The Research Blindspot Checklist
Fifteen questions to ask before you trust a study, ours included. A market research checklist that follows a project from the brief to the decision, in about ten minutes.
Your score
/15
Answer the 15 questions to see where you land.
In short
Three points | 30 seconds- Bad research rarely looks bad. It arrives with a clean sample, tidy charts and a confident recommendation, and the flaw sits in a step nobody checked.
- These 15 questions follow a study from design to decision, covering who was recruited and how, what happened in the field, and what the analysis can honestly claim.
- Answer yes only when you could show the evidence. With fewer than eight yes answers, the study is not ready to carry a costly decision.
Why we wrote it
Most teams aren’t short of data
They have dashboards, survey tools and a folder of past studies. What they lack is a quick way to tell which of those studies deserves their trust before the meeting where the money is decided.
We wrote this checklist for that moment. It’s the list we run on our own projects, and the one we’d want a client to run on us. Our full standards are public too, in Research quality standards.
Research looks expensive until you price the alternatives
A study with blind spots
It gives a wrong answer with the confidence of a right one. Nobody goes looking for a mistake they trust, which makes this the most expensive option of the three.
No study at all
The decision rests on experience and on the loudest opinion in the room. Sometimes that works, and when it fails, nobody can say why.
A summary nobody checked
AI research tools write fluent reports in minutes. If no one checked who was asked and how, the report carries every error underneath it, in very good English.
How to use it
Pick one study
The last one that shaped a real decision works best: a launch, a market entry, a new price or a repositioning.
Answer with evidence
Say yes only if you could show it: the screener, the guide, the raw data, the decision it changed. If you’re not sure, it counts as a no.
Count, then use it as a brief
Read where you land. Then send the 15 questions to your next research supplier, us included, and ask for the evidence before fieldwork starts.
Phase 01 | Design and recruitment
The foundation
One honest answer from one participant depends on much more than the question. Good design controls what it can and records what it can’t. Three things decide the quality of an answer before anyone asks anything:
- Trust In your brand, in the study and in the person asking.
- Timing Christmas, the summer, an economic shock or a hard week in someone’s life.
- Fit Whether the person in front of you behaves like the customers your decision depends on.
Can you name the decision this study will change?
Before anyone writes a question, someone should be able to name the decision, the budget or the KPI the study feeds, and say which result would change it. Without that anchor, research produces interesting facts that nobody acts on.
Evidence to ask forThe brief, with the decision written in its first paragraph.
Are participants recruited for what they do, as well as who they are?
“Women, 25 to 40, urban” says little about how anyone buys. Strong screeners ask about recent behaviour: what people bought, switched or gave up in the last few months, and why.
Evidence to ask forThe screener, and the behaviour questions in it.
Does the screener keep out professional participants?
Every panel has people who take part for the incentive and know which answers get them in. Rigorous recruitment checks past participation, cross-checks answers for consistency and, for interviews, adds a short call before confirming anyone.
Evidence to ask forThe rules used to reject people, and how many were rejected.
Has someone outside the project checked the questions for bias?
“How much do you like this feature?” invites a polite yes. “Show me how you solved this the last time” invites what really happens. A social scientist with no stake in the hypothesis should review the guide and the questionnaire before fieldwork.
Evidence to ask forThe reviewed guide, with the changes marked.
Does the design check one method against another?
Interviews explain why but cannot tell you how many. Surveys count but rarely explain. Social listening hears what people say when nobody is asking. A strong design says where each method will test the others, so no single one gets the last word.
Evidence to ask forThe plan that shows where the methods meet.
This is the idea behind our H3™ method.
Phase 02 | Depth and rigour
The fieldwork
A lot of research fails before the analysis starts. The same three problems turn up in almost every weak study we’ve seen:
- Bias enters through the field A loose sample or a leading question builds an error into the data that no model can remove later.
- People perform In front of an interviewer they tidy up their habits, rationalise their choices and leave out what embarrasses them.
- Insight needs judgement in the room A good interviewer notices the contradiction and follows it while the participant is still there.
Can your researchers tell the person people want to be from the person they are?
