How to Evaluate Logo Recognition Before and After a Redesign

What you are evaluating

Logo recognition is correct brand identification from the mark. Evaluating it before and after a redesign means showing the current logo and the proposed logo under the same conditions and recording whether viewers name the right brand, confuse it with another, or cannot identify it. Preference, novelty, and “it looks more modern” are separate judgments and do not substitute for identification.

Run that comparison on the vector artwork that will actually be delivered. A vector logo is built from paths: outlines defined by anchor points and curves, which scale without the blockiness of a pixel (raster) image. A lockup is the approved arrangement of symbol, wordmark, spacing, and clear space. Construction details in those paths and in the type—stroke weight, counters, letter spacing, and how the symbol sits with the name—change how the mark reads at size. A decorative mockup, a compressed screenshot, or a different lockup from the one you will ship will not tell you how the real mark performs.

Identify the correct before and after artwork

A fair test needs the same class of asset on both sides: the authorized current logo and the proposed redesign in the lockup, colorway, and crop intended for the use you are judging. Icon-only at small size, a horizontal lockup, a stacked lockup, and a one-color or reversed (light-on-dark) version are different marks for recognition purposes. Mixing those versions treats a presentation change as if it were a redesign.

Open each file in a vector editor and inspect the objects, not only the filename. True vector artwork shows selectable paths and anchors. If type is still live, the editor reports a font rather than a set of outlines that merely look like letters. A placed photograph, a flattened pixel layer, or an embedded bitmap remains raster even when it sits inside a vector document. An .ai, .eps, .pdf, or .svg extension is not proof that the content is vector-only, fully editable, or cleared for reuse.

• Confirm you are looking at paths you can select and, where relevant, live type you can still edit, rather than a single embedded image.
• Note whether letters are live type or converted outlines, and whether the intended font is present if type is live.
• Record the color variant: full color, single color, reversed, and any simplified small-size drawing.
• Match the lockup to the use under test. Do not extract an icon from a wide lockup on one side and use a full wordmark on the other unless that extraction is the actual change.
• When the mark belongs to a named organization, verify identity and the current official artwork against that organization’s own brand source. Third-party copies, auto-traced downloads, and old lockups are common mismatches.

If the editor cannot show path structure, treat the file as a picture of the logo, not as production artwork. You can still use a faithful rasterization of a verified vector file for on-screen testing, provided both the before and after images are produced the same way from the path originals. Do not treat that rasterization as evidence that the source is editable or licensed.

Match the viewing conditions

Recognition depends on size, contrast, crop, and how long the mark is in view. Hold those factors constant so a drop or gain can be attributed to the redesign rather than to the test setup. The following controls are design recommendations; they are not a documented industry threshold or a substitute for a brand’s own research protocol if one exists.

• Show both marks at the same physical or pixel size, including a small-size pass that matches the tightest real use (app icon, embroidery, favicon-scale, or footer lockup).
• Use the same background and contrast. A mark that relies on color alone often fails when reversed or reduced to one color.
• Keep the crop consistent. If the redesign changes the lockup, test the old lockup against the new lockup as they will appear in that use, and separately test symbol-only if the symbol will stand alone.
• If you use timed exposure, use the same duration for both marks.
• Do not coach the second viewing with the first. Use separate viewer groups, or counterbalance order, and do not display the brand name beside the mark during identification.

What to record

Ask viewers to name the brand from the mark alone. Record the outcome in categories you can compare, not only a single yes-or-no score. No universal pass percentage, sample size, or exposure time is supplied here; if an organization publishes its own research rules, those documented rules govern that brand. Otherwise treat the categories below as a practical comparison, not as a performance claim.

Outcome	What it tells you about the artwork
Correct identification	The mark still carries distinctive structure at the size and colorway tested.
Wrong brand or near-miss	The new paths or letterforms overlap another known mark, or the distinctive old features were removed.
Familiar but unnamed	General category is readable; unique geometry or the wordmark is not.
Fails only at small size	Stroke weight, counters, or spacing collapse when scaled; the large presentation hid the problem.
Fails only in one color or reverse	Recognition was carried by color or by a background the shipping file will not have.

Turn the result into an artwork decision

If identification holds across the sizes and colorways you will actually ship, the proposed vector file is a candidate for delivery in those uses. If identification drops, inspect the path geometry and typography of the redesign against the current mark instead of adjusting the test until the new mark looks better.

• Stroke weights that look even at large size may merge or disappear when scaled down.
• Counters—the enclosed interior spaces of letters and symbols—fill in when outlines are too heavy or when a simplified drawing closes gaps that used to identify the form.
• Tracking and kerning in the wordmark can break a readable name into an unfamiliar texture.
• A simplified symbol that drops a distinctive cut, join, or proportion may still feel related in a side-by-side layout and fail when shown alone.
• Live type that will be outlined for delivery should be tested as outlines, because hinting and font availability will not travel with the file.

Decide among keeping the current artwork, revising the proposed paths and type, or shipping a size- or color-specific variant only after those files—not stand-in pictures—have been through the same identification pass. A redesigned lockup that is recognizable in full color at generous size is not automatically the mark to deliver for small, one-color, or symbol-only use.