Updated 15 November 2026 · about 9 minutes
A security questionnaire is a buyer's way of deciding whether you're safe to trust with their data. It's rarely a pass/fail test. Reviewers want clear, consistent, verifiable answers — and they want to see you understand your own risks. Here's a method that works for small vendors without a security team.
1. Before you open the spreadsheet: triage
Spend ten minutes on these questions first. They decide how much effort the questionnaire deserves.
- When is it due, and who asked? Ask your champion at the customer for the reviewer's name and the real deadline. Deadlines often have slack; reviewers rarely do.
- What's the deal worth and what data will you hold? A vendor that stores customer HR records gets a harder review than one that sends anonymous metrics. Calibrate.
- Will they accept existing evidence instead? Many reviewers will take a SOC 2 report, a completed CAIQ or a link to your trust page in place of some or all of the questions. Always ask.
- What format do they need back? Usually their own Excel file. Don't paste answers into a new document — reviewers compare against their template, and a reformatted file slows them down.
2. Start from your library, not from a blank cell
The biggest time sink is rewriting answers you've already given. Keep one approved answer per question, with an owner and a review date (we explain how in Build an RFP answer library). Then the work for a new questionnaire becomes:
- Map the customer's columns — which column holds the question, which takes the answer.
- Match each question to your library. Expect 60–80% of a typical questionnaire to match well once your library covers a couple of past questionnaires.
- Read every matched answer once in context. The same approved answer can need a small tweak when the question is narrower or broader.
- Spend your real effort on the remaining questions.
Wording varies more than meaning: “Is data encrypted at rest?”, “How is stored information protected?” and “Describe your encryption of customer data” are the same question. Store alternate phrasings with each answer so they all find it.
3. Write answers a reviewer can verify
A good security answer has three parts: a direct response, the specific how, and where the evidence is.
Weak: “We take security very seriously and follow industry best practices.”
Strong: “Yes. MFA is required for all access to production systems, cloud consoles and source code; administrators use hardware security keys. Our access-control policy is available under NDA.”
- Lead with yes, no or partly. Many questionnaires have a Yes/No column next to a comments column. Fill both.
- Use numbers and names. “Reviewed quarterly”, “TLS 1.2 or higher”, “restored from backup in a test twice a year”.
- Say who does it. If your cloud provider handles physical security, say so and point to their report. Claiming inherited controls as your own is a common way to lose trust.
- Keep it short. Two to four sentences. Long essays hide the answer and invite follow-up questions.
- Be consistent. The same question answered differently in two places is the fastest way to trigger a deeper review. A single source of approved answers fixes this.
4. Handle gaps honestly
Every small company has gaps. Reviewers expect that; what they're checking is whether you know about yours and have a plan. When the true answer is no:
- Say “No” plainly, then what you do instead (a compensating control) and, if true, when you'll close the gap.
- Example: “No, we don't yet have a SOC 2 report. Our controls follow the same criteria; we share our policies, latest penetration test summary and architecture overview under NDA, and plan to begin a SOC 2 Type I in Q2.”
- Never answer yes to something you don't do. Questionnaire answers often end up attached to the contract.
If nothing in your library covers a question, don't let a tool — or a colleague — improvise. Mark it as a gap and assign it to the person who actually knows. In RFPKit, AI drafts are only written from approved answers for exactly this reason: when there's no approved source, there's no draft.
5. Pull in experts without a meeting
Route each open question to one named person — engineering for infrastructure, legal for contract terms, operations for continuity — with the question, the deadline and what you've drafted so far. Ask them to edit and approve, not to write from scratch. Then add their approved answer to the library so you never have to ask again.
6. Review before you send
Have one person read the whole questionnaire top to bottom before it goes out. They're checking for contradictions, answers that over-promise, internal names or URLs that shouldn't be shared, and cells left blank by mistake. Keep a record of which approved answer and version you used for each response, so if a reviewer asks “has this changed since?”, you know.
7. Return it in their format
Write your answers into the customer's original file. Keep their sheets, formatting, drop-downs and column widths untouched; a reviewer who has to clean up your export is a reviewer in a worse mood. If a cell already has text (sometimes customers pre-fill guidance), decide deliberately whether to overwrite it.
8. After you send it
- Put follow-up questions and your answers back into the library.
- Note any promises you made (“we'll complete X by June”) and make sure someone owns them.
- Set review dates on answers that change often — subprocessors, certifications, pen test dates — so they don't go stale.
- Consider a trust page so the next prospect can self-serve your overview and documents before sending a questionnaire at all.
A checklist to keep
- ☐ Asked whether existing evidence (SOC 2, CAIQ, trust page) can replace some questions
- ☐ Matched every question against the approved library
- ☐ Read each matched answer in context
- ☐ Every “yes” is true today and specific
- ☐ Gaps answered honestly with compensating controls
- ☐ Expert-written answers approved and added to the library
- ☐ One full read-through for contradictions and over-promises
- ☐ Exported into the customer's original file