B2B market research agency: what they deliver, the five kinds, and how to judge the method

A business to business market research agency gets you evidence from people who are hard to reach: the finance director who signs off your category, the plant engineer who specifies it, the procurement lead who blocks it. It does that through interviews, surveys, segmentation studies, brand tracking, win and loss analysis and market sizing, and the hard part of the job is not the analysis but the sample. In consumer research you can field a thousand responses in a weekend. In business research your entire universe might be four hundred companies worldwide, so recruitment, screening and incentives are where the money and the skill go.

This page covers what these agencies do, the five kinds you can hire, how to write a brief that produces a decision, how to judge the methodology and what drives the cost. For buying marketing execution rather than research, see what to hire a B2B marketing agency for.

What they actually deliver

Study type	Question it answers	Typical method
Market sizing and opportunity	How big is this market, by segment and geography?	Desk research, expert interviews, bottom up modelling
Segmentation	Which groups of buyers exist and which should we serve?	Quantitative survey plus statistical clustering, validated qualitatively
Buyer journey and decision mapping	Who is involved, in what order, and what do they each need?	Depth interviews across the buying committee
Pricing and feature trade off	What will they pay, and what would they give up?	Conjoint analysis, MaxDiff, Van Westendorp style price testing
Brand and category tracking	Are we known, considered and preferred, and is that changing?	Repeating survey on a fixed sample frame and fixed questions
Win and loss analysis	Why did we win or lose that deal, in the buyer's words?	Short interviews with recent buyers, run by a third party
Concept and message testing	Which of these propositions lands, and with whom?	Qualitative rounds then a quantitative read

Win and loss analysis is the one most companies underuse and the one that pays back fastest. A neutral interviewer gets answers that your own salesperson never will.

The five kinds of supplier

• Full service research agencies. Design, recruit, field, analyse and present. Best when the study matters and you want one accountable owner.
• Sector boutiques. Deep in one industry, so they already know the job titles, the vocabulary and often the people. Fastest route to credible recruitment in a narrow market.
• Panel and recruitment providers. Supply the respondents and the fieldwork while you or another partner design and analyse. Cost effective when your internal team can do the thinking.
• Research platforms. Software for running your own surveys and interviews, with panel access bought alongside. Good for repeatable tracking once the design is settled.
• Strategy consultancies with a research arm. Research bundled with recommendations and change work. You pay consultancy rates and you get consultancy level stakeholder handling.

The common mistake is buying a platform when the real problem is recruitment, or buying a consultancy when you only needed fieldwork. Decide which part of the chain you are short of before you brief anybody.

Writing a brief that produces a decision

The best research briefs start from the decision, not the topic. Put these seven things in writing:

• The decision this will inform, stated as a choice between named options, with the date it has to be made.
• What you already believe and why. Including the belief you most want tested, which is usually the one nobody has written down.
• Who counts as a respondent, by role, by company size, by sector and by country, with the screening question that proves it.
• Sample you consider sufficient, and your reasoning. In business research a well recruited thirty interviews can be stronger than a loosely recruited three hundred responses.
• Existing data you hold, from CRM, support, sales calls and past studies, so the agency does not resell you what you already know.
• Deliverables, named: topline, full report, raw data, verbatim transcripts, segment profiles, workshop. Raw data and transcripts are worth insisting on.
• Who will act on it, and who presents it to them.

How to judge the methodology

Ask these in the pitch and listen for specifics rather than reassurance:

• Where will the respondents come from? Named panels, client lists, trade associations, LinkedIn outreach and expert networks each have different biases. A supplier who tells you the source and its limits is the one doing it properly.
• What are the screening questions? Ask to see them. The screener is where business research is won or lost, because it is what separates the person who really specifies your product from the person who says they do to earn the incentive.
• What is the incentive, and does it distort the sample? Senior business respondents need a real incentive, so ask how they keep it from attracting professional respondents.
• How will the data be weighted, and against what external benchmark?
• Will you show the questionnaire and the full data tables? Yes is the only acceptable answer.
• Which code of conduct do you work to? Reputable agencies name one: the Market Research Society code in the United Kingdom, the Insights Association in the United States, or the ESOMAR and International Chamber of Commerce international code. These cover respondent consent, data handling and how findings may be published.
• How will small base sizes be reported? You want a supplier who flags a subgroup of twelve as indicative rather than presenting it as a percentage.

What drives the price

• Incidence rate. How many people you have to approach to find one who qualifies. This is the biggest single cost driver in business research.
• Seniority of respondent. A chief financial officer costs several times what a general manager costs to recruit and incentivise.
• Method. Depth interviews are priced per interview. Surveys are priced by complete, with the screener and length affecting the rate. Conjoint and segmentation add analysis time.
• Geography and language. Each market adds recruitment, translation and local fieldwork.
• Deliverables. A topline deck is cheaper than a full report, segment profiles and a workshop. Decide what you will actually use.
• Repeat waves. Tracking is cheaper per wave after the first, because the design is already built.

Ask for the quote broken out by those lines. Two proposals that look similar in total often differ enormously in sample quality, and the breakout is where you see it.

Where research and competitive analysis meet

Research tells you what buyers think. Competitive analysis tells you what rivals are doing. Run them together and you get the one output that changes decisions: a clear view of what your market values, who currently owns each of those positions and which one is available to you. Our repeatable competitor analysis method is the desk research half of that pairing, and it is worth completing before the fieldwork so the interviews spend their time on what you do not already know.

Questions buyers ask

How many interviews are enough?

For qualitative work, run until new interviews stop producing new themes, which in a defined role and market commonly happens somewhere between fifteen and thirty. For quantitative work the number follows from how finely you need to cut the data: every subgroup you want to report on needs its own usable base.

Can we just survey our own customers?

Your customer list answers questions about satisfaction, expansion and advocacy very well. It cannot tell you why people who did not buy chose somebody else, which is usually the question with the most money attached. Mix your own list with externally recruited non customers.

How long does a study take?

A focused interview study runs in weeks: design, recruitment, fieldwork, analysis, presentation. A segmentation or conjoint study runs in months because the design and the analysis are heavier. Build recruitment time in generously for senior or niche audiences.

Should the agency know our industry already?

Sector knowledge speeds up recruitment and screener design, which is where it pays. Methodological strength matters more for anything statistical. Where you have to choose, buy method strength for a segmentation study and sector depth for an interview study.

Do we own the data?

Settle it in the contract. Ask for the raw data, the data tables, the questionnaire and the anonymised transcripts as named deliverables, and confirm you may reuse findings in your own marketing, subject to the code of conduct the agency works to.

Is DIY research ever the right answer?

Yes, for repeat tracking on a settled questionnaire, for quick concept checks with your own customers, and for pilot work before you commission something larger. Bring in an agency when recruitment is hard, when the analysis is statistical, or when the finding has to withstand challenge from your own board.

Your brief to a research agency gets answered twice

A buyer hiring a B2B research agency writes a brief, sends it to a few firms and compares. Before that, more and more of them ask an assistant who does this work, and the names it gives set the long list. A firm with excellent fieldwork but a thin description online does not make it. Look at your method pages: do they state sectors, sample types and where respondents are recruited from?

AI Knows Us records which agencies assistants name for such questions.