Key points
- A sample vendor sells completed interviews that match a specification. Judge them on whether respondents are who they claim, whether the quoted numbers hold in field, and what happens when something goes wrong.
- Write a definition a screener can enforce. Include target completes, a measured interview length, expected incidence, field window, platform and exclusions.
- Ask what incidence the quote is built on and what happens to the price if it comes in lower. The answer separates providers faster than anything else.
- Red flags: incidence quoted above published figures, a price well below the market model, quality described without naming a method, and a fast field estimate on a senior audience.
- Contract for held pricing, free replacement of rejected completes, a rejection log delivered with the data, and source disclosure at category level.
What are you actually buying?
A B2B sample vendor sells you completed interviews with people who match a specification. Everything else, the panel, the platform, the account manager, is how they get there. Judge vendors on three things: whether the respondents are who they say they are, whether the numbers quoted at feasibility hold in field, and what happens when something goes wrong.
New to this? Start with what B2B survey sample is, then come back. Every term below is defined in the glossary.
How does buying B2B sample actually work?
Seven steps from brief to delivered data. Most of the risk sits in steps two and three, before anyone has fielded anything.
- 1
Write the brief
Define the audience in terms a screener can enforce, and state target completes, measured interview length, field window, platform and exclusions.
- 2
Get feasibility
Reachable n, expected incidence, field window and a cost per complete. A good vendor returns this the same business day and tells you what will cause trouble.
- 3
Interrogate the quote
Ask what incidence it assumes, what happens to price if incidence comes in lower, and how job title is verified. Three questions, three very different answers across vendors.
- 4
Contract
Held pricing, free replacement of rejected completes, a rejection log with the data, and source disclosure at category level.
- 5
Soft launch
Field about ten percent. Confirm real incidence and real median length against the quote before releasing the rest.
- 6
Field and monitor
Daily counts by quota cell. Verification should run continuously, not as a clean up pass at the end.
- 7
Check the delivery
Read the rejection log before the data. Run your client's cuts and check the base sizes against the margin of error.
How do you write a brief a vendor can quote?
A vendor can only quote what you specify. Vague briefs get optimistic quotes, then revisions. This is the checklist.
- Audience definition in terms a screener can enforce. “IT decision makers” is not a spec. “Director or above with final or shared authority over IT purchases, at companies with 500 or more employees, US” is.
- Target completes in total and by quota cell. Say which cells are must fill and which are nice to have.
- Interview length as a median from a test, not a guess. Every minute past ten changes price and completion.
- Expected incidence if you have a view, and the basis for it. Compare it against published figures.
- Field window you need, and whether it is fixed by a client deadline.
- Platform the survey is hosted on, and any link parameters you need passed through.
- Exclusions: respondents from a previous wave, competitors, agencies, anyone already in a related study.
What should you ask at feasibility?
Seven questions. A vendor who answers all seven clearly is one you can work with. A vendor who cannot answer the second is going to revise the price in week two.
| Ask this | What a good answer sounds like |
|---|---|
| What incidence is this quote built on, and what is it based on? | A specific figure with a source, and a willingness to compare it to published numbers |
| What happens to the CPI if incidence comes in lower? | The price holds, or a clearly stated tolerance band agreed before field |
| How do you verify job title and company size? | A named method: corporate domain match, third party record, category knowledge question |
| What is your rejection policy? | Removed and replaced at no cost, inside an agreed window |
| Will I get a rejection log with the data? | Yes, with reasons per removal |
| Is this one panel or several, and can I see the source mix? | Category level disclosure at minimum |
| What happens if the first 20% comes in slow? | A named escalation, not silence until the deadline |
What are the red flags in a quote?
Two quotes for the same specification can differ by half and describe entirely different products. This is how to tell which one you are holding.
| Red flag | What it usually means | What good looks like |
|---|---|---|
| Incidence quoted well above published figures | The screener will qualify people who should have screened out | An incidence at or below the published range, with the reasoning |
| A CPI far below the market model | No verification, or a price that will be revised in field | A price inside the market range, and a commitment that it holds |
| “Industry standard quality checks” | Identity checks only, which do not catch claim fraud | A named method for verifying the employment claim |
| Nothing said about price if incidence drops | You will find out in week two | Held pricing, in writing, tied to the specification |
| A very fast field estimate on a senior audience | The vendor has not checked supply | A realistic window, and a smaller reachable n if that is the truth |
| Zero rejections reported on past studies | They are not checking, or not telling you | A rejection rate they can quote from memory, usually 8 to 15% |
Which contract clauses actually protect you?
| Clause | What it protects |
|---|---|
| Held pricing | The CPI quoted at feasibility holds through delivery unless you change the spec |
| Free replacement | Any complete you reject on documented quality grounds is refielded at no cost |
| Rejection log | Every removal and replacement listed with a reason, delivered with the data |
| Source disclosure | The vendor names the supply sources used, at least at category level |
| Respondent data handling | GDPR and CCPA consistent, with no personal data delivered unless the study requires it and the respondent consented |
| Soft launch gate | Full field only released after the first ten percent confirms incidence and length |
What should you check after field?
Read the rejection log before you read the data. It tells you how hard the vendor worked and what the raw supply looked like. Then check the open ends for category knowledge, run the cuts your client will run, and check the base sizes against the margin of error before you present a difference as a finding.
Buying B2B sample: frequently asked questions
How far in advance should I brief a sample vendor?
For a standard B2B audience, a week before field is comfortable. For C-suite, healthcare or narrow screens, two to three weeks lets the vendor confirm supply rather than guess. Same business day feasibility is normal for a good vendor on standard audiences, but confirmed supply on rare audiences takes longer.
Should I use one sample vendor or several?
One vendor with multiple sources behind them gives you a single standard of verification and one point of accountability. Splitting a study across vendors yourself means reconciling different quality standards and duplicate respondents across sources.
What is a fair rejection rate?
For verified B2B sample, expect 8 to 15% of raw completes to be removed and replaced before delivery, higher on executives. A vendor reporting zero rejections is either not checking or not telling you.
Can I get respondent level verification data?
You can get the rejection log and a description of the verification method for every delivered complete. Respondent identity data is not delivered because it is personal information, and you do not want the liability of holding it.
How do I know if a quote is too cheap?
Run the specification through the cost per complete estimator. If a quote sits well below the market range for that incidence and seniority, ask what verification is included and what happens to the price if incidence comes in lower. One of those two answers usually explains the gap.
What should I do if incidence comes in lower than quoted?
Stop full field at soft launch and renegotiate before spending the budget. A vendor with held pricing absorbs it. A vendor without it will ask for more money, and the time to find that out is at ten percent fielded, not eighty.
Do I need a soft launch on every project?
On anything above about 100 completes, yes. It costs a day and it is the only cheap moment to discover that the incidence assumption or the interview length was wrong.
Send the spec. Get real numbers back.
Audience, market, target n, expected interview length. Feasibility the same business day for standard audiences, with reachable counts, an estimated qualification rate we believe and a quoted cost per complete that stays fixed for the agreed brief.