The first time I laid out side by side the handover photos of a network built over two years, I didn't look at what was wrong. I looked at what had changed without anyone deciding it. The red was a shade lighter. The dimensional letters had a 3 cm profile, not 5. The light box sat 12 cm lower than the drawing called for, because the beam was where it was.
No one had broken the brand manual. Every deviation had been rational on the day it happened, made by a competent person, on scaffolding, with the information they had at that hour.
This is where the misunderstanding lies. Deviation in a network doesn't come from negligence. It comes from the fact that the brand manual and execution answer two different questions, and no one notices until the number of locations makes the difference visible.
Intent and fabrication
The manual says how it should look. Production needs to know how it's achieved — every time, on façades that don't resemble each other, with crews that change, in different years.
An identity manual, however well made, normally doesn't contain execution tolerances, approved suppliers for critical materials, procedures differentiated by substrate, or acceptance criteria that two people can apply and reach the same verdict. That's not a shortcoming of the manual. That's not what it does.
The distance between the intent specification and the fabrication specification is exactly the space where deviation accumulates. Whoever doesn't close it explicitly leaves it to interpretation — and interpretation is spread across dozens of people who don't know each other.
"Pantone 485 C" is not a specification
It's a color code. What ends up on the façade is a film or a paint, made by a specific manufacturer, with a specific finish and a UV behavior declared by that manufacturer. Two films from different manufacturers, both sold as the equivalent of the same Pantone, look different after a year and a half of southern exposure. Not dramatically. Enough to show when the new location is 200 meters from an old one, on the same commercial street.
The test is simple: if the specification allows two products from two suppliers, you don't have a specification — you have a preference. The same applies to sheet thickness, LED color temperature, ACP class, the fixing system, the type of silicone.
The reference has to be an object
The PDF doesn't settle disputes. The first executed and validated location does — if it's treated as the reference standard: handover photos from fixed angles, with a color reference in frame, physical samples kept (film, a corner of ACP, a segment of profile), a signed data sheet.
When at location 60 someone says it's not the same color, the comparison is made against a piece of material, not against a memory. The practical difference isn't in accuracy. It's in how long the discussion lasts and who pays for the remediation.
Variation isn't eliminated, it's anticipated
No network of 100 locations has 100 identical façades. Standardization doesn't mean identical installation — it means that, faced with variation, the decision is predictable. The situations repeat and can be written down in advance.
The available height is below the nominal one: it's reduced proportionally down to a defined threshold, below which you switch to the compact version of the logo. The substrate can't take direct anchoring: an independent metal support structure, with a standard detail already drawn. A building in a protected area: an identity variant prepared before it's asked for, not improvised after the permit is refused. A ventilated façade: a separate set of rules for penetrations and sealing.
The rule I apply: any exception solved once becomes a written rule. Otherwise you solve it a hundred times, slightly differently each time.
The four-year test
Four years from now, an element deteriorates. How long does it take to produce an identical replacement? If the answer starts with "we'll send someone to measure," there's no archive.
The per-location technical dossier — dimensioned survey, execution drawings, the materials actually used with batches where it matters, fixing details, electrical diagram, handover photos, permits — costs almost nothing on installation day. Reconstituting it later costs a trip, a survey and a delay the client experiences as an incomplete façade in the window.
Documented handover is part of the same logic, but it has a side effect more valuable than the stated one: when the installation crew knows exactly what is checked and in which photos, it checks itself beforehand. The remediation rate drops before handover even takes place.
Who holds the system
This is where the difference between a supplier and a network implementation partner shows, and it's a question that should be asked at qualification, not discovered after two years.
If the system lives only with the client, as documents no one operates, every change of manufacturer resets the network. The new contractor reinterprets and the deviation starts over from zero.
If it lives only with the manufacturer, the client is captive. They can't change supplier without losing the visual continuity of their own network. No serious expansion manager should accept that position, and no manufacturer who knows the trade needs it.
The correct position is the third: documented with the client, operated by the manufacturer. In a nationwide visual-identity rollout, the specifications, technical dossiers and reference standards belong to the client and stay with them. The manufacturer builds, maintains and applies them. The relationship rests on competence.
The qualification question that says the most about a contractor: what stays with the client if we end the collaboration tomorrow? A good answer contains objects and documents. A weak answer contains the installed locations.
What shows up in the metrics
A working system reads in data, not in perceptions. Time from contract to installation compared between the second location and the fiftieth — in a system, it drops; without a system, it stays constant. The post-handover remediation rate. The cost of post-warranty intervention, directly proportional to the quality of the archive. The duration of a franchise audit. The onboarding time of a local contractor when the network expands geographically.
If none of these improve as the number of locations grows, you don't have a system. You have a series of independent projects executed by the same supplier — which works up to a point, and that point comes sooner than everyone expects.
Design is solved once, at the start, usually well. What isn't solved once is keeping the same correct decision across a hundred contexts, with different people, year after year. Networks that look coherent don't have better brand manuals than the rest.
WOX FAD implements visual identity for multi-location networks nationwide — commercial façades, illuminated signage, ACP cladding, dimensional letters. The technical dossier, specifications and physical reference samples stay with the client.
If you have an expanding network and want to see, concretely, what a production specification looks like starting from your own brand manual, send us one real element and we'll show you its translation — not as a quote, as a work sample.