Steps in creating a brand font: from brief to hinting
Brief: abstract wishes versus specific questions
There are two ways to compile a font brief, and they yield different results. The first one is to ask the client to describe the desired character of the font using words like 'modern', 'premium', 'friendly'. The problem is that different people imagine opposite things under the word 'modern', and the designer receives 5-6 rounds of revisions simply because the parties were speaking different languages.
The second approach is to ask specific, emotion- and context-bound questions. This method reduces the number of iterations and accelerates alignment on direction. Before starting work, the designer needs to know:
- In which media the font will mostly be used — packaging, signage, interface, large headline or small text in an app;
- Which languages and alphabets should be supported — Cyrillic, Latin, diacritics for specific markets;
- Whether there are references marked as 'like/dislike for a specific reason', not just a collection of attractive logos;
- What width of signage composition is needed at the start — a minimal test set or immediately a full alphabet with numbers and punctuation;
- Is a static font or a variable one needed, with interpolation between letterforms;
- What budget of time does the team have for approvals — this determines how many concepts to show simultaneously.
A good brief excludes the 'telephone effect' — distortion of the task during transfer between departments. It was precisely clear specifications and their absence that allowed TypeType to create O'STIN Sans in three months — this is a benchmark for timelines for a structured project, not a marketing slogan.
Conceptualization: one path or several parallel directions
The linear approach looks logical on paper: the designer draws one concept, shows it to the client, receives corrections, and refines it. In practice, it drags out the project because the client only sees alternatives after the first idea has already been partially confirmed and it's painful to discard it.
The working alternative is to show 2-3 conceptual directions at once, but not across the full alphabet, but on a limited set of 15-20 characters that cover the main forms of letters. This is exactly how the process was organized in the Citymobil project: three directions based on a limited set allowed the client to choose the final one before investing resources into scaling it up. After the choice, the direction is expanded to the full character set — with numbers, Cyrillic, diacritics, special characters.
Mastering: what it actually consists of
The word 'mastering' is often used in articles as a synonym for 'refinements,' without explaining its content. In practice, it consists of four sequential blocks of work, and you can't skip one — an error at the previous stage carries over into all subsequent ones.
- Visual mastering. Aligning all characters of the alphabet to a unified logic of construction — consistent optical stroke weight, coordinated proportions, a unified rhythm within-letter spacing.
- Technical audit. Validation of contours (removing self-intersections and unnecessary nodes), checking the correctness of the outlines and setting up OpenType features — ligatures, alternative characters, fractions.
- Manual kerning by system. Checking and editing of pairs of signs in four groups: UPPERCASE+UPPERCASE, UPPERCASE+lowercase, lowercase+lowercase, digits. This is not one operation, but a full grid of checks for hundreds of pairs.
- TrueType hinting. Rendering setup for specific platforms — Grayscale, ClearType, DirectWrite.
Professional studios automate part of this routine with Python scripts for file building and validation — this reduces post-production by 30% and more, but does not replace manual kerning checks: the algorithm catches technical errors, not aesthetic misalignments of pairs.
Spacing, kerning, tracking: different tasks that are confused with each other
These three terms are regularly used as synonyms, although they solve different tasks and work at different levels of font preparation.
| Parameter | What it regulates | When it is applied |
|---|---|---|
| Spacing (approximate) | Base side margin for each individual sign | At the start of technical preparation, before any paired settings |
| Kerning | Individual adjustment of spacing for specific problematic pairs of signs (for example, "AV", "To") | After the aproushes are set and the half-aproushes are checked |
| Tracking | Uniform increase or decrease of letter spacing throughout the entire text or block | Applied by the font user-designer, not the developer, at the level of layout |
Without proper spacing, kerning is meaningless — it will correct the symptom, not the cause of uneven rhythm in the line. Without kerning, individual letter pairs visually "stick together" or spread apart, even if the base aproushes are perfect. And tracking is no longer the developer's concern, but a designer's tool that uses the finished file in a layout. Confusing these three levels is a typical mistake in briefs, where the client asks to "adjust tracking," meaning the problem of kerning for specific pairs.
