Branding typography: how to choose a font pair
Contrast, complementarity, coherence: three approaches in comparison
Most guides stop at the advice to 'take serif and sans-serif' — and this works in about thirty percent of cases, no more. In reality, a pair works only if all three principles are followed simultaneously, not separately.
| Principle | What is checked | How a violation looks like |
|---|---|---|
| Contrast | Fonts are visually distinct — in weight, width, and style | Two similar medium-weight grotesques — the eye cannot distinguish where the heading ends and the text begins |
| Complementarity | Fonts share the same era, proportions, or emotional tone | Strict geometric grotesk + handwritten decorative font — a conflict of mood |
| Coherence | The pair establishes hierarchy: heading, subheading, text are read as a system | The heading font 'pulls' attention, while the body text gets lost in its background |
An important nuance that is rarely discussed: a sixteenth-century serif typeface and a twentieth-century sans serif can work together just as well as two contemporary typefaces if their proportions and height ratios match. The difference in when they were created is no obstacle when the structural parameters of the letters align.
What parameters of letters to look at when selecting
Readability is more important than recognizability — this applies primarily to text fonts. For the main font, you need at least two variations: Regular and Bold; without them, the pair won't withstand real layout with accents.
- x-height — the height of lowercase letters. In a pair, it should be approximately the same: if one font visually appears 'higher' than the other at the same point size, the text will look uneven on a single line.
- Contrast of strokes — the difference between thick and thin elements of a letter. A font with strong contrast (antiqua) paired with a monolinear grotesque creates the necessary tonal difference, but if both fonts are contrastive, they start to compete.
- Proportions — the width of letters relative to their height. A narrow headline font and a wide body font work if the difference is intentional, not accidental.
- Aperture and the shape of the light fixture letter — the openness of such letters as 'c', 'e', 'a'. Fonts built on the same logic have similar apertures, and this is subconsciously perceived as 'similarity'.
How to test a font pair before implementing it in a brand
Here begins what almost nobody describes specifically. Testing fonts 'by eye' in a layout with ideal lighting is a bad idea: the real medium is almost always worse.
- Create a test layout not with lorem ipsum, but with the brand's actual text — a slogan, product name, paragraph from the future website. Abstract text hides problems with specific Cyrillic letter combinations.
- Check the body text at 12–16px size — this is the minimum size at which the font should remain readable without losing details. If at this size letters 'merge', the font is not suitable for body text.
- Compare the x-height of both typefaces at the same point size — place words side by side on one line and check whether the height of lowercase letters matches visually, not by numbers in the specification.
- Test the pair on different backgrounds: white, dark, brand's official color. Contrast that was readable on white may get lost on a colored background.
- Render the layout on actual media — print on paper, smartphone screen in daylight, favicon 16×16px, sign in scale. A monitor in a designer's office is not a representative sample.
- Make a prototype with a real hierarchy: heading, subheading, text, caption — and check whether the structure is readable within 2–3 seconds of viewing, without special analysis.
For this approach, tools such as Figma and Adobe XD with ready-made text styles work well, as well as a regular browser — insert fonts via @font-face and view the page on different devices through the inspector. Print mockups are made in any graphic editor with output to a real printer, which is cheaper than it seems and saves months of revisions after launch.
Licenses for corporate fonts: what to buy and where
This is a topic that almost nobody examines in detail, yet it costs more than it appears at the start of a project. The difference between licenses is not formalities but a real legal risk for the brand.
| License type | What it permits | Where to get them |
|---|---|---|
| Open source (SIL OFL 1.1) | Free use, including commercial, without separate payment | Google Fonts — over 1500 fonts with open licenses |
| Commercial (desktop/web) | Use in a limited number of products, often with traffic or device limits | Adobe Fonts (150 000 fonts for Creative Cloud subscribers), MyFonts, foundry websites |
| Corporate (Enterprise/Brand license) | Use on unlimited number of media, editing, embedding in apps and packaging | Direct negotiations with foundries — Monotype, Linotype, Fontsmith, Dalton Maag |
For corporate branding, a corporate license is almost always required — a regular web license restricts website traffic or the number of installations, and when scaling the brand, additional rights will have to be purchased. It is also necessary to check whether the font supports the OpenType format — it provides cross-platform compatibility, automatic ligatures, and proper operation on any operating system, unlike outdated TrueType files.
