---
name: Market Research & Intelligence
description: Design research that yields decisions, not decks — sound experiments and unbiased surveys, plus qualitative depth (interviews, observation) fused with quantitative scale — so you understand real customers from evidence instead of the loudest opinion in the room.
audience: product manager · UX researcher · growth marketer · founder
---

# Market Research & Intelligence

## What this is
A method for gathering decision-grade evidence about a market and its customers: designing experiments and surveys that don't bias their own answers, running qualitative work (interviews, observation) for depth, and triangulating the two into an insight you can act on.

## What this is NOT
- **Not leading-question theatre.** A survey engineered to confirm a prior is worse than no survey; this skill hunts and removes bias in the instrument, the sample, and the analysis.
- **Not extractive or covert.** Research is consented, participants' data is protected (GDPR), and it does not deceive, over-incentivise, or surveil.
- **Not a substitute for a trained researcher** on high-stakes studies or regulated claims; it designs and triangulates, and routes rigor-critical work to a specialist.

## When to use
Sizing or segmenting a market; designing a survey or an experiment; planning and synthesising interviews; combining qual and quant to explain a "what" with a "why"; deciding whether a signal is real before you build.

## Operating principle
The design decides the truth. A biased instrument or sample produces confident nonsense — so bias control comes before data collection, and every finding names its sample, its method, and what it cannot claim. Qual explains, quant confirms; neither alone is enough.

## Capabilities
- **Experiment & survey design** — Goal: instruments that don't lie. Method: define the decision first, choose method to match, write neutral non-leading questions, control order/scale effects, plan sampling and size for the precision needed, pilot before fielding. Output: a research plan + the instrument + a bias-check note. Quality bar: questions pass a leading-language check; the sample's representativeness (and its limits) is stated up front.
- **Qualitative depth** — Goal: the why behind the numbers. Method: semi-structured interviews and behavioural observation over stated preference, probe for actual behaviour and unmet needs, code transcripts for themes with saturation tracked. Output: themed findings with verbatim evidence + confidence by theme. Quality bar: insights are grounded in what people did/said, labelled by strength of evidence, not the one memorable quote.
- **Triangulation & sizing** — Goal: one trustworthy read. Method: fuse qual themes with quant scale, size the market/segment from defensible inputs (TAM/SAM/SOM with sources), reconcile where they disagree rather than picking the convenient one. Output: a synthesised insight + a sized opportunity with assumptions. Quality bar: qual and quant corroborate or the tension is explained; every size is labelled estimate with its inputs.

## A worked example
"Everyone's asking for feature X." → Interviews (n=12) reveal the real job is *trust*, and X is one imagined fix; a neutral survey (representativeness stated) shows only 18% would use X, while 60% cite the underlying trust gap. Triangulated: build for trust, not X. The write-up names the sample, flags the survey's coverage limit, and labels the sizing as a modelled estimate.

## Guardrails & escalation
Regulated or high-stakes claims → a trained researcher + defensible methodology. Personal data → consent + GDPR handling. Persona artifacts → the Persona Development skill; competitive intel → the Competitive Analysis and Market Position skills. Every size labelled estimate.

## References
Survey-methodology and experimental-design canon (question wording, sampling, bias control); qualitative methods (grounded theory, ethnographic observation, saturation); TAM/SAM/SOM sizing; GDPR Art. 5–6 for participant data. Verify rigor-critical designs with a specialist.