Participants describe their ideal selves: healthier, more rational and more loyal than their receipts show. Anthropology trains you to notice the gap, and to ask about the last time something happened rather than the usual time.
Evidence to ask forOne example where the team caught a socially desirable answer, and what it changed.
Do you listen to what people say when nobody is asking?
Reviews, forums and social posts are written for other customers. Read with method, they show the words people use, the complaints they repeat and the comparisons that never come up in a survey.
Evidence to ask forThe sources read, the period covered and how bots and noise were filtered out.
How we do it: social listening research.
Does someone keep asking why until the answer stops being polite?
“It’s too expensive” is the first answer people give. The useful one usually comes three or four questions later: it doesn’t match what they think that kind of product should cost, or what their partner will say when they see it. That takes a senior interviewer who can improvise.
Evidence to ask forTwo or three transcripts, unedited.
Do you record what people do, as well as what they say?
Pauses, workarounds, the step they skip and the shelf they walk past often tell you more than their words. Observation, diaries and shop-alongs catch the behaviour people would never think to mention.
Evidence to ask forThe observation notes, next to the interview summary.
Does the survey analysis look beyond the average?
An average score can hide two groups who feel opposite things. Growth often sits in the segment that loves you for a reason you never advertised, or in the minority that leaves for one avoidable reason.
Evidence to ask forThe results by segment, with the size of each base.
Phase 03 | Decisions and return
The analysis
Analysis is where research earns its fee, or ends up in a folder. Three conditions make the difference:
- Senior interpretation Patterns, cultural logic and business consequences need someone with years in the field to see them and sign them off.
- Connected data When research, sales and service data use different definitions, the patterns between them stay invisible.
- Clear next steps A finding explains what happened. A recommendation says what to do, why it should work and how you will know.
Is a senior researcher accountable for every insight?
AI tools summarise transcripts in minutes, and they also produce confident mistakes that read well. Junior teams spot problems but often miss what they mean for the business. Someone senior should check every insight against the raw material and put their name to it.
Evidence to ask forWho signs off the findings, and how they checked them.
Do your data sources talk to each other?
Service logs, sales notes and research reports often describe the same customers with different words, samples and definitions. Until they share a common frame, each team trusts its own source and the patterns between them stay hidden.
Evidence to ask forHow the study connects to the data you already hold.
Does the report tell you what to do next?
“Users struggle with onboarding” describes a problem. “Users drop at step three because it asks for a document most of them keep at home; move it after the first payment and test the change” gives you a next step. Every key finding should end in an action you can measure.
Evidence to ask forThe recommendations, each tied to a finding, an owner and a measure.
Can you point to a decision your last study changed?
A launch moved, a price held, a market dropped, a message rewritten. If nobody can name one, research is being bought as a cost and filed as a report. Its return is the decision it improved or the mistake it prevented.
Evidence to ask forThe decisions your last two studies informed, written down.
Can the people who decide find the findings when they need them?
Insight locked in a 120-page PDF reaches the people who sat through the presentation. Product owners and marketers need the findings, the quotes and the data in a place they can search in the week they make the call.
Evidence to ask forWhere the findings will live after the presentation.
How we deliver: the reporting ecosystem.
Your score
Where you land
Answer the 15 questions to see where you land.
/15
- 13 to 15
Decision-grade
Your research is built to carry costly decisions. Keep running the check: the weak point moves as teams, suppliers and markets change.
- 8 to 12
Partial truths
You do research and it helps, but some steps let bias in, most often recruitment and analysis. Fix your no answers in order, because a flaw in the foundation spreads to everything after it.
- 0 to 7
The illusion of knowledge
The research makes decisions feel safer than they are. That’s more dangerous than having no study, because it gives a wrong answer the authority of evidence. Start with the foundation before you commission anything new.
What to do with your no answers
Fix the foundation first
A no in the first five questions affects everything after it. Rework the brief, the screener and the guide before the next round of fieldwork.
Use the questions as your brief
Send the 15 questions to any supplier, us included, and ask for the evidence before you sign. A good team can answer them in a page.
Build the skill in-house
If research runs inside your company, train the team on the steps where you answered no. We teach this on LinkedIn Learning and in company workshops.
See the Academy
First Tuesday