How to Test a Logo Before Launch: Six Checks That Catch Almost Everything

A mark is nearly always judged wrongly

The logo is shown large, centred, on a light ground. That is the one place almost any shape looks fine.

The real life of a mark looks different: sixteen pixels in a browser tab, one-colour embossing, overlaid on a busy photograph, printed on a cheap printer.

The six checks below take half an hour and catch almost every error that otherwise surfaces after printing.

Reduction

The first and most productive test.

Shrink the mark to sixteen pixels — favicon size. Thin lines merge, counters close, serifs vanish. If a recognisable shape survives at that size, the construction is sound.

If it turns to mush, a compact version is needed: simplified, not merely scaled down.

One colour

Colour can hold a form that does not hold on its own, and that hides structural problems.

Convert the mark to pure black on white and back. If it falls apart without gradients or colour separation, the problem is the form, not the palette.

The same test prepares the mark for engraving, embossing and screen printing, where there is no colour by definition.

Ground

A mark rarely lives on white.

Put it on black, then on the lightest and darkest areas of one photograph. A mark that reads on white often disappears on dark: the outline merges with the ground, or fine elements drop out.

That is where light and dark versions come from — not as a nicety, but as the result of this check.

Print and distance

Print the mark on an ordinary office printer and step back a metre.

Cheap printing thickens lines and fills small counters, much as embroidery on a uniform or print on packaging will.

Looking from a distance shows what survives of the mark once detail is no longer visible. That is what a person actually remembers.

Rotation and a stranger

The last two tests are the cheapest and the most uncomfortable.

Rotate the mark by a hundred and eighty degrees and mirror it: this sometimes reveals readings nobody intended, and it is better to notice them before the run.

Then show it to someone who was not on the project. Do not ask whether they like it — ask what they see. The gap between intention and answer is the real result of the test.

Who runs the checks, and when

The checks divide between two parties, and confusion about that is why some of them get run by nobody.

On the studio side sit the technical ones: reduction, one colour, inversion, behaviour on photography, rotation. They need access to the source and happen before the mark is shown at all.

On your side sit the contextual ones: how the mark looks beside your competitors on a shelf, how it reads in your industry, whether the form carries a local meaning the supplier could not have known about.

The similarity check sits apart: technically a studio can run it, but the decision about how deep to go belongs to the client, because the cost of being wrong is theirs.

This division is worth agreeing before the start. Otherwise each side assumes the other checked, and nobody does.

When the checking should happen

Timing matters as much as content. Checks run after approval barely affect the result: the mark is already chosen, and findings land as nitpicking.

The right place is before the presentation. Directions that reach a client should already have passed reduction and one colour. Then the discussion is about which direction is right rather than whether it works at all.

That changes the character of the conversation too. A client shown three options, each already verified, chooses on meaning. A client shown attractive pictures chooses on taste, and half of what they choose falls apart later.

The second checking moment comes after the chosen direction is refined and before files are assembled. That catches what shifted while proportions were adjusted.

There is no third moment: once the files are assembled, a correction means assembling them again.

The checks people forget most often

Beyond the six main ones there are several that get remembered rarely and regretted often.

First, the mark on a coloured ground rather than only on black and white. Many marks are built on contrast with something light, and on the company's own brand colour they suddenly disappear.

Second, the mark beside other marks: on a partners page, in a site footer, on a sponsor wall. There it sits among other people's forms and must neither vanish nor argue.

Third, vertical application: a banner, a bookmark, a spine. A mark designed only for horizontal use either rotates awkwardly or does not fit.

Fourth, the name set in running text: how the brand looks written in an ordinary typeface inside a sentence. Sometimes the mark and the written form turn out to diverge, which creates confusion in documents.

What to do with the findings: A dark ground exposes marks that exist only on white.

A check produces not a verdict but a list of decisions. They usually come at three levels.

First: the mark gets refined. Lines thicken, counters open, a detail is simplified. The form stays recognisable and the problems go. That is the most common and the cheapest outcome.

Second: an additional version is made. Compact for small sizes, single-colour for engraving, vertical for narrow carriers. The original mark does not change; a family is added around it.

Third: the direction is rejected. That happens rarely and only where the problem is structural — for instance, a form that reads only in colour and only large.

What matters is that all three outcomes are normal and all three cost less than discovering the problem after printing. A check that finds nothing is useful too: it means the mark can launch without caveats.

Writing the check down so it can be repeated

A one-off check is useful once. A written one works for years, because it can be applied to the next mark, to an updated version, and to another supplier's work.

Writing it down is simple: one page listing the tests with a box beside each. Reduction to sixteen pixels — yes or no. One colour — yes or no. Dark ground, the light part of a photograph, rotation, a stranger's reading.

Attaching a pass and a fail example to each test helps — using your own mark where one exists. Then the sheet makes sense to someone seeing it for the first time.

That sheet belongs beside the file set and should be handed over with it. It is one page that makes the handover checkable.

And it converts a subjective conversation about whether the mark is finished into a list with yes and no answers.

Why the author cannot check their own work

There is a limit experience does not overcome: the person who invented a form sees the intention in it rather than the result. They know a stylised letter is in there, so they see it; a stranger sees a shape.

This is not a sign of a weak designer. It is how perception works: knowledge of what is depicted cannot be switched off on request. Which is why the final check on meaning is always done by somebody else.

The practical method is simple and effective: show the mark to three people outside the project for a few seconds each, and write down what they named. Not whether they liked it, but what it is.

If two of the three answers match the intention, the form reads. If all three differ, the mark communicates nothing specific, and that is worth knowing before printing rather than after.

The same method catches unintended readings: an outsider will say out loud what the author failed to notice across a week of work.