How to concept test a product pack, claim, or message this week

Junu Yang · 2026-10-01

You have a pack, a claim, or a message, and a decision that will not wait for a six-week study. This is the week: what to put in front of people, how many interviews to run, and what to do with the quotes when they come back.

The questions live in concept testing questions. Use that list when you write the guide. This is the plan around it, not a second question bank.

Write the decision in one sentence

Before you pick a stimulus, write the decision the quotes have to change. Which pack direction do you lock? Do you keep this claim on the front? Which line goes in the ad? If you cannot name the decision, the interviews come back as interesting comments and nobody knows what to do on Friday.

One decision, one stimulus. A pack, a claim, and a message are three tests. Putting all three in one interview mixes the reasons together, and you will not know which one caused the hesitation.

Show one thing at a time

A pack is judged in a glance. People say what they think the product is, what they expect it to cost, and whether they would pick it up. A claim has to be believed. People say what it is saying, what they doubt, and what proof they would want. A message has to be understood and chosen. People say what they think it wants them to do, what they would remember, and which line they would actually use.

Show one concept at a time. If you have two routes, give each one its own conversation. A single score across both routes hides which one people could say back.

The shape of the week

Day one is the decision and the stimulus. Write the sentence, pick the pack or the claim or the message, and stop adding variants. Two routes are fine if each one gets its own set of interviews. Five routes are a different study.

Day two is the guide and the fielding. Take a short set from the question bank, matched to what you are testing, and send the interviews out. Do not spend the day polishing a discussion guide that tries to cover product, pack, claim, and message at once.

Then you wait. Quotes come back in 48 to 72 hours, so interviews that go out on Tuesday are back Thursday or Friday. Use that gap to decide, in advance, what would make you keep the work, rewrite it, or kill it. If you invent the rule after you have read the quotes, you will talk yourself into the option you already liked.

The last day is the read. Sort what people said against the decision you wrote on day one, and write down the change before the week ends. A readout that slips into the next sprint is how a concept test becomes a file nobody opens.

What you put in front of them

A pack test needs the front of the pack, large enough to read. Not a strategy slide, not a mood board, not three routes on one page. People should be able to say what they think the product is from that one image.

A claim test needs the line as it would appear on the pack or the ad. Not a paragraph that explains why the claim is true. If they need your explanation to understand it, the claim is not ready.

A message test needs the line, or the ad, as a person would meet it. The headline, the line under it, and what you want them to do next. Cut the internal language before you show it. People cannot react to a message they only meet inside your deck.

Keep the guide short

Pull a short set from concept testing questions, matched to the thing you are testing. You need enough to hear three things: what they understood, what they doubted, and what they would actually choose. You do not need the whole bank in one sitting.

Ask them to say it back before you ask whether they like it. A yes to "would you buy this?" is easy to give. The useful answer is what would stop them.

Run 15 to 30 interviews

For a pack, a claim, or a message, run 15 to 30 interviews. That is enough to hear the same confusion, the same objection, or the same reason for choosing, more than once. It is not a scored survey, not a monadic benchmark, and not a large quant test. It will not tell you market share. It will tell you whether people can explain the idea and whether they would choose it.

Quotes come back in 48 to 72 hours. Read them while the decision is still open. A study that lands after the pack is locked is a recap, not a test.

Talk to people who would actually choose, not whoever answers first. 15 to 30 of the right people will repeat a real objection. A larger group of the wrong people will repeat a polite one. The sprint does not get more true because the number gets bigger. It gets more true because the same reason shows up again.

Read the quotes against the decision you wrote

On a pack, listen for the first takeaway and the miss. If people describe a different product than the one you designed, the direction is not ready. On a claim, listen for the objection. A claim people like and do not believe will not survive the shelf. On a message, listen for the line they can repeat and the line they would cut. If they remember the joke and not the point, the message failed.

Sort the quotes by what you would change: keep, rewrite, or kill. If five people describe the concept five different ways, that is the finding.

Decide before the week ends

Write down what you will change, and what you will leave alone. A pack people cannot place, a claim they will not trust, or a message they cannot say back should not go forward unchanged. The point of the week is a sharper decision, not a prettier slide.

Run the week on concept testing: 15 to 30 AI-moderated interviews, with quotes back in 48 to 72 hours.