Four formulations were reviewed. One was selected. Three were set aside, and the reasons are worth writing down, because the decisions were more interesting than the one we kept.
The first set aside had more allergen complexity than we were comfortable with. Not anything prohibited, and not anything a reasonable person would object to seeing on a label. Simply more components that would have to be named separately, and therefore more ways for someone to be unable to work out what their skin was responding to. For a product whose whole point is being readable, that was the wrong direction.
The second depended on an alcohol level that was not disclosed to us clearly at the point we asked. The question was straightforward and the answer was not, and we decided that a specification we could not state plainly was a specification we could not stand behind. It may well have been a perfectly good formulation. We were never in a position to say so, and that is the problem.
The third came with claims attached that could not yet be evidenced. There is a version of this business where you accept the claims, print them, and hope nobody asks. We would then have been responsible for statements we had no file to support, which is both a legal exposure and a slow way to lose the only thing a new brand actually has.
The one we chose was the plainest of the four. It does the least on paper. It has the shortest story, the fewest things to say about it, and the least interesting description. It is also the one where every line of the specification could be stated, checked and defended, and where nothing had to be softened or left vague to make it sound better than it is.
We are not naming any product, code or supplier here. The point is not to comment on anyone else's work, and the three set aside were rejected for our purposes rather than judged in general. Two of the three could suit a different brand with different priorities perfectly well.
What the process changed was our sense of where the work actually is. It is not in finding something impressive. Impressive is easy to find and hard to verify. It is in narrowing down to the thing you can describe completely, and then resisting the urge to describe it as more than it is.
The interesting part of building a brand is what you leave out.
Update, 17 August 2026. The alcohol figure was subsequently provided: 7.5%. It took six days from the question. We are recording it here because a brand that publishes what it did not know should also publish what it later learned. The formulation remains set aside, for the reasons above, but the specification is no longer unstated.