Concept testing for products and communications

We begin with a testable business question, not a preferred technique. We align the action that follows the result and standardize stimuli so comparisons are fair. Monadic exposure reduces contamination; a sequential design can support diagnosis when order is controlled. Before fieldwork, we turn assumptions into hypotheses, variables and interpretation rules. This makes it possible to separate an actionable signal from a superficial difference and prevents the team from collecting information that would not change the course of the project.

Scope is aligned with leadership, marketing, insights, innovation or product stakeholders. We document the target population, geography, competitive context, downstream actions and operational constraints. For multi-country Latin American work, we preserve a comparable core while adapting language, stimuli and quotas to each market. A literal translation is never treated as evidence of cultural equivalence.

The research method follows the required level of evidence. In-depth interviews can refine language and uncover objections. Quantitative testing then estimates levels and differences. MaxDiff may rank benefits or messages, while conjoint analysis is useful when acceptance depends on combinations of attributes, formats and price. CAWI research with an online panel is often efficient for digitally reachable audiences; CATI can improve coverage when access or incidence makes an online-only design weak. Ethnography and in-depth interviews reveal context, tensions and vocabulary before those patterns are quantified.

Sample size is based on population heterogeneity, the number of cells or segments, expected incidence and the comparisons the study must support. Margin of error belongs to a specific design and estimate, so we state its assumptions rather than attach a generic figure to every result. Small subgroups are labelled exploratory instead of being presented as representative.

Questionnaires and discussion guides are piloted. We monitor quotas, survey speed, internal consistency, duplicates and open-end quality. When stimuli are tested, exposure is randomized and order effects are controlled. The analysis plan is agreed before results are reviewed, including weighting, exclusions, missing values and thresholds for interpreting differences.

The core deliverable is not a collection of charts. It is a traceable recommendation: each conclusion points to supporting evidence, states the level of confidence and identifies the action it can inform. We provide an executive narrative for alignment and a technical appendix covering the instrument, sample, quality controls and definitions.

A typical package includes concept scorecard, promise and support diagnosis, comparison of alternatives, revision guidance, criteria for the next stage. When the client needs to continue the analysis internally, data files, codebooks and tables are specified at the proposal stage. Materials separate observed facts, interpretation and implications so that a recommendation cannot drift into an unsupported claim.

The final session is a working meeting. We connect evidence with internal knowledge, answer technical questions and prioritize next steps by impact, effort and risk. If material uncertainty remains, it is made explicit along with a proportionate way to resolve it; we do not manufacture certainty that the design cannot provide.

A focused quantitative concept test usually takes four to six weeks. A qualitative-to-quantitative sequence may require six to nine weeks. Timing depends on incidence, number of markets, recruitment complexity, stimuli and approval speed. A realistic plan separates design, programming or preparation, pilot, fieldwork, quality control, analysis and stakeholder readout. Compressing any stage changes risk and should be an explicit trade-off.

The same senior team defines the study, supervises fieldwork, interprets the evidence and presents the recommendation. Clients have one accountable lead and clear checkpoints: validated brief, approved instrument, field update and final findings session. This model reduces the context loss that often occurs between sales, operations and analysis.