Services · Design-stage reviews
A Brand Palette Review Settles the Contrast Once, for Every Page
For the brand owner picking colors and type once, who wants every page built on them to start out passing rather than get repainted later. The whole kit gets tested pair by pair, and each failing pair comes back with the smallest shade shift that clears it.
What We Keep Finding
Brand palettes get chosen for feeling, and contrast never gets a vote. Then the kit ships, and the one light gray that felt so calm in the brand deck becomes the body text on every page you will ever publish. A palette problem is never one finding. It is the same finding everywhere, forever, until the kit itself changes. Which is also why it pays to fix at this level. Low contrast is the most common failure a machine finds anywhere on the web, sitting on 83.9% of home pages surveyed and turning up an average of 34 times on each one.
There are three thresholds, and most kits get built against one. Body text needs 4.5:1. Large text drops to 3:1, and large is defined in points rather than pixels, at 18 point or 14 point bold, which lands near 18.5px and 24px. Then buttons, borders, focus rings and icons need 3:1 of their own, under a separate rule covering the parts of a page that are not text. Build against the first threshold alone and you throw away accent colors you could have kept. Forget the third and every control you draw fails from birth.
Then there is the stranded accent, the brand color too light to carry white text and too dark to sit on your own background, so nothing built from it works either way round. The fix is usually a shade shift small enough that nobody outside the design team will ever notice, and we can show you which shifts get you there. Two things worth not repainting, by the way. Text that is part of a logo or a brand name carries no contrast requirement at all, and neither does a control that is switched off. Placeholder text does, and that is the one people expect to be excused.
What We Check
- Test every pair in the kit against all three thresholds rather than the 4.5:1 one alone
- Map which pairs carry body text, which carry large text only, and which carry neither
- Check buttons, links, borders, and focus rings against the 3:1 bar for non-text
- Review typefaces and weights at the sizes your pages actually use, because thin strokes and unfamiliar letterforms get harder to read as contrast drops
- Leave the wordmark and the switched-off states alone, since neither one carries a contrast requirement
- Recommend the smallest shade shift that turns each failing pair into a passing one
What You Get
A palette review is one pass rather than three. Every pair in your kit gets measured, and what comes back is a written grid. This pair carries body text. This one carries large text only. This one carries nothing, and here is the shade that would change that.
Failing pairs come back with a suggested value rather than a red cross, because a review that only says no leaves you exactly where you started. There is no automated scan, because a kit is not a page, and no screen-reader session for the same reason.
The Honest Limit
One honest limit. A brand kit cannot sit through a screen-reader session, so this is an expert review with no blind tester on it. It also cannot be tested for whether it survives a reader turning up their own letter and word spacing, because rule 1.4.12 measures built pages rather than a starting value in a style guide. What a review can settle is every pair in the kit, which is the decision that repeats on every page you will ever build.
What It Costs
A palette review is quoted rather than packaged, because brand kits vary in size a good deal more than websites do. Send the palette and the type scale, and a fixed number comes back by email before anybody opens anything. The two prices on our pricing page buy audits of a built site, and a kit is not one yet.
Worth Reading Next
Related on this site
Guides, checklists, tools, and terms that go with this service.