Begin with the decision
Define what a viewer must do and where that decision must be made. “Read the sign” is not the operational goal. The goal might be to recognize the correct entrance before passing it, select a lane without abrupt movement, confirm arrival, or identify a room while moving down a corridor. Mark the earliest useful recognition point and measure from there. A measurement taken directly in front of the panel describes inspection distance, not decision distance.
Separate detection, recognition and reading
A viewer first detects that an object may carry information, then recognizes its category or destination, and finally reads detail. Color, silhouette, placement and a familiar arrow can aid detection, but they cannot replace readable words where words are required. Decide which layer must work at which distance. A large destination name and smaller secondary instruction can be appropriate if the hierarchy matches the journey.
Use ratios only to start a mock-up
Rules of thumb linking letter height to distance are planning shortcuts, not universal performance standards. Stroke width, width-to-height proportion, spacing, capitalization, message length, illumination, visual acuity and speed all matter. Use a conservative estimate to create the first full-size proof, then test that proof in context. Where accessibility or traffic-control standards apply, the governing requirement is the starting floor, not an optional reference.
Build at full scale
Print tiled paper letters, cut inexpensive vinyl, project the layout at calibrated scale, or mount a temporary panel. Include the actual border, background, line breaks, arrows and competing messages. Place the mock-up at the intended height and angle. A reduced drawing on a monitor cannot reproduce viewing distance or peripheral competition. Photograph the test from eye-level positions, but do not rely on a wide-angle phone image alone; stand in the space and observe.
Test movement and exposure time
A stationary reader can search and reread. A walking visitor has less time. A driver must divide attention and should not be asked to process an improvised dense message. Conduct passes at realistic speed with a second person recording the first point of correct recognition. Never create unsafe testing conditions, enter traffic, or distract a driver. Road signs and traffic-control devices require specialized standards and qualified professionals.
Check contrast in light, not in a swatch book
High chromatic difference does not guarantee high luminance contrast. Gloss can turn a dark field into a bright reflection, while backlighting can change apparent stroke weight. Test morning and afternoon light, interior lights, nighttime conditions where relevant, and wet surfaces if exterior exposure matters. Look for veiling glare, shadow, bright windows behind the sign and uneven illumination.
Edit message load
When a layout fails, shrinking secondary copy is usually the wrong first response. Remove information that belongs elsewhere, shorten names consistently, separate overview from direction and use progressive disclosure. Keep arrows close to the destinations they govern. Avoid several competing alignments that force the eye to reconstruct relationships.
Record the result
Create a test sheet with location, date, light, viewing points, panel size, letter heights, typeface, colors, finish, observers and errors. Note not only whether someone eventually read the message, but when they first made the correct decision. Revise one variable at a time when possible. The resulting record makes design review more useful than a preference debate.
Know the limit
A field test supports a decision; it does not certify compliance, accessibility, structural safety or suitability for a particular user population. Confirm current local codes, adopted accessibility standards, landlord criteria and transportation requirements with the authorities and qualified professionals responsible for the project.