Cyrillic script, multilingualism, and mobile devices
This is where teams often lose money: the team approves a pair in Latin script, but then it turns out that the Cyrillic version of the font was drawn formally, with letters like 'yo', 'y', and 'hard sign' looking foreign in relation to the rest of the alphabet.
- Check the letters 'yo', 'y', 'hard sign', and 'shch' separately — these are the most problematic Cyrillic characters; they are often added later to the main set and not always done with quality.
- If a brand operates in multiple countries, look at diacritic support for the necessary languages. SST supports 93 languages, Avenir Next World and Neue Frutiger — more than 150, which are reasonable benchmarks for international brands.
- Check the pair on mobile screens separately from desktop — pixel density and small type render differently, and the contrast of strokes may appear differently, with fine details disappearing during rendering.
- Test with real long words of the language — Russian text is on average 15–20% longer than English, and what looked beautiful in the layout on English may be distorted in Cyrillic.
Common mistakes when choosing a pair
The most common mistake is chasing a trendy font instead of a classic one. Corporate typography should remain relevant for at least 5 years, while trendy typefaces become outdated within 1–2 years and pull the entire visual brand image into the past.
- Two similar grotesks of medium weight in one pair — the eye does not see the difference, and text hierarchy collapses already on the second screen of the website.
- Decorative script in headings with a strict corporate grotesk in the body text — an emotional disconnect that readers feel even if they don’t understand why.
- Choosing a font only by the appearance of the heading without checking the Bold and Italic styles — in practice, these styles will be needed in the first week of website layout.
- Ignoring the Cyrillic version of the font when choosing an international typeface — a beautiful Latin font may have a weak Cyrillic version with a different weight and width perception.
- Purchasing a web license without checking traffic limit — as the site grows, the license simply stops covering the usage, and this becomes apparent at the worst possible moment.
The logic of selecting a font pair should not be based on a ready-made list of 'trendy combinations,' but rather on the task of a specific brand: a techno-company benefits from a grotesk pairing with difference in weight and width, whereas a brand with history and artisanal character suits a pairing of old-style and grotesk fonts, where x-height and letter proportions match. There are no universal pairs; there are only pairs tested on real text, real size, and real medium.
Contrast, complementarity, coherence: three approaches in comparison
Most guides stop at the advice to 'take serif and sans-serif' — and this works in about thirty percent of cases, no more. In reality, a pair works only if all three principles are followed simultaneously, not separately.
| Principle | What is checked | How a violation looks like |
|---|---|---|
| Contrast | Fonts are visually distinct — in weight, width, and style | Two similar medium-weight grotesques — the eye cannot distinguish where the heading ends and the text begins |
| Complementarity | Fonts share the same era, proportions, or emotional tone | Strict geometric grotesk + handwritten decorative font — a conflict of mood |
| Coherence | The pair establishes hierarchy: heading, subheading, text are read as a system | The heading font 'pulls' attention, while the body text gets lost in its background |
An important nuance that is rarely discussed: a sixteenth-century serif typeface and a twentieth-century sans serif can work together just as well as two contemporary typefaces if their proportions and height ratios match. The difference in when they were created is no obstacle when the structural parameters of the letters align.
What parameters of letters to look at when selecting
Readability is more important than recognizability — this applies primarily to text fonts. For the main font, you need at least two variations: Regular and Bold; without them, the pair won't withstand real layout with accents.
- x-height — the height of lowercase letters. In a pair, it should be approximately the same: if one font visually appears 'higher' than the other at the same point size, the text will look uneven on a single line.
- Contrast of strokes — the difference between thick and thin elements of a letter. A font with strong contrast (antiqua) paired with a monolinear grotesque creates the necessary tonal difference, but if both fonts are contrastive, they start to compete.
- Proportions — the width of letters relative to their height. A narrow headline font and a wide body font work if the difference is intentional, not accidental.
- Aperture and the shape of the light fixture letter — the openness of such letters as 'c', 'e', 'a'. Fonts built on the same logic have similar apertures, and this is subconsciously perceived as 'similarity'.
How to test a font pair before implementing it in a brand
Here begins what almost nobody describes specifically. Testing fonts 'by eye' in a layout with ideal lighting is a bad idea: the real medium is almost always worse.
- Create a test layout not with lorem ipsum, but with the brand's actual text — a slogan, product name, paragraph from the future website. Abstract text hides problems with specific Cyrillic letter combinations.
