B2B teams copy consumer-survey sample rules, overspend on hard-to-reach respondents, and still recruit the wrong buying roles.
For many directional B2B decisions, 100 qualified respondents can be enough. The required sample depends on the segments you need to read, the precision you need, how rare the audience is, and what happens if the decision is wrong.
A 400-response study might cost about $40,000 while a 100-response study might cost about $10,000. Treat these as planning examples, not market benchmarks. Replace them with an actual quote for the ICP, incidence rate, interview length, and fieldwork method.
Consumer studies may need broad samples to represent wide differences in age, income, location, and behavior. A B2B study often targets a narrow set of roles inside a defined company profile.
That does not make every group homogeneous. The right question is whether the sample represents the buying roles and segments needed for the decision.
One hundred qualified people can be more useful than 1,000 respondents who do not match the ICP or participate in the decision.
Use a screener for role, seniority, company size, industry, geography, and buying involvement. Set quotas before fieldwork so an easy-to-recruit role does not dominate the sample.
Niche senior respondents usually cost more and take longer to recruit. The actual price and timing depend on incidence, geography, quotas, survey length, and recruitment source.
At a 95% confidence level, with simple random sampling, a large population, and a response split near 50/50, the approximate margins of error are 9.8 percentage points for 100 responses, 6.9 points for 200, and 4.9 points for 400.
Those calculations describe sampling error under stated assumptions. They do not correct a weak panel, nonresponse, weighting, fraud, low-incidence audiences, or unstable subgroup cuts. A precise answer from the wrong respondents is still wrong.
If a result is 70% at a base of 100, the approximate interval under those assumptions is about 60% to 80%. Decide before fieldwork whether more precision would change the business decision.
Larger samples can consume budget that might be more valuable in better screening, repeated measurement, qualitative follow-up, or a second ICP.
Treat $10,000, $20,000, and $40,000 sample-cost examples as illustrations, not universal prices. Compare the extra precision with the decision value before approving the spend.
Set the requirement before fieldwork:
Each reported segment needs an adequate base. Do not collect 100 responses, divide them into three industries after fieldwork, and report groups of 20 or 30 as complete studies.
Use larger samples when you need reliable segment comparisons, need to detect small differences, or plan to make consequential public claims. Academic, clinical, and regulatory work may require different designs and standards.
A focused 100-person study can often field faster than a 400-person study, but fixed timing promises are misleading. Incidence, geography, quotas, role seniority, survey length, and recruitment source determine the schedule.
Speed has value when the decision is waiting, but do not relax respondent-quality checks to finish sooner.
Generalize only to the frame sampled. One hundred US SaaS marketing leaders cannot represent every B2B buyer, country, or buying role.
Ask four questions before someone insists on a larger sample: Which decision changes? Which segment needs a separate read? What difference must the study detect? How much uncertainty is acceptable?
If 100 qualified respondents answer the directional question and the result will not be split into unstable subgroups, stop there. If the stakes, segmentation, or precision requirement is higher, increase the sample or change the research design.
Respondent fit matters more than a round-number rule. Design the sample around the decision.
.jpg)
.png)
