Market Research Checklist

The PMM or insights lead drafts a one-page brief: business question, decision the research will inform, scope, audience, deliverable format, deadline. Avoid kicking off without a named decision-maker — research without a decision attached tends to produce decks no one reads.

Examples: positioning for the Q3 product launch, repricing a tier, prioritizing a new ICP segment, choosing between two campaign concepts. If the answer is "we want to learn more about customers," push back until there is a concrete decision and decision date.

Specify firmographics (industry, employee count, revenue band) and persona (title, function, seniority). Tight criteria reduce screen-out rates during recruitment and protect the cost-per-complete in panel-based studies.

What does "done" look like? Examples: directional signal at n=200 with ±7% margin of error, three persona archetypes documented, top-five purchase drivers ranked by segment. Tie KPIs back to the decision in step 11.

Confirm budget for panel/incentives, tooling (Qualtrics, SurveyMonkey, dscout, UserTesting, Respondent.io), and any external moderator fees. Lock the readout date with the stakeholder before fieldwork begins.

Qualitative (IDIs, focus groups, diary studies) for "why" questions and exploratory work. Quantitative (surveys, conjoint, MaxDiff) for sizing, ranking, and significance testing. Mixed methods when the decision needs both — for example, a qual phase to surface drivers, then a quant phase to size them.

Set quotas by segment, region, and any required cuts. Document screen-in and screen-out criteria for the panel. For B2B work, expect 8–15× incidence rate in cost-per-complete estimates; budget accordingly.

Order questions least-to-most leading. Use validated scales (NPS, CSAT, intent) where comparability matters. Avoid double-barreled questions and unbalanced response options. For B2B IDIs, target a 45-minute guide with branching probes.

Watch for confused respondents, straight-lining, premature drop-offs, or items where the answer distribution is degenerate (everyone picks 5). Time the survey — anything over 12 minutes will tank completion rates on a panel.

Address every issue surfaced in the pilot before going to full field. Re-version the instrument and circulate to the stakeholder for sign-off; once fielding starts, mid-stream changes invalidate prior responses.

For first-party customer research, suppress anyone in active sales cycles or open support escalations. For panel-sourced studies (Respondent, User Interviews, Prolific), confirm the screener catches fraudulent or AI-generated responses — attention-check questions and trap items.

Soft-launch the survey to 10–20% of sample first; check completion time, drop-off points, and quota fill before opening fully. For IDIs, schedule no more than 4 per day per moderator — fatigue degrades probing quality.

Flag respondents who fail attention checks, complete in under one-third the median time, or straight-line Likert grids. Track quotas by segment so you don't end up with 80% one industry. Adjust the panel send if a cell is lagging.

Remove flagged respondents, recode open-ends, reverse-code negatively-worded items, and weight the dataset if quotas didn't fill exactly. Document every exclusion rule in the methods appendix — reviewers will ask.

Run the cross-tabs and significance tests the decision actually requires — not every cut imaginable. For qual, work from the discussion guide back to themes with a coding framework (two coders, reconcile disagreements). Note effect sizes, not just p-values.

Open with the decision, the recommendation, and the top three supporting findings. Stakeholders skim — front-load the answer. Detailed cuts and methods belong in the appendix.

One chart per finding, titled with the takeaway ("Buyers rank security #1") not the metric ("Top-box rankings by attribute"). Show base sizes on every chart. Use brand colors from the visual identity guide; avoid red/green encodings for accessibility.

Standard structure: executive summary, decision and recommendation, key findings, segment cuts, persona profiles, implications, methods appendix, questionnaire appendix. Internal review pass before the executive summary is written.

One page maximum. Decision asked, recommendation, top three findings, three implications. The CMO or VP should be able to brief the CEO from this page alone without opening the full deck.

30-minute readout: 10 on findings, 20 on what to do about them. Capture pushback, alternative interpretations, and follow-up requests in the discussion notes — those become the activation backlog.

Work through requested revisions, document any analytical changes in the methods appendix, and re-circulate to the stakeholder group with a short change log. Do not silently rewrite findings — the audit trail matters when results inform a launch decision.

An insight ties a finding to a specific change: "Mid-market buyers cite integration breadth as the #1 driver, so the integrations page should be promoted to primary nav." Drop findings that do not produce an insight; they belong in the archive, not the activation plan.

Map recommendations to owners across PMM, demand-gen, content, and brand. Each recommendation gets a target ship date and a measurable outcome (e.g., "ship the new positioning page by EOQ; track organic conversion rate vs. control").

Bring the prioritized list to the next planning ceremony (quarterly planning, sprint planning, or campaign review). Recommendations that don't make it into a planning doc within two weeks of readout almost never ship.

Track the activation work in the same system as the rest of marketing's roadmap (Asana, Monday, Jira). Tag tickets with the research project ID so impact can be attributed back at the 90-day review.

Pull the metrics tied to each recommendation in GA4, the MAP, and the CRM. Write a one-page memo: what shipped, what moved, what didn't, and what to study next. This memo is what funds the next research request.