How to choose between two designs without a committee

Two designs, a deadline, and a group of people with opinions. The usual failure is not picking the worse option — it is spending three days not picking either, then choosing the one belonging to whoever argued longest.

Here is a method that takes about an hour.

First, decide whether this is a taste question

Some design choices have a right answer and should not be voted on:

• Does the text meet contrast requirements?
• Does it work at the size it will actually be seen?
• Is the trademark clear of the other one?
• Does it fall apart on a phone?

Check these first and eliminate anything that fails. Voting on an option that fails an accessibility check just launders a bad decision through a group.

What remains — which of two acceptable options works better — is a genuine taste question, and that is where asking people helps.

Ask the right question

The most common mistake is asking "which do you prefer?" That gets you a popularity contest, and popularity is not what you want unless the design's job is to be liked.

Ask instead about the job the design has to do:

The right-hand column produces answers you can act on, because it measures the thing you actually care about. The left-hand column measures whose aesthetic matches the room's.

Ask people who resemble the audience

Your team is the worst possible sample. They have seen every iteration, they know which one the designer prefers, and they have opinions about the process. Whoever the design is for is the only group whose reaction predicts anything.

If you cannot reach them, a general audience is still better than your colleagues, because at least they are seeing it fresh.

Show one thing at a time

Two options side by side, nothing else. No process explanation, no "we tried this and rejected it", no naming the designer. Every piece of context you add moves the answer away from what a real viewer would do, because a real viewer gets one second and no story.

Read the result properly

Anything inside about ten points is a tie. With 40 responses, 55/45 is noise. Treat it as "no meaningful difference" and choose on other grounds — which is easier to produce, which extends better, which the designer thinks is stronger.

A big split is real information. 75/25 across a decent sample is worth listening to even if you preferred the other one. That is the situation the exercise exists for.

Watch for the polarised result. Some designs split a room evenly because half love them and half hate them. That is different from an even split caused by indifference, and often a polarised design is the right choice — it means the work is doing something. If you can, read the comments as well as the counts.

Decide, then say what you decided

A vote is an input. You can override it, and sometimes you should — the audience is not the art director. What you cannot do is run the vote, ignore it silently, and expect people to take part next time.

Say which you picked and what the deciding factor was. If the vote went the other way and you overruled it, say that too. It costs one sentence and it is the difference between "we asked" and "we asked and it mattered".

The fastest version

If you have twenty minutes rather than an hour:

1. Eliminate anything that fails an objective check.
2. Write one question about the job, not preference.
3. Put the two options in front of thirty people outside the project.
4. If the gap is under ten points, let the designer choose.
5. Post the result and the decision.

You can run step three with a two-option poll — put both images in, share the link wherever your audience is, and people can vote without signing up, which matters a great deal for response rates.

For more on phrasing the question so the answer means something, see how to write a poll question people actually answer.