Concept testing companies: choose who owns the decision

Choose a concept testing company by the work it will own: bespoke design, a repeatable program, serviced diagnostics, portfolio context or choice against current alternatives. Assess Decision Analyst, Zappi, Kantar, NIQ BASES and Ipsos against one brief. Require a source-to-recommendation walkthrough and included iteration, especially when a high-interest concept may be misunderstood.

Sapien publishes this comparison and is included separately.

Commission a breakfast development decision

Imagine you run a food brand and must choose a single-serve breakfast pot or a family-size kit for prototype development. In this illustrative brief, the buyer’s breakfast occasion and preparation needs should shape the choice. You need a provider that connects appeal, product feasibility and the next investment, including any preparation or value barrier to resolve first.

The study-ownership ledger for a breakfast decision

Use the responsibility map to see which work the provider accepts and which inputs your team still owns.

A proposed commissioning ledger for comparing a single-serve breakfast pot with a family-size kit. Each responsibility has a provider owner and an explicit brand-team input; the deliverable is a development decision, not just a score.

Both cards should state occasion, contents, preparation, pack size, proposed price and comparable image detail. Include what eligible buyers currently use when the task concerns switching. Beating an implausible alternative can produce a winner with little practical meaning. Keep a separate name-only test separate from reactions to the complete breakfast offer.

Write the progression rule before fieldwork. Understanding, feasibility and sufficient appeal under the team’s economics belong together. A universal launch cutoff cannot replace those conditions, and a forecast needs an explicit account of the launch assumptions.

Five providers imply different operating arrangements

Use this as a shortlist by research job. The provider sections below explain which design, audience, interpretation and iteration responsibilities still need agreement in the proposal.

Decision Analyst: bespoke design and interpretation

Decision Analyst offers researcher-led design and analytical consulting for a custom concept decision. Confirm audience, fieldwork, analysis and who resolves a preparation misunderstanding.

Zappi: a repeatable testing program

Zappi combines a platform with professional services for tailored standard solutions, custom audiences and program support. Separate licence, sample, custom work and interpretation in the quote, and specify human or synthetic response sources.

Kantar ConceptEvaluate: standardized diagnostics with a serviced route

Kantar ConceptEvaluate offers study setup, a dashboard, written evaluation and a human researcher readout in its serviced scope. Self-serve and serviced arrangements differ: confirm diagnostic modules, audience, concepts and included iteration.

NIQ BASES: portfolio and reference context

NIQ BASES combines concept optimization, portfolio context and expert consulting with databases and predictive analytics. Check whether the category and launch conditions fit the reference evidence. A forecast still depends on its assumptions.

Ipsos InnoTest: choice against current alternatives

Ipsos InnoTest offers choice-based diagnostics, acceptor profiles and optional modules against the products or solutions consumers currently use. Assisted and self-service routes differ. Confirm design, recruitment, interpretation and the evidence behind a Trial Index before treating assisted access as a fully managed project.

Primary pages were checked on October 6, 2026. The comparison describes advertised scope and decision fit; it is not a hands-on quality ranking. Decision Analyst is relevant when custom design and analytical consulting dominate. Zappi is relevant when a repeatable platform program needs tailored questions, audiences or services.

Kantar ConceptEvaluate explicitly distinguishes serviced and self-serve routes. In the serviced scope, compare the setup, dashboard, written evaluation and researcher readout. NIQ BASES supplies innovation context supported by databases and predictive analytics; inspect reference fit and launch assumptions. Ipsos InnoTest emphasizes current-alternative choice and diagnostic outputs, with assisted and self-service access. Assisted access alone does not establish full project ownership.

Start with the responsibility the team cannot supply

If the two offers need a bespoke instrument or difficult interpretation, begin with a researcher-led proposal. If the team is building a recurring program, ask where platform services begin and end. If a standard diagnostic study fits, compare exactly which setup and readout are included. If portfolio or forecast implications drive the investment, ask which reference evidence and commercial assumptions support them.

Do not use the number of services on a website as a substitute for a written responsibility map. A license, purchased responses, assisted setup and full interpretation are different scope items. Zappi also advertises a separate synthetic screening route, so specify the response source used by the quoted solution.

Use the ledger to expose the unfinished work

The ledger above assigns design, audience, interpretation, iteration and acceptance. The brand supplies feasible cards and product facts; the partner should state how it handles assignment, quality, analysis and recommendations. Agree which revision is included and what happens if the first concept is misunderstood.

Ask for an illustrative or redacted walkthrough that connects one unfavorable answer to an action. A provider can supply polished scores while leaving the team to resolve the only fact that matters. Inspect who owns the preparation question and whether a follow-up is part of the project rather than a new purchase discovered at the end.

A favorable family-kit score can hide the wrong experience

Suppose the family kit leads on interest, but unaided answers describe it as ready to eat. The intended product requires preparation. The scored stimulus and the imagined experience are no longer aligned. High interest cannot justify prototype investment without resolving that mismatch.

Request the exact card, preparation question, coding rule and comprehension base. Check whether people who understood preparation reacted differently from those imagining immediate consumption. That diagnostic comparison can help locate the problem, but it does not turn a post-hoc subgroup into a new randomized test. Show its role as explanation and specify the comparison needed next.

The complete next step is a bounded repair

First confirm that the card accurately and visibly states preparation. If the fact is missing, repair the card. If the complete offer has an unattractive preparation burden, the development team may need to revise the product or positioning rather than add persuasive copy. Those are different actions, so the report should identify which evidence supports each.

For the breakfast case, revise the family-kit description and reassess comprehension before progressing it on the strength of the original interest score. Keep the underlying product and other relevant facts fixed for a communication repair. If the product changes, save a new concept version and treat the follow-up as evaluation of that new offer.

A separate managed modeled route

Our managed research service compares options with simulated populations. Assess grounding, validation and the agreed recommendation for the concept decision. Physical preparation, taste and repeat purchase need corresponding evidence before supporting launch claims. A modeled comparison can guide a scoped development choice while leaving those next questions explicit.

Accept the decision with its working record

Retain the instrument, cards, exposure, eligible segment bases, exclusions, material assumptions and next action in the agreed format. Keep human and modeled response sources labeled. For a standardized diagnostic index, inspect what the index means and which reference evidence fits; do not rename it as actual purchase probability.

Our published foldable-iPhone report illustrates modeled qualification and switching. It is an example of a report, rather than observed product demand. Use any sample report to inspect the path from source to finding to recommendation and the evidence still required.

Send two suitable partners the same breakfast brief

Choose the first conversations around the hardest task, then send the matched cards, ownership ledger and progression criteria. Request audience feasibility, evidence route, included iteration and complete price. Compare which partner finishes the development decision at a workload and scope the brand can support.

For the preparation-misunderstanding case, accept a provider that names the unresolved experience and owns a bounded reassessment. Ordering a broader survey without resolving that fact adds volume while leaving the prototype choice unfinished.

Further reading

• our name-testing guide
• our concept-testing guide
• our concept testing tools guide
• our guide to market research services and costs
• our synthetic versus human respondents guide
• our market research companies guide

For a managed modeled breakfast comparison, bring both complete cards, the preparation question and the required validation and reassessment to a scoped Sapien study discussion.