Six months after a redesign, the question always comes, often at a board meeting: “did it work?”. And the team that carried the project finds itself without a solid answer. It talks about impressions, comments received, a general feeling. It is not a lack of rigour. It is that nobody ever told them what to observe, or when.
The two false measures
The first false measure is revenue. It is tempting because it is available and because everyone cares about it. It is unusable because dozens of other factors move at the same time: the economy, a competitor closing, a new product, a price change, a grant, the weather for a cultural venue. Attributing a rise or a fall to the visual identity is a matter of belief, not measurement. When the brand does well, the agency gets wrongly congratulated. When the brand does badly, it gets just as wrongly blamed.
The second false measure is the preference survey. “Do you like our new logo?” measures nothing useful, for a well-documented reason: people resist a change to a sign they knew, and that resistance fades within a few months. A survey taken three weeks after a launch measures surprise, not quality. Almost every landmark redesign in history was poorly received at first, and nobody would want to go back today.
There is, by the way, a better question than “do you like it?”, and it changes everything: “what does it make you think of?”. The first asks for a verdict of taste. The second produces usable content, because it reveals the gap between what you wanted to say and what is perceived. That gap, and it alone, constitutes a measure of identity.
Measure before, or nothing is comparable
Here is the most frequent and most definitive mistake: deciding to measure after the launch. By then it is too late, because there is no longer any point of comparison. All you can produce are raw numbers that nobody knows are good or bad.
Good practice takes half a day, before the project begins. You record the starting state on a few simple indicators, you archive them, and you will take them again later with exactly the same protocol. These are not sophisticated measurements. They are benchmarks, and an imperfect benchmark is infinitely better than no benchmark at all.
This initial measurement has a second benefit, often more valuable than the first: it forces the organization to state what it really expects from the project. It is very common to discover, at that moment, that senior management and the communications team were not expecting the same thing at all. Better to find out before the first sketch than at the review.
Five signals that say something
None of these signals is enough on its own. Together, they paint an honest portrait, and they are all accessible without a research budget.
- Recognition without the name. Show your materials with the name masked. How many people in your audience identify you?
- Five-second recall. Show the home page for five seconds, remove it, ask what the organization does.
- Consistency of in-house work. Gather ten documents made without the agency. Do they look like they come from the same organization?
- Production time for a routine piece. How long does it take to get an event poster out, from request to final version?
- Frequency of exception requests. How many times a month does someone ask to step outside the system?
The first signal is the closest to what people mean by “awareness”, but it is far more demanding than name recall. An identity that works ends up being recognizable even without its logo, through its colour, its typography, its framing, its tone. That is the threshold that separates a brand from mere dressing. It is rarely crossed before two years, and it is never crossed if the organization changes visual direction every eighteen months.
The second signal tests clarity rather than aesthetics. An organization whose activity nobody can sum up after five seconds of its home page has a hierarchy problem, not a style problem. This test is brutal, it costs ten minutes, and it almost always produces an unpleasant surprise the first time.
The best indicator is internal
The most reliable signal of all is not in the public: it is in your own files. Take ten documents produced by your teams without outside help over the last six months. A department presentation, an event poster, a form, a social post, an annual report. Lay them side by side.
If these ten documents look like they come from the same organization, your system works. It is understood, it is usable, it holds without supervision. If they look like they come from five different organizations, the system does not work, however beautiful the guidelines. It is not necessarily a design problem: very often it is a tooling problem. A rulebook was delivered, without the templates that make it possible to apply the rules.
The last two signals follow the same principle. If producing a poster takes three weeks and four rounds of revisions, the system costs money every month. If someone asks to make an exception every week, the system does not cover a real need, and it is better to extend it than to defend it. A living identity gets corrected. An identity defended against its own users dies slowly.
The reading calendar
The timing of the measurement matters as much as the measurement itself, because an identity does not produce its effects right away. Three horizons stand out clearly.
At three months, nothing significant can be measured about perception. On the other hand, internal adoption can be usefully observed: are the teams using the new templates, or are they carrying on with the old files? It is the moment to fix the tooling, not to judge the design.
At one year, the novelty has passed and the comparison with the initial state becomes valid. It is the right time to take the five signals again with the same protocol, and to read the gap. It is also the moment when you can honestly say whether the system holds in the real life of the organization.
At three years, you measure what really counts: durability. Does the identity still hold without having been patched up everywhere? Has it absorbed the new needs that appeared, a new department, a new platform, a new audience? An identity that goes through three years without emergency repairs has succeeded, regardless of what people thought of it at launch.
What an identity cannot fix
It also has to be said what an identity will not do, and said before starting. It does not fix an offering that cannot find its market. It does not make up for failing customer service. It does not replace a go-to-market strategy. An organization that expects a redesign to solve a fundamental problem will be disappointed, and will redo its identity in three years thinking the first one was not the right one.
What an identity does, on the other hand, is perfectly real: it makes an organization recognizable, it reduces the effort needed to be understood, it gives its teams a framework for producing quickly and well, and it accumulates value with every exposure instead of starting from scratch. These are modest effects taken one by one, considerable over five years, and all observable.
Measuring an identity is therefore not about looking for a single number that would justify the invoice. It is about giving yourself a few honest benchmarks, taken before and again after, to know whether the system holds and where it needs adjusting. It is also the best protection against the reflex redesign, the one launched because people got bored. Our articles on logo redesigns and on the real price of a logo deal with the decision upstream. If you are about to launch an identity project, write to us: the initial half-day of measurement is easy to schedule, and it will be worth a great deal when the time comes to take stock.