Blog Content
Is Your B2B Market Survey Getting Honest Answers?
You have three weeks to field a B2B market survey, a list of procurement managers who have agreed to take part, and a questionnaire that looks perfectly reasonable. Almost all of them will respond. Almost none of them will lie to you. And the results can still push your pricing strategy in the wrong direction.
That is the uncomfortable part of primary research. Surveys rarely fail because respondents are dishonest. They fail because respondents are helpful, busy and human and because the questionnaire made the polite answer the easiest one to give. So we’re opening the hood on how questions get designed so the answers can be checked.
What does a dishonest answer actually look like?
Almost never like a lie. In business-to-business research it arrives in three recognisable shapes.
The professional answer. A procurement manager describes the process their job description says they follow competitive tender, three quotes, a scoring matrix rather than the shortcut they took in March because a line was down and one supplier could ship on Thursday.
The aspirational answer. Ask anyone what they will do next year and you get a plan, not a behaviour. Plans are sincere. They are also free, and they get revised the first time a delivery slips.
The reconstructed answer. People rebuild memories into a tidy story. Asked why they changed supplier, many supply the reason that sounds most rational in hindsight, and price is the most respectable reason available to anyone who buys for a living.
Key idea: A survey does not measure what people do. It measures what people are willing and able to tell you. That is a different variable, and it has to be designed for.
How do you design a B2B market survey that gets the truth?
The principle is easy to state and awkward to execute: ask about things that already happened, in enough detail that the answer could in theory be checked against a record.
Five design moves do most of the work.
- Anchor every question to a dated event. Not “how often do you review suppliers” but “when did you last review one, and what prompted it”.
- Ask for the artefact, not the attitude. Shortlists, quote counts, lead times, payment terms. A respondent who cannot recall the artefact has just told you something useful about how much that decision really mattered.
- Make trade-offs cost something. If price, quality, lead time and service are each offered as separate virtues, everyone selects all four. Force an allocation or a ranked substitution so choosing one means giving up another.
- Sample the roles that contradict each other. Our analysts run primary research with manufacturers, distributors and traders, procurement managers, technology providers, regulators, subject-matter consultants and end users — not for coverage, but because each role has a different incentive to shade the same fact. A distributor’s account of lead-time performance is a live check on the manufacturer’s.
- Design the follow-up, not just the form. Often the most useful minute in B2B primary research is the one after the answer, when an analyst asks what happened next.
Why ask about the past when the forecast is about the future? Because past behaviour is the only part of a respondent’s account that has already survived contact with a budget, a deadline and a boss. An opinion about next year has survived nothing yet. Forecasts built on behaviour are forecasts built on something that has already been tested once.
Why does one badly worded question cost you a forecast?
Take a survey of 150 procurement managers in an industrial components category. Asked which factor matters most when selecting a supplier, 78% choose price. That single figure is enough to build a commercial strategy on, and plenty of strategies have been built on less.
Now ask the behavioural version of the same question. Of those 150, 50 had actually changed supplier in the previous 18 months. Asked what triggered the change: 36% lead-time failure, 26% a quality or specification miss, 24% price, 14% service and support.
Price falls from 78% to 24% a 54-point gap and 76% of real switching was caused by something other than price.
The cost of trusting the first number is specific. A manufacturer reading 78% concludes the category is price-elastic, approves a discount programme and gives away margin — then keeps losing the same accounts, because those accounts were never leaving over price. The investment that would have held them, in delivery reliability, never gets funded, because no question in the survey pointed at it.
One survey, two decisions. The stated figure funds a discount. The revealed figure funds a delivery upgrade. Only one of them keeps the customer.
Which questions should you rewrite before you field?
Most questionnaires can be improved in an afternoon, because the same few habits cause most of the damage. Three rewrites are worth making every time.
Asking which supplier was last replaced, and what triggered it, produces a dated event with a cause attached. Asking how many suppliers made the last shortlist, and who was cut first, exposes the real screening criterion rather than the stated one. Asking what was actually paid the last time delivery was needed inside a week replaces a hypothetical willingness-to-pay number with a transaction.
None of these are harder to answer. They are harder to answer casually — which is the entire point.
One test for any question before it goes into field: could the respondent, in principle, check their own answer against a record? If not, you are collecting an opinion and should price it as one.
What stops a weak answer becoming a published number?
Better questions reduce the problem. They do not remove it, which is why no survey finding enters one of our reports on its own authority.
Under The Full Truth, a figure is accepted only when four kinds of evidence agree: primary evidence from people active in the market, paid and credible secondary sources, quantitative analysis and modelling, and expert review by practitioners still working in the category. No single source is trusted alone, and every number must survive cross-examination from at least one independent angle.
That sits inside a nine-stage workflow: primary research, paid and credible sources, data cleaning, market estimation, triangulation, gap analysis, expert validation, internal quality control, and final market insights — with sensitivity analysis and scenario testing running alongside it. Where the angles disagree, the finding goes back to gap analysis. Estimates that do not meet our confidence standards are excluded rather than published with a footnote.
A finding we cannot defend from an independent angle does not get a caveat. It gets left out.
Documentation is the other half of the answer. Reports carry data sources, key-assumption logs, model structure, analytical boundaries and limitations, and confidence intervals where applicable, so a reader can see which conclusions rest on which evidence. The work is carried out under ISO 20252 for market, opinion and social research and ISO 9001:2015 for quality management, and we are members of ESOMAR and the Market Research Society of India.
What this means for manufacturers and buyers
For manufacturers. Your exposure is on the supply side. A stated-preference survey will almost always tell you to compete on price. A behavioural one tells you whether you are losing accounts on lead time, specification tolerance or support response — and those three carry very different capital costs. That distinction decides where capacity, inventory buffer and service headcount should go, and it is the difference between protecting margin and discounting into it.
For buyers, investors and strategists. Your exposure is on the demand side: why customers buy, what actually triggers switching, and where unmet needs are opening up. When you commission or purchase research, the question worth asking a provider is not how large the sample was but how the switching question was worded, and what independent evidence the answer was tested against. A switching figure from a hypothetical question is an opinion poll with a market-size label on it.
Ready to put your questionnaire through this?
A B2B market survey earns its budget when its answers can be checked, not when its response rate is high. Anchor questions to dated events, ask for artefacts instead of attitudes, sample the roles whose accounts contradict each other, and test every finding against independent evidence before it becomes a number anyone plans around.
