Concept Testing

What it is

A research method that shows target customers a short description of a product, feature or offer that does not yet exist (or exists only as a prototype) and measures how they react. A concept is usually one page or one screen: the problem, the proposed solution, the main benefits, and often a price. Respondents rate it on appeal, relevance to their needs, uniqueness, believability and purchase intent, then explain their ratings in their own words. The method has no single originator; it grew out of consumer goods research and is now standard practice in survey platforms. There are two main designs. In a monadic test, each respondent sees only one concept, which gives the cleanest score but needs a separate sample per concept. In a sequential monadic test, each respondent rates several concepts in random order, which needs fewer people but lets one concept colour the next. Concept testing differs from message testing: concept testing asks whether customers want the offer, while message testing asks which words best describe an offer that already exists.

When to use it

• Before committing build budget to a new product or major feature, when product and PMM need evidence of demand, not just an internal belief.
• Choosing between two or three product or packaging ideas aimed at the same segment, where a monadic score per concept gives a fair comparison.
• Entering a new segment (a market development move), to check that the new buyers want the existing product framed for them before a full launch.
• Testing a new bundle, tier or pricing model with customers before it goes on the price page.
• A product team is split on which of several directions to pursue, and a structured customer read would settle the argument faster than more internal debate.

How to run it

1. Write the decision first. State what you will decide with the result: build or not, which concept to pursue, which segment to lead with. Set the threshold that would change the decision, such as "at least 30% of respondents say they would definitely or probably buy".
2. Write each concept on one page. Include the problem, the solution in plain language, three main benefits, and a price if price is part of the decision. Keep every concept the same length and format so differences in score reflect the idea, not the writing.
3. Choose the design. Use monadic when concepts are similar, the stakes are high, or you need a score to benchmark against future tests. Use sequential monadic for early screening of three to five ideas on a tight budget, and randomise the order.
4. Recruit the right respondents. Screen for the target segment and buying role using your ICP. Aim for at least 100 to 150 qualified respondents per concept in a monadic design; fewer than that and differences between concepts are often noise.
5. Ask the standard measures. Rate appeal, relevance, uniqueness and believability on a five-point scale, then purchase intent ("definitely would buy" to "definitely would not"). Add one open question: "What, if anything, would stop you buying this?"
6. Analyse top-two-box scores by segment. Report the share of respondents choosing the top two points on each scale, overall and by segment. Read the open answers for the objections that recur most.
7. Compare against a benchmark. Compare scores with an existing product tested the same way, or with your own past tests. A purchase-intent score means little on its own: researchers widely agree that people overstate what they will buy.
8. Hand the result to the decision owner. Give product the scores, the main objections and a recommendation. Feed the winning concept's customer language into positioning and launch messaging.

Cadence & ownership

PMM commonly runs concept tests, especially at growth-stage companies without a dedicated research team, because it holds the segment definitions, the positioning and the launch plan the result feeds. The build decision itself belongs to product. At a scaled company, a consumer insights or user research team may run the fieldwork, with PMM writing the concepts and defining the target sample. Run a test whenever a significant new product, tier or segment entry is proposed; a typical online concept test takes two to three weeks from brief to result.