Vetting a CI Vendor: 12 Questions That Matter
Buying CI is a procurement decision with asymmetric information: vendors overpromise, and buyers can't verify until it's too late. Twelve questions across four dimensions — plus three red flags and a pilot structure — turn the decision from a leap of faith into due diligence.
Why Vendor Vetting Fails
Asymmetric information: the vendor knows exactly how deep the methodology goes. The buyer finds out later.
🎭 The Pitch Is a Deliverable in Disguise
The pitch deck is produced by the sales team, not the analysts who will serve you. It's engineered to close, not to inform. Every CI buyer has signed after a confident demo and then received work that doesn't match the sample. The failure is structural: you're evaluating marketing, not methodology.
🧩 You Can't Verify What You Can't Define
Buyers rarely write down what “good intelligence” means before evaluating vendors. Without a definition — signal types, cadence, confidence standards, format — every vendor looks plausible. The vetting process is where you discover your own requirements, but only if you arrive with questions, not just a budget.
📉 Procurement Logic Doesn't Fit Research
RFPs and price comparisons reward the cheapest confident answer. But CI is a research operation: value lives in sourcing discipline, verification, and analyst judgment — invisible in a quote. The same procurement logic that works for hