A logo is a compression algorithm, not a story
A mark has to work at 16 pixels, in a single color, and be recognized before it is read. Meaning arrives through years of repetition, not through the drawing. What that means for the brief you write.
Contents
A logo's job is compression. It has to carry an identity at 16 pixels, in one color, on a receipt and on a screen, and be recognized before it is read. Meaning attaches to it through years of repeated exposure. It does not arrive inside the drawing, and no brief can put it there.
Key takeaways
- A 128-participant study measured 39% brand recall for left-positioned website logos against 21% for right-positioned ones (Nielsen Norman Group, 21 February 2016).
- WCAG 2.1 exempts logotypes from its 3:1 non-text contrast requirement, so the mark carries no legibility guarantee and its surroundings must compensate (W3C, June 2018).
- WIPO's Vienna Classification sorts figurative marks into 29 categories of shapes, applied by around 60 trademark offices. A registry sees geometry, not narrative (WIPO).
- Distinctive assets are measured on fame and uniqueness: how many category buyers link an asset to a brand, and how many link it to that brand alone (Jenni Romaniuk, *Building Distinctive Brand Assets*, Oxford University Press, 2018).
What a mark has to survive
Write the list of surfaces before drawing anything. A favicon at 16 pixels. An app icon on a crowded home screen. Embroidery, where thin strokes vanish into thread. A one-color stamp on a corrugated box. A sponsor wall at 12 mm.
Each of those is a lossy channel. Detail, gradient, thin counters and tight kerning are the first things it throws away. A mark that only holds at hero size on a slide has not been tested; it has been admired.
WIPO's Vienna Classification, established by the Vienna Agreement of 12 June 1973, sorts the figurative elements of trademarks into 29 categories, from the general to the particular. Around 60 trademark offices apply it, including the EUIPO. Source: WIPO.
Registries index marks by shape because shape is what survives reproduction. If a mark cannot be described in three shape terms, it carries detail it will lose anyway.
Recognition is repetition, not invention
Jenni Romaniuk's framework at the Ehrenberg-Bass Institute measures distinctive assets on two axes: fame, meaning the share of category buyers who link the asset to the brand, and uniqueness, meaning the share who link it to that brand and no other. Both are outcomes of consistent use over time.
Neither axis rewards how much thought went into the drawing. A mark can be elegant and internally meaningful while scoring near zero on both, because nobody has seen it enough times. That is an accumulation problem, and the only fix is years of consistent application.
Every redraw spends that accumulation, which is why "refresh the logo" is an expensive default answer to a problem that usually sits somewhere else in the business.
The one-color test comes before the color test
Reduce the mark to a single flat color on white, invert it, then rasterize it to 16 pixels square. Three failures show up almost immediately: the silhouette stops reading as one object, the counters fill in, or the shape resolves into something else at small size.
None is fixable with color. Color is the last variable added, because it is the one the channel removes most often — embroidery, foil, etching, a monochrome partner lockup.
WCAG 2.1 Success Criterion 1.4.11 requires a 3:1 contrast ratio for graphical objects needed to understand content, but the W3C exempts logos and logotypes: a mark shown in the brand's own colors counts as essential. Source: W3C, June 2018.
The exemption is permission, not endorsement. The standard will not catch a mark that is illegible against its own header, so the studio has to. Design against the worst background before designing the palette.
Why "tell our story" is the wrong brief
The story brief asks one glyph to encode positioning, three product lines, a founding year and a set of values. What comes back is a mark with a legend: the arc is growth, the negative space is the client, the two tones are heritage and innovation.
Marks built this way carry detail that exists to justify the explanation, and that detail is exactly what the 16-pixel test deletes. The explanation itself only lives inside the studio deck.
Story belongs in the system around the mark: naming, tone of voice, photography, the way an interface behaves. A brief that pushes narrative into the glyph is usually one that has not defined the actual problem.
What to put in the brief instead
A brief that produces testable work states constraints, not adjectives. Four lines are usually enough:
- Minimum reproduction size in millimeters and in pixels, with the surfaces it comes from.
- Reproduction methods it must survive: one color, reversed out, embroidered, etched, printed on the client's substrate.
- The confusion set — the marks in the category it must not resemble, listed by name.
- The asset it is meant to become, so fame and uniqueness can be measured later.
Every line is checkable by someone who was not in the room, which is why clear constraints beat a persuasive narrative when a project starts going wrong.
FAQ
Does a logo need to explain what the company does?
No. A mark that depicts the product locks the brand to a category it may leave, and descriptive shapes are the least distinctive ones in any registry. Explanation belongs to the name, the tagline and the site. The mark's job is to be identified quickly and reproduced anywhere.
How small should we test a logo?
Test at the smallest size the brand will actually appear: 16 pixels for a favicon, and whatever physical size the client's packaging, embroidery or signage uses. Rasterize it rather than viewing a vector scaled down, because vector previews hide the pixel-level failures that real rendering produces.
How long before a new mark is recognized?
Longer than any campaign. Fame and uniqueness accumulate through consistent exposure across years, not through a launch. Plan the first eighteen months as application discipline rather than promotion, and measure recognition periodically so the curve is visible before someone proposes redrawing it.
What to do next
Print the current mark at its smallest real size in one color, next to the five marks it competes with on a shelf or in a marketplace listing. If it fails there, the fix is structural and comes before any palette decision. Then write down what the mark should become as an asset, so in six months there is something to measure beyond whether the room liked it. That is a harder sale than a story, and the same argument as charging for judgment rather than output.