- Check the body text at 12–16px size — this is the minimum size at which the font should remain readable without losing details. If at this size letters 'merge', the font is not suitable for body text.
- Compare the x-height of both typefaces at the same point size — place words side by side on one line and check whether the height of lowercase letters matches visually, not by numbers in the specification.
- Test the pair on different backgrounds: white, dark, brand's official color. Contrast that was readable on white may get lost on a colored background.
- Render the layout on actual media — print on paper, smartphone screen in daylight, favicon 16×16px, sign in scale. A monitor in a designer's office is not a representative sample.
- Make a prototype with a real hierarchy: heading, subheading, text, caption — and check whether the structure is readable within 2–3 seconds of viewing, without special analysis.
For this approach, tools such as Figma and Adobe XD with ready-made text styles work well, as well as a regular browser — insert fonts via @font-face and view the page on different devices through the inspector. Print mockups are made in any graphic editor with output to a real printer, which is cheaper than it seems and saves months of revisions after launch.
Licenses for corporate fonts: what to buy and where
This is a topic that almost nobody examines in detail, yet it costs more than it appears at the start of a project. The difference between licenses is not formalities but a real legal risk for the brand.
| License type | What it permits | Where to get them |
|---|---|---|
| Open source (SIL OFL 1.1) | Free use, including commercial, without separate payment | Google Fonts — over 1500 fonts with open licenses |
| Commercial (desktop/web) | Use in a limited number of products, often with traffic or device limits | Adobe Fonts (150 000 fonts for Creative Cloud subscribers), MyFonts, foundry websites |
| Corporate (Enterprise/Brand license) | Use on unlimited number of media, editing, embedding in apps and packaging | Direct negotiations with foundries — Monotype, Linotype, Fontsmith, Dalton Maag |
For corporate branding, a corporate license is almost always required — a regular web license restricts website traffic or the number of installations, and when scaling the brand, additional rights will have to be purchased. It is also necessary to check whether the font supports the OpenType format — it provides cross-platform compatibility, automatic ligatures, and proper operation on any operating system, unlike outdated TrueType files.
Cyrillic script, multilingualism, and mobile devices
This is where teams often lose money: the team approves a pair in Latin script, but then it turns out that the Cyrillic version of the font was drawn formally, with letters like 'yo', 'y', and 'hard sign' looking foreign in relation to the rest of the alphabet.
- Check the letters 'yo', 'y', 'hard sign', and 'shch' separately — these are the most problematic Cyrillic characters; they are often added later to the main set and not always done with quality.
- If a brand operates in multiple countries, look at diacritic support for the necessary languages. SST supports 93 languages, Avenir Next World and Neue Frutiger — more than 150, which are reasonable benchmarks for international brands.
- Check the pair on mobile screens separately from desktop — pixel density and small type render differently, and the contrast of strokes may appear differently, with fine details disappearing during rendering.
- Test with real long words of the language — Russian text is on average 15–20% longer than English, and what looked beautiful in the layout on English may be distorted in Cyrillic.
Common mistakes when choosing a pair
The most common mistake is chasing a trendy font instead of a classic one. Corporate typography should remain relevant for at least 5 years, while trendy typefaces become outdated within 1–2 years and pull the entire visual brand image into the past.
- Two similar grotesks of medium weight in one pair — the eye does not see the difference, and text hierarchy collapses already on the second screen of the website.
- Decorative script in headings with a strict corporate grotesk in the body text — an emotional disconnect that readers feel even if they don’t understand why.
- Choosing a font only by the appearance of the heading without checking the Bold and Italic styles — in practice, these styles will be needed in the first week of website layout.
- Ignoring the Cyrillic version of the font when choosing an international typeface — a beautiful Latin font may have a weak Cyrillic version with a different weight and width perception.
- Purchasing a web license without checking traffic limit — as the site grows, the license simply stops covering the usage, and this becomes apparent at the worst possible moment.
The logic of selecting a font pair should not be based on a ready-made list of 'trendy combinations,' but rather on the task of a specific brand: a techno-company benefits from a grotesk pairing with difference in weight and width, whereas a brand with history and artisanal character suits a pairing of old-style and grotesk fonts, where x-height and letter proportions match. There are no universal pairs; there are only pairs tested on real text, real size, and real medium.
Recent articles
MegaFon, BMW, VTB and more 240+ companies. See how it looks in the showcase reel:
Our branding agency is approached to order:
Branding agency services
