Sheet 09No. 0454 entries

Logo design principles, in the order they are applied

A logo is not judged on whether it is liked. It is judged on what survives: being shrunk, printed in one ink, stitched, embroidered, faxed, favicon-ed and glanced at. The principles below are ordered by the test that kills the most candidates first, which is almost always the smallest size.

A precision drawing compass and a clear circle template lying on a drafting board, a violet triangle at the edge.
No. 045 A compass makes a circle repeatable. A principle makes a judgement repeatable, which is the same trick.

Entries in this section

Register
  • No. 077 How to make a good logo 15 Aug 2026 Open How to make a good logo Why a mark survives being shrunk to sixteen pixels, and the five tests that decide it before any colour is chosen.
  • No. 602 Logo formats and files 08 Aug 2026 Open Logo formats and files The vector master, the SVG, the one-colour pair and the one-page note that stops a mark degrading.
  • No. 519 Logo presentation 06 Aug 2026 Open Logo presentation Show the mark, then show it small, then explain, and why the format decides the question the room answers.
  • No. 331 Pain and gain in a mark 04 Aug 2026 Open Pain and gain in a mark Three devices for holding two opposed ideas in one shape, and the four ways that kind of mark breaks.

Principles are tests, not adjectives

Most published lists of logo design principles are adjectives: simple, memorable, timeless, versatile. Adjectives cannot be failed, and that is exactly what makes them comfortable and useless. Each one below is written as a test with a pass condition, because a principle that cannot reject a design is a compliment.

Test one: sixteen pixels

Render the mark at sixteen pixels square and look at it. This is the favicon size, the notification size, and the size at which most people will encounter it most often. If the counters have filled in, if two shapes have merged into one, or if it has become an indistinct blob, the mark has failed the test that matters most in practice.

The usual cause is detail that exists at design size and cannot exist at reading size. The usual fix is a reduced variant with fewer elements, which is a legitimate answer rather than an admission of defeat; most durable identities ship two or three variants for exactly this reason.

Test two: one ink

Set the mark in solid black on white, then in solid white on black. Both have to work, and the second is the one that catches problems, because a mark built from a light shape sitting on a dark background frequently disappears when the relationship inverts.

Colour is the last decision, not the first. A mark that only works in its brand colour is a mark that stops working on an invoice, a fax, an embroidered shirt and a laser-etched panel.

Test three: silhouette

Fill the whole mark in flat black, remove every internal detail, and look at the outline. If the silhouette is generic, the mark is generic, and no amount of internal detail will rescue it. This is the fastest test for the swoosh-and-circle family of marks that dominate the middle of the market.

Test four: recall at six seconds

Show it, take it away, and ask someone to draw it. What comes back is what the mark actually is. This is uncomfortable and cheap and worth doing before a presentation rather than after one.

Test five: the field it will live in

Place the mark among the marks it will appear beside: competitors on a shelf, applications in a launcher, sponsors on a hoarding. Distinctiveness is relative and cannot be assessed on a white artboard. A mark that is beautiful alone and invisible in its field has failed.

What the principles do not decide

They do not decide whether the mark is a wordmark, a lettermark, a pictorial mark, an abstract mark or a combination. That follows from the name, a long name is a poor wordmark and a strong lettermark, and from where the mark has to appear. Nor do they decide the typeface, which is a separate craft covered on the type side of this reference.

They also do not decide the process. The main entry sets out a working order from brief to final files; formats and files covers what actually gets delivered; presentation covers how a mark is shown so that it is judged on the right thing; and pain and gain works through the specific case of a mark built on a contradiction, which is the hardest brief in the discipline.

Questions about marks

Queries

How many logo design principles are there really?

Five that can be tested, as above. Longer lists tend to be the same five restated as adjectives, plus a few that are properties of good design generally rather than of marks specifically.

Should a logo be designed in black and white first?

Yes, and not for tradition. Colour hides structural problems: two shapes that read as separate only because they are different colours will merge the moment the mark is printed in one ink or rendered in a system that strips colour.

Is a simple logo always better?

Simpler is not the goal; surviving is. Simplicity is usually what survives, which is why the two get confused. A mark with real detail that still reads at sixteen pixels has passed the test, and the number of anchor points is nobody’s business.

How is a mark tested against the principles before it ships?

At size, in the wrong conditions, and by somebody who has not seen it before. Print it at sixteen pixels and at the largest size it will ever appear, view it in greyscale, view it reversed out, and photocopy the page once. Then hand it to a reader cold and ask what it is, not whether they like it. Each of those answers a specific principle, and a mark that passes all of them has been tested rather than approved.