Hinting: why one font file renders differently
Hinting is not a final touch for a checkbox, but a separate technical discipline, because the same letter contour is physically rendered differently on different rendering systems.
| Technology | Where applied | Feature of rendering |
|---|---|---|
| Grayscale | Older versions of macOS, some PDF renderers | Grayscale smoothing without binding to color subpixels |
| ClearType | Windows, LCD matrix displays | Subpixel smoothing, requires separate hinting settings for horizontal strokes |
| DirectWrite | Modern Windows applications | Softer rendering without strict binding to the pixel grid, but sensitive to errors in contours |
If you skip setting up even one of the three technologies, some users will see letters with jagged strokes or uneven lines—despite the fact that the designer did not touch the contours, the problem is specifically with rendering. A separate headache is variable fonts: even after a successful export, interpolation between extreme weights (for example, Light and Black) can reveal errors in the contours that were not visible on static weights. Therefore, testing intermediate states is a mandatory, not optional, step.
Timelines and cost: what determines the price
According to data from type.today, the cost of developing a single font character ranges from 10 to 18 dollars. This is the basic mathematics that serves as the foundation for calculations: a set of 250-300 characters of Cyrillic and Latin scripts with digits and punctuation gives a range, and then the price increases or decreases depending on the following factors:
- Number of weights — one weight is cheaper than a family of 6-9 weights with variational interpolation;
- OpenType features complexity — ligatures and alternate characters require separate technical development;
- Number of conceptual directions at the start — parallel development of 2-3 variants increases the time required for the sketch stage;
- Brief quality — the more specific the questions are answered at the start, the fewer paid iterations of revisions are required.
As for timelines, the benchmark is three months for a font family with a clear brief, as in the case of O'STIN Sans. Without a structured brief and with constant changes in requirements during the process, the same volume of work can easily stretch to twice as long, and this is not the fault of the executor, but a direct consequence of the quality of task definition at the initial stage.
Brief: abstract wishes versus specific questions
There are two ways to compile a font brief, and they yield different results. The first one is to ask the client to describe the desired character of the font using words like 'modern', 'premium', 'friendly'. The problem is that different people imagine opposite things under the word 'modern', and the designer receives 5-6 rounds of revisions simply because the parties were speaking different languages.
The second approach is to ask specific, emotion- and context-bound questions. This method reduces the number of iterations and accelerates alignment on direction. Before starting work, the designer needs to know:
- In which media the font will mostly be used — packaging, signage, interface, large headline or small text in an app;
- Which languages and alphabets should be supported — Cyrillic, Latin, diacritics for specific markets;
- Whether there are references marked as 'like/dislike for a specific reason', not just a collection of attractive logos;
- What width of signage composition is needed at the start — a minimal test set or immediately a full alphabet with numbers and punctuation;
- Is a static font or a variable one needed, with interpolation between letterforms;
- What budget of time does the team have for approvals — this determines how many concepts to show simultaneously.
A good brief excludes the 'telephone effect' — distortion of the task during transfer between departments. It was precisely clear specifications and their absence that allowed TypeType to create O'STIN Sans in three months — this is a benchmark for timelines for a structured project, not a marketing slogan.
Conceptualization: one path or several parallel directions
The linear approach looks logical on paper: the designer draws one concept, shows it to the client, receives corrections, and refines it. In practice, it drags out the project because the client only sees alternatives after the first idea has already been partially confirmed and it's painful to discard it.
The working alternative is to show 2-3 conceptual directions at once, but not across the full alphabet, but on a limited set of 15-20 characters that cover the main forms of letters. This is exactly how the process was organized in the Citymobil project: three directions based on a limited set allowed the client to choose the final one before investing resources into scaling it up. After the choice, the direction is expanded to the full character set — with numbers, Cyrillic, diacritics, special characters.
Mastering: what it actually consists of
The word 'mastering' is often used in articles as a synonym for 'refinements,' without explaining its content. In practice, it consists of four sequential blocks of work, and you can't skip one — an error at the previous stage carries over into all subsequent ones.
- Visual mastering. Aligning all characters of the alphabet to a unified logic of construction — consistent optical stroke weight, coordinated proportions, a unified rhythm within-letter spacing.
- Technical audit. Validation of contours (removing self-intersections and unnecessary nodes), checking the correctness of the outlines and setting up OpenType features — ligatures, alternative characters, fractions.
- Manual kerning by system. Checking and editing of pairs of signs in four groups: UPPERCASE+UPPERCASE, UPPERCASE+lowercase, lowercase+lowercase, digits. This is not one operation, but a full grid of checks for hundreds of pairs.
- TrueType hinting. Rendering setup for specific platforms — Grayscale, ClearType, DirectWrite.
Professional studios automate part of this routine with Python scripts for file building and validation — this reduces post-production by 30% and more, but does not replace manual kerning checks: the algorithm catches technical errors, not aesthetic misalignments of pairs.
Spacing, kerning, tracking: different tasks that are confused with each other
These three terms are regularly used as synonyms, although they solve different tasks and work at different levels of font preparation.
| Parameter | What it regulates | When it is applied |
|---|---|---|
| Spacing (approximate) | Base side margin for each individual sign | At the start of technical preparation, before any paired settings |
| Kerning | Individual adjustment of spacing for specific problematic pairs of signs (for example, "AV", "To") | After the aproushes are set and the half-aproushes are checked |
| Tracking | Uniform increase or decrease of letter spacing throughout the entire text or block | Applied by the font user-designer, not the developer, at the level of layout |
Without proper spacing, kerning is meaningless — it will correct the symptom, not the cause of uneven rhythm in the line. Without kerning, individual letter pairs visually "stick together" or spread apart, even if the base aproushes are perfect. And tracking is no longer the developer's concern, but a designer's tool that uses the finished file in a layout. Confusing these three levels is a typical mistake in briefs, where the client asks to "adjust tracking," meaning the problem of kerning for specific pairs.
Hinting: why one font file renders differently
Hinting is not a final touch for a checkbox, but a separate technical discipline, because the same letter contour is physically rendered differently on different rendering systems.
| Technology | Where applied | Feature of rendering |
|---|---|---|
| Grayscale | Older versions of macOS, some PDF renderers | Grayscale smoothing without binding to color subpixels |
| ClearType | Windows, LCD matrix displays | Subpixel smoothing, requires separate hinting settings for horizontal strokes |
| DirectWrite | Modern Windows applications | Softer rendering without strict binding to the pixel grid, but sensitive to errors in contours |
If you skip setting up even one of the three technologies, some users will see letters with jagged strokes or uneven lines—despite the fact that the designer did not touch the contours, the problem is specifically with rendering. A separate headache is variable fonts: even after a successful export, interpolation between extreme weights (for example, Light and Black) can reveal errors in the contours that were not visible on static weights. Therefore, testing intermediate states is a mandatory, not optional, step.
Timelines and cost: what determines the price
According to data from type.today, the cost of developing a single font character ranges from 10 to 18 dollars. This is the basic mathematics that serves as the foundation for calculations: a set of 250-300 characters of Cyrillic and Latin scripts with digits and punctuation gives a range, and then the price increases or decreases depending on the following factors:
- Number of weights — one weight is cheaper than a family of 6-9 weights with variational interpolation;
- OpenType features complexity — ligatures and alternate characters require separate technical development;
- Number of conceptual directions at the start — parallel development of 2-3 variants increases the time required for the sketch stage;
- Brief quality — the more specific the questions are answered at the start, the fewer paid iterations of revisions are required.
As for timelines, the benchmark is three months for a font family with a clear brief, as in the case of O'STIN Sans. Without a structured brief and with constant changes in requirements during the process, the same volume of work can easily stretch to twice as long, and this is not the fault of the executor, but a direct consequence of the quality of task definition at the initial stage.
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
