Most logo wordmarks ship with default kerning and hope for the best. The few that don't — FedEx, IBM, Mobil — hold up for decades because someone moved every letter pair by hand. The skill is going away faster than people notice.
If you look at the FedEx wordmark long enough you'll see the arrow hidden between the E and the x. That arrow is famous, talked about every other year in some "hidden details in famous logos" listicle, and it's deserved attention. What gets less attention is the fact that the arrow only works because every letter pair in the word FedEx was kerned by hand to within a thousandth of an inch. The arrow exists because the negative space between the E and the x was sized exactly right, and that didn't happen by accident.
The wordmark was designed by Lindon Leader at Landor in 1994. Leader hand-kerned it. Custom drew certain glyphs to make the spacing work. Tested it printed at every size from a business card to the side of a truck. Adjusted some terminals to keep the arrow legible. None of that is a step you find in a brand-identity tutorial in 2026. Most of what brand designers learn now stops at "pick the typeface and set the word in it."
Hand-kerning a wordmark — the practice of moving every letter pair manually until the spacing reads as visually even rather than mathematically even — is a dying skill. The reasons it's dying are interesting. The cost of letting it die is real.
The kerning baked into a typeface — the kerning the type designer drew — is good for body text. It assumes the typeface will be set at a normal reading size, in long passages, with the eye sweeping past each letter in milliseconds. The default kerning is a compromise that has to work everywhere.
A logo is the opposite of normal reading. A logo is set huge. The eye spends time on it. Every letter is being looked at directly, often in isolation. The compromise that works for a paragraph at 11pt does not work for a five-letter word at 200pt on the side of a building.
Hand-kerning a wordmark means going through every adjacent letter pair and asking: does the white space between these two letters feel like the same amount as the white space between every other pair? Not measured. Felt. The eye is doing the math, not a tool. Some pairs need negative kerning because their shapes interlock. Some pairs need positive kerning because their shapes flee from each other. Some pairs need a custom alteration to one of the letters so the spacing works at all.
IBM (Paul Rand, 1972). Three letters, eight horizontal stripes, and one of the most-imitated identities in history. The stripes are designed to read as IBM at any size. The kerning between the I, B, and M is so tight that the negative space between letters is the same width as the negative space between stripes within letters. Look at it once with that observation in mind and you can't unsee it. The wordmark reads as a unified mark rather than three separate letters because of that decision.
Mobil (Tom Geismar at Chermayeff & Geismar, 1965). The famous round-O Mobil mark. The o is custom-drawn to be a perfect circle, slightly larger than the cap height of the other letters so it reads as the same height optically. The l is hand-kerned in close to the i. The whole word reads as a single round, friendly word rather than five letters set in a row. Try setting Mobil in any typeface in the world without hand-kerning. It looks like a stock photo of the word.
Spotify (Christian Wilsson and team, 2015 update). The current Spotify wordmark uses a custom typeface (Spotify Circular, drawn by Lineto) and even with that custom drawing, the wordmark itself was hand-spaced. The S has a slightly tighter relationship to the p than the p does to the o. The terminals were adjusted. None of that ships in the typeface itself — it's specific to the wordmark.
Two forces, mostly.
The first is the rise of "good enough" typefaces. Modern typefaces ship with much better kerning than typefaces did in 1990. A designer setting a wordmark in a 2024-vintage typeface from a top foundry can leave the kerning at default and get a result that's, honestly, fine. It won't be great. It won't have the unified-shape quality of a hand-kerned mark. But it'll pass internal review. And once internal review is the bar, hand-kerning becomes a thing nobody asks for.
The second is the speed of the modern brand-design process. Wordmarks used to take months. The designer sketched, kerned, printed, kerned again, walked away for a week, came back, kerned again. Now wordmarks get pumped out in two-week sprints because the agency is also designing the brand book, the social templates, the launch animation, and the email templates in the same engagement. Hand-kerning is the first thing to get cut when the timeline tightens. It's not visible to clients in the way the type choice is visible.
The tools have gotten worse for hand-kerning too, in a small way. Figma's kerning controls are not as precise as the kerning controls in Illustrator or FontLab. The keyboard shortcut for nudging a letter by 1 unit takes more keystrokes. None of these are dealbreakers. But they're friction, and friction in a tool you use 80 hours a week is the difference between doing something and not doing it.
Default-kerned wordmarks have a quality you start to spot once you're looking. The word reads as a string of letters rather than as a single shape. The eye picks up the rhythm of the typeface (which is a paragraph rhythm, not a logo rhythm). The spacing between certain pairs — usually around the V, A, T, W, and the round letters — looks slightly looser than the spacing around the straight verticals. The wordmark looks like the typeface was set in it, which is the opposite of what a wordmark is for.
The cost is not catastrophic. A default-kerned wordmark is not a broken wordmark. But it has a shorter useful life. Brands with hand-kerned wordmarks tend to survive type-revival cycles, design-trend cycles, and aesthetic shifts because the wordmark is a unified mark, not a typesetting choice. Brands with default-kerned wordmarks tend to look outdated faster, because every shift in default-kerning conventions makes them look subtly off.
If you are designing a wordmark and you do not normally hand-kern, three things will get you most of the way there:
One: print the wordmark at multiple sizes. Wordmark on a business card, wordmark on a letterhead, wordmark blown up to 30 inches across. Pin them to a wall. Walk away. Come back the next morning. The pairs that look wrong will jump out at you in a way they don't on screen.
Two: make the spacing visible. Drop a thin colored rectangle into every gap between letters. Make the rectangles all the same color. Look at the rectangles, not the letters. They should all look like the same amount of color. If some rectangles look bigger or smaller than others, the kerning needs work.
Three: kern with your keyboard, not your mouse. Click into the word, set kerning to optical first (which is better than metric in most cases), then nudge each pair manually with the keyboard kerning shortcuts. Optical is a starting point, not an end point. The eye still has to confirm.
None of this takes special skill. It takes patience and the willingness to do something slow that nobody is going to thank you for. The thanking comes later, when the wordmark has lasted twenty years and still looks like a single deliberate object instead of five letters typed into a font.
FedEx, IBM, Mobil, and a few other identities have lasted because someone moved every letter pair by hand and looked at the result and moved them again. That work was invisible to everyone except the designer. The designer probably didn't get a bonus for it. The client probably didn't notice. But the wordmark has spent the last fifty years not embarrassing itself, in every size, on every surface, in every era. That is the return on hand-kerning.
The next wordmark you design — and the one after that — will probably go through default kerning, get approved, and ship. Sometimes that's the right call. The brief is small, the timeline is tight, the wordmark will be replaced in three years anyway. But occasionally — once or twice a year — you'll get a wordmark that you suspect will outlive everything else you make. That's the wordmark to hand-kern. The skill isn't dying because nobody can do it. It's dying because nobody is asking. Be the one who asks.
Be the first to mark this page