes, a merger, a legal problem. Tokyo 2020 is the starkest example of the legal kind. Kenjiro Sano’s emblem was unveiled on 24 July 2015 and dropped on 1 September, after a Belgian designer argued it too closely resembled his 2013 logo for the Théâtre de Liège. Sano denied copying, but the mark was finished.

Its replacement, picked through a nationwide contest and unveiled in April 2016, was Asao Tokolo’s indigo chequered ring, drawn from ichimatsu moyo, a pattern popular in Edo-period Japan. It looked rooted rather than borrowed.

Asao Tokolo’s chequered emblems drew on an Edo-period pattern, so they looked rooted rather than borrowed.

Then there is the change that isn’t really a redesign. When Twitter’s bird, known as Larry, gave way to a black X in July 2023, nobody was refreshing a logo. The new owner, Elon Musk, who had bought the company for $44 billion the previous October, was announcing that the old company was gone. Judge that one on business results, not on the letterform.

How to judge a logo without asking your mother-in-law

Personal taste is the worst instrument for judging a logo, and the one everyone reaches for first. Try these tests instead.

Shrink it. View it at favicon size and inside a round WhatsApp profile picture. If it turns to mush, it fails.

Strip the colour. Print it in black on a thermal receipt. A strong mark survives one colour and a bad printer.

Line it up with the competition. Put it beside your five closest rivals. If you could swap the names around without anyone noticing, you have designed camouflage.

Make sure it’s yours. Run trademark searches before anyone falls in love. Tokyo proved that a resemblance, even an accidental one, can kill a mark that has already launched.

Picture it in twenty years. Mitsubishi’s diamonds and Shell’s scallop have lasted because they are too simple to go out of fashion. Trendy gradients age like milk.

Look for a root. The strongest marks often grow from something specific to the organisation. Expo 2020 Dubai’s emblem, unveiled in March 2016, was modelled on a finely worked gold ring found at a 4,000-year-old archaeological site in Dubai’s desert. It gave a global event a local anchor without a single literal skyline.

Notice that “Do I like it?” isn’t on the list. In week one, nobody likes a new logo, sometimes including the person who signed it off.

Brief the house, not the bell

If the logo is the doorbell, the brief should describe the house. Before anyone talks about shapes, settle what the business stands for, who it serves and how it differs from the alternatives. That is brand strategy. Done properly, it makes the logo brief shorter, not longer.

Then budget for what you are actually buying. Owners are often baffled that logo prices run from less than lunch to more than a car. The gap is mostly scope: a drawing, versus a mark built on research, tested at every size, cleared legally and delivered with rules for using it everywhere. A logo without that system is a doorbell with no wiring.

Finally, plan for the reaction. Show your own staff first. Tell your most loyal customers why before you tell everyone else. Expect a wobble, don’t panic in week one, and whatever you do, don’t crowdsource your way out of it.

Before you ring

Is a logo the same as a brand?
No. A logo is a visual mark that identifies an organisation, while a brand is the sum of what people think, feel and expect about it. The logo triggers the brand; it cannot replace it.

How long does it take for people to accept a new logo?
Backlash is usually loudest in the first days and fades as the mark becomes tied to real experiences. London 2012’s logo was mocked at launch in 2007, yet it went on to serve the Games unchanged in 2012.

When should a business redesign its logo?
When the mark fails practically, for instance it breaks at small sizes or can’t be legally protected, after a merger, or when the business has genuinely changed. Keeping up with design fashion is rarely reason enough.

If you’d rather design the house before choosing the doorbell, start with a proper brand conversation.