Variant governance

Compare logo variants without confusing similarity with approval

Two logo files can be visually close and still serve different jobs. A useful comparison explains what changed, why it matters and whether the variation is authorized.

OutcomeA named master plus documented lockups, colorways and exceptions.

Workflow guide. The private comparison tool accepts one master and up to 60 selected reference images from files, a folder or ZIP. Optional saved fingerprints stay in your browser. Export JSON, CSV or a print review; no files are changed or deleted.

Compare construction before decoration

Start with silhouette, spacing, alignment and proportion. A color change can distract from a geometry change, while a different export background can make identical artwork appear unrelated.

Review color as a separate signal

Ask whether the variation is an approved colorway, a reproduction accommodation or accidental drift. Compare actual palette values and transparency rather than relying only on appearance inside one viewer.

Check the production role

Horizontal, stacked, symbol-only, small-size and monochrome variants should be retained when their purpose is documented. A filename such as logo-final-3.svg describes history, not role. Prefer names that communicate intended use.

Use relationship tiers carefully

  • Exact: likely duplicate export.
  • Close variant: same identity with a meaningful or accidental treatment change.
  • Related: shared construction that needs context.
  • Different: no strong fingerprint relationship; still inspect manually.

Record the final decision

For every retained variant, document the approved name, role, background, color mode and minimum-size guidance. A visual score is only useful when it leads to a clear operational decision.

Optional usage measurement

Help us understand which catalogues and file formats are useful. We do not record search text, uploaded images, names, email addresses or IP addresses in usage events. Events expire after 30 days. Local, staff and recognised bot traffic is separate.

Measurement is off until you opt in. Downloads have separate operational delivery counts.