Jean Technologies Icon
Jean Technologies Text

Test ten concepts, keep the best two

Score every idea on digital twins in minutes. Take the strongest to a hundred real interviews overnight, and see which one clears your bar and why.

Run a studyBook a demo

Plant-based snack bar, six conceptsIllustrative · Monadic, 300 digital twins per concept

Percent top-2 box. Letters mark a score significantly higher than that concept at 95%.

To real interviewsTo real interviews
AProtein firstBKids lunchboxCLow sugarDUpcycled oatsEDessert swapFGym bag
Would buyAction standard 45%49BDF3354BDEF2741D35
Uniqueness42BF3147BEF58ABCEF3629
Relevance63BDEF46D66BDEF3851D52D
Believability66DEF59D63DE4152D55D
Value for money44D50D52D3547D48D
Top likeKeeps me fullEasy for schoolNot too sweetFeels good to buyTastes like a treatFits in a pocket
Top dislikeChalky textureKids won't eat itProbably blandWhat is upcycled?Not really healthyNothing special

Action standard: would buy 45% or more

  • A Protein firstTo real interviews

    Would buy49BDF
    Uniqueness42BF
    Relevance63BDEF
    Believability66DEF
    Value for money44D
    Top like
    Keeps me full
    Top dislike
    Chalky texture

  • B Kids lunchbox

    Would buy33
    Uniqueness31
    Relevance46D
    Believability59D
    Value for money50D
    Top like
    Easy for school
    Top dislike
    Kids won't eat it

  • C Low sugarTo real interviews

    Would buy54BDEF
    Uniqueness47BEF
    Relevance66BDEF
    Believability63DE
    Value for money52D
    Top like
    Not too sweet
    Top dislike
    Probably bland

  • D Upcycled oats

    Would buy27
    Uniqueness58ABCEF
    Relevance38
    Believability41
    Value for money35
    Top like
    Feels good to buy
    Top dislike
    What is upcycled?

  • E Dessert swap

    Would buy41D
    Uniqueness36
    Relevance51D
    Believability52D
    Value for money47D
    Top like
    Tastes like a treat
    Top dislike
    Not really healthy

  • F Gym bag

    Would buy35
    Uniqueness29
    Relevance52D
    Believability55D
    Value for money48D
    Top like
    Fits in a pocket
    Top dislike
    Nothing special

Overnight: 100 real interviews, protomonadic

A Protein firstWould buy 46%

Fix before launch: the texture worry came up in 31 of 100 interviews.

C Low sugarWould buy 52%

“Most bars taste like candy. This one sounds like food.”

Final choice, head to head

Run it with

Methods

  • Idea screening

    Rank a long list of rough ideas on appeal and uniqueness, so only the best few get a full test.

  • Monadic test

    Each person sees one concept, which gives the cleanest read to hold against your action standard.

  • Sequential monadic test

    Each person rates several concepts in rotated order, so you need fewer people per concept.

  • Protomonadic test

    Rate concepts one at a time, then pick a favorite head to head, so you get both scores and a winner.

  • Concept scorecard

    Top-2 box on purchase intent, uniqueness, relevance, believability and value, with significance letters and the top like and dislike.

  • Screen the whole shortlist first

    An agency test is priced per concept, so teams cut ideas before anyone sees them. Twins let you score all ten, and real interviews go to the few that earn them.

  • Scores with the why attached

    A rating tells you a concept is weak. The interviewer asks what drove it, so each concept comes back with its top like, its top dislike and the words people used.

  • Fix it and test again

    A concept that misses is often one phrase away. Reword the claim, rerun it on twins in minutes, and send the new version to people the same night.

How you run it

  1. 1Minutes

    Screen

    A thousand digital twins answer first, so the weak options drop out.

  2. 2Overnight

    Interview

    The AI interviewer talks to real people and follows up on what they say.

  3. 3By morning

    Decide

    The number, the reasons behind it and the quotes that carry it.

Questions

What is the difference between monadic and sequential monadic testing?

In a monadic test each person sees one concept, so there is no comparison bias but you need a separate group per concept. In sequential monadic each person rates several concepts in rotated order, which needs fewer people but lets earlier concepts color later ones.

What metrics does a concept test measure?

Purchase intent is the main one, usually reported as top-2 box: the share who say they definitely or probably would buy. Most tests add uniqueness, relevance, believability and value for money, plus open questions on what people liked and disliked.

What is an action standard?

It is the bar a concept must clear to move forward, set before fieldwork. For example: top-2 box purchase intent of 45% or more, and significantly ahead of the current product. Jean has no category norms of its own, so most teams set the bar from their own past tests.

How many people do I need per concept?

For a monadic read, 100 to 150 people per concept is common. At that size, gaps under about ten points are hard to call. Twins let you add hundreds per concept at the screening stage at little cost.

What kinds of concepts can I test?

Anything you can show as text, an image or a video: product ideas, packaging, claims, names and positioning lines. Jean does not run in-store or in-home product use tests.

Are digital twin scores accurate enough to make the call?

Use twins to rank and cut a long list. Make the final call on the concepts that survive with real interviews, which Jean runs overnight.

Talk to us.