Evidence on how people respond to your ideas, why, and how to make them stronger, so your team can choose which to develop.
A norm-based score is often the right tool for a repeat test of a familiar format. For a new idea, your team also needs the reasons behind the score.
Say your team is testing ideas for a new interior paint. Here is one, read against the paints people already buy.
An illustration, not a client project.
The score, in words
Above the comparison set on appeal; below it on believability.
Wanted, and doubted: your team needs to know why.
The concept as testedInterior wall paint
Paint it today. Sleep in it tonight.
Understood
People got the promise at once: a room back in use the same day.
A water-based wall paint for bedrooms, with almost no smell, even with the windows closed.
Doubted
Almost no smell went further than people’s own experience of painting, and the closed windows read as a health promise.
Why it works
Made without the solvents that give paint its smell.
Believed
The reason made sense, but on its own it didn’t settle the doubt.
So your team keeps what people understood and believed, and rewrites what they doubted, with a way to check.
Rewritten with your teamInterior wall paint
Paint it today. Sleep in it tonight.
Kept
The promise people wanted, and understood.
Rewritten
A water-based wall paint for bedrooms, with a low smell that fades as it dries, so the room is ready by bedtime.
Why it changed
A smell that fades matches what people know from painting. The closed windows came out.
Rewritten
Why it works
Made without the solvents that give paint its smell. Smell the open tester in store before you buy.
Why it changed
The reason stays, with a check people can make for themselves.
Read again, after the rewrite
Still above the set on appeal; now in line with it on believability.
Whether a tester in every store is worth its cost is for your team to decide.
When the choice is a whole range, or the features and prices that go with it, a choice model reads those trade-offs directly: Conjoint and choice modeling.
In the same illustration, here is the paint beside your team’s other ideas, read against the same comparison set.
| Idea | Appeal | Believability | The reason, written down | What your team decides |
|---|---|---|---|---|
| Low-smell paint, rewritten | Above the set | In line with the set | Wanted, and now believed. | Develop it |
| Peel-off color for renters | Above the set | Below the set | Renters doubted it would come off cleanly; only real walls can show it. | Test it in use first |
| Paint with primer built in | In line with the set | Above the set | Easy to believe, but little that people don’t already buy. | Set it aside |
The comparison
A norm compares your idea with the average of concepts tested before it, most of them other companies’. Intentions predict purchases of existing products better than new ones (Morwitz and colleagues, 2007), so we read a score against products that already sell.
For a food company, volume estimates set each screened concept against products already on sale. Products in the program’s opportunity areas later reached market.
Claims
Consumer research shows which claims people understand, believe and care about. Whether a product can substantiate a claim is a separate question, for your regulatory and legal specialists.
For a personal care company’s new hair care product, women in focus groups in two markets rebuilt the claims themselves, putting the benefit areas in order and prioritizing the claims within each.
A workwear brand had garments made up in a new fabric, and had to decide whether to go ahead with it. Its lab tests could say how the fabric performed, but not whether someone working in a warehouse would feel the difference over a long working day.
Workers in physically active jobs across five sectors wore the garments through their working days, with a short video report and survey each day. The brand received their accounts of how the garments performed and of their comfort, fit, styling and laundering, to weigh alongside its own lab and wash tests.
What happened next
The brand later launched a workwear range built on the fabric we tested.
Match the option to the question. Repeat tests of familiar formats work well in-house or on a platform. A new kind of idea, a large development investment, or a portfolio choice that several teams must agree on benefits from help setting the comparison before fieldwork and working through the results together.
We model volume where the plan needs a number, for example by combining penetration, frequency and units, and read it against products that already sell. A launch forecast needs more: the awareness, distribution and sampling the plan will actually achieve (Urban and Katz, 1983), which go into the model as inputs from your marketing plan.
When launching is cheap and reversible, test in market. When the risk is in use after purchase, run a product test first. When the idea is so new that people struggle to judge it from a description (Hoeffler, 2003), start with qualitative work and a prototype. And when no result would change your next step, don’t buy the test.
Hoeffler, Measuring Preferences for Really New Products, 2003
Our own senior team runs the work, from trends through ideation, and our own data scientists and statisticians design the screens and build the volume models, in-house.
Send us the concepts as they stand, or the one idea everyone is attached to, and when your team needs to choose. We’ll suggest what each idea should be compared with, and where qualitative work or a product test would help first.
A partner or director will reply.