Custom Logo Design That Works at Every Size

What is the difference between a logo and a brand identity?

A logo is one mark. A brand identity is the whole system it lives inside — palette, typography, layout rules, imagery style, voice and the guidelines that keep them consistent.

This distinction is worth understanding before you commission anything, because they are different purchases at different scales. A logo project produces a mark and its variants. An identity project produces everything needed to make a website, a document, a social post and a business card all recognizably belong to the same company.

A logo on its own is genuinely enough for many businesses, particularly early on. What causes trouble is expecting a logo project to deliver identity outcomes — commissioning a mark and then wondering why the website still does not feel coherent.

Why do some logos stop working after a few years?

Usually because they were designed for a context that changed, or designed for one context only. The most common version is a mark designed for print or a website header that later has to be an app icon, and cannot be.

Detail is the recurring culprit. A mark with fine lines, layered gradients or a wordmark integrated into an illustration will look excellent at large sizes and become an indistinct smudge at 32 pixels. If it cannot be reduced to a clean icon, every small placement becomes a compromise.

The other cause is following a visual trend too closely. Marks that adopt whatever styling was fashionable in a given year date precisely as fast as that style does. Simpler, more structural marks tend to survive because there is less about them to go out of fashion.

How many concepts should a logo project include?

Fewer, developed further, is better than many presented shallowly. Three genuinely different and properly developed directions beat ten variations that are mostly the same idea in different colors.

A large number of initial concepts sounds like better value and usually is not. It means each one received less thought, and it makes the decision harder — clients presented with ten options tend to pick by preference rather than by fit, and often ask to combine elements from several, which rarely improves any of them.

What matters more than concept count is how many rounds of refinement the chosen direction gets. That is where a promising idea becomes a finished mark, and it is the part most often cut when a project is priced on concept volume.

What files should you receive?

Vector masters plus every raster export you will realistically need — and the vector files are the important part, because everything else can be regenerated from them.

A complete handover includes vector formats (SVG, EPS and the editable source), raster exports at the sizes your website and platforms require, favicon and app icon files, single-color versions in both black and white, and variants for light and dark backgrounds.

The single most common gap is receiving only PNGs. It means you cannot scale the mark for signage, cannot generate a new size cleanly, and have to go back to the designer for changes that should take minutes. Insist on vector source files and confirm the rights transfer in writing, whoever designs it.

What makes a logo work at every size?

Restraint at the small end, decided at the start rather than discovered at delivery. A mark that only works large has been designed for the presentation rather than for use.

The practical constraint is that fine detail disappears. Thin strokes fill in when printed and vanish on screen; enclosed counters close up; two shapes that read as separate at large sizes merge at small ones. Designing while checking at favicon size prevents all of it.

The usual answer is a system rather than one drawing. A full lockup with the name for wide placements, a compact stacked version for narrow ones, and a symbol alone for the smallest — each drawn deliberately rather than scaled down and hoped for.

The symbol-alone version needs the most honesty. It only works if the business is recognizable without its name, which most are not initially. Until then it belongs in places where context supplies the name, and the full lockup does the work everywhere else.

How many concepts should a project include?

Few, developed properly, rather than many presented thinly. A page of twenty options is a way of transferring the decision to the client rather than making it.

Two or three genuinely different directions, each explored far enough to be judged, is more useful than a wall of variations. Each should represent a real strategic choice — a different argument about what the business is — not the same idea in three typefaces.

Each direction needs to be shown in use. A mark on a white slide tells you very little; the same mark on a sign, an app icon, a business card and a dark website tells you whether it works. Approving something that has only been seen in isolation is how logos get approved and then quietly redrawn.

Revisions should refine rather than restart. A brief that produces three directions and then requests three more is usually a brief that was never settled, and the fix is upstream — in the strategy — rather than in more drawing.

What files should you receive?

Vector masters first, because everything else can be produced from them and they can be produced from nothing else.

That means editable vector files in the standard interchange formats, with the type converted to outlines so the mark renders correctly on a machine without your font licensed.

Then the raster exports each context needs at the sizes it needs — favicon sizes, app icon sizes, social profile dimensions, and transparent versions for placing on color. Providing these prevents someone in your team exporting their own at the wrong dimensions, which is how inconsistency starts.

And color specifications in every system you will use: screen values, print values and, where relevant, a spot color. A brand blue that differs between the website and the printed brochure is nearly always a missing specification rather than a printing error.

Why do some logos stop working?

Because the contexts changed and the mark did not. Most logos that feel dated were designed correctly for a world with different constraints.

The clearest example is small-screen use. A mark designed when the smallest reproduction was a business card now has to work as a 32-pixel favicon and a circular avatar, and detail that was appropriate then is illegible now.

Dark interfaces are the more recent version of the same problem. A logo designed only for white backgrounds needs a redraw, an outline version or a container — and if none was provided, everyone improvises their own.

The other cause is drift. Without a master file and clear rules, each use is copied from the last, and after a few years several slightly different versions are in circulation with nobody sure which is correct. That is a governance problem more than a design one, and it is what brand identity work exists to prevent.

How should a logo be presented for approval?

In use, at real sizes, on the surfaces it will actually appear on — never as a large mark centered on a white slide.

The reason is that a logo approved in isolation is being judged on the wrong criteria. What matters is whether it works as a favicon, on a sign, on a dark website and beside competitors, and none of that is visible at presentation size on white.

The presentation should therefore include the awkward cases deliberately: the smallest size it will be used at, a single-color version, reversed out of a dark background, and placed on a photograph.

It also helps to present fewer options in more depth. A wall of variations transfers the decision to the client without giving them the information to make it, and the version chosen that way frequently fails a test nobody ran.

And approval should involve whoever will apply it afterwards. A mark approved by leadership and handed to a team who find it unusable in half their contexts is a common and avoidable outcome.