On this page
People approve logos on a large screen, in full colour, in isolation, at the end of a process they are tired of. That is the least representative viewing condition imaginable. Here are five tests that predict whether a mark will still be working in five years.
1. The squint test
Shrink the logo to 16 pixels — the size of a browser tab favicon — and look at it. Can you tell what it is? Most failed logos fail here, because they were designed at 400 pixels wide and contain detail that dissolves. If the mark becomes a grey smudge, you need a simplified variant at minimum, and possibly a simpler mark.
2. The one-colour test
Print it in solid black. Then reverse it out of black. A logo that only works in its gradient is a logo you cannot embroider, emboss, fax, stamp, engrave, or print in a single-colour newspaper advertisement. Colour should be a reinforcement, never the thing carrying the identity.
3. The memory test
Show it to someone for five seconds, take it away, and ask them to describe it. If they cannot, it is not distinctive enough to be recognised, which is the only job a logo has. Note that "describe it" is a much lower bar than "draw it" — almost nobody can accurately draw a logo from memory, including famous ones.
4. The category test
Put it beside your five closest competitors. Does it look like it belongs in the category, and does it look different from the others in it? Both matter. A law firm that looks like a toy brand loses clients; a law firm indistinguishable from four others gains nothing from its design at all.
5. The application test
Mock it up on the three things it will actually appear on most. For a builder that is a van, a hi-vis jacket and an invoice. For a SaaS company it is a favicon, an app store listing and a slide footer. Approving a logo on a white background is approving it in the one place it will almost never appear.
Common failures
- Too much detail. Illustrative marks with fine linework fail the squint test immediately.
- Dependence on a gradient or a photograph. Both break in single-colour reproduction.
- A typeface doing all the work. If the name set in a font is the whole logo, you own nothing distinctive.
- Following a trend precisely. The 2010s flat-and-geometric wave now dates work as reliably as a 1990s bevel does.
- Clip-art symbolism. A globe for international, a leaf for eco, a lightbulb for ideas. These say nothing because everyone uses them.
- Text too small in the lockup. Company names in logos are routinely set at sizes that become unreadable on signage.
What does not matter as much as people think
Whether it "means" something. Most iconic marks are abstract and acquired meaning through use rather than carrying it inherently. A clever hidden concept is a nice bonus and a terrible primary criterion — nobody but you will ever notice it.
Colour psychology. The claimed effects are far weaker and more culturally variable than marketing articles suggest. Choose colour for distinctiveness within your category and for legibility, not because a chart says blue means trust.
Want this done for you?
We will tell you honestly whether you need us for it. Send a two-line description of the business and we come back with a plan, a timeline and one fixed price.