How we actually choose a font pairing

Typeface pairing is the part of an identity project clients expect to be pure taste, and it's the part we've systematized the most. After the Solstice project taught us what a week of unplanned type exploration costs, we wrote down the process we now run every time. It has four steps, and the first two have nothing to do with how a typeface looks on a poster.

Start from the logo mark, not the mood board

A pairing that flatters the mood board but fights the logo mark will lose every real-world layout it appears in. So we set candidates directly against the approved mark at three sizes — business card, browser tab, storefront — before we judge anything else. Roughly a third of candidates drop out here, usually because their geometry argues with the mark's construction in ways a specimen page never shows.

Two tests that reject most of the rest

The surviving candidates then face two deliberately unglamorous tests:

• The 14px screen test. We set two paragraphs of real client copy — never lorem ipsum — at 14px on a mid-range Android phone. If the body face needs 16px and generous letter-spacing to stay readable, it fails, no matter how good the specimen looks printed.
• The all-caps label test. Navigation, buttons, form labels, and packaging side-panels all end up in small caps or all-caps eventually. Faces with a single case style, tight apertures, or ambiguous I/l/1 shapes fail here. This test rejects more "beautiful" typefaces than any other step we run.

Pair for contrast of role, not contrast of style

The pairings that survive both tests get combined by role: one voice for headlines, one workhorse for everything else. We want the two faces to do different jobs, not merely look different — a display face with real personality over a text face that disappears. When both faces compete for attention, every layout becomes a negotiation, and layouts that need negotiating don't survive handoff to a client's in-house team.

When to stop looking

We cap explorations at 14 pairings — the number we hit on Solstice before realizing the extra candidates were re-litigating decisions the tests had already made. If nothing survives after 14, the problem is almost never the font library; it's the brief, and we go back one step instead of downloading more typefaces.

If you're scoping an identity project and want to see this process applied to your brand, our brand identity service includes the full type exploration with every rejected candidate documented, so you see why the final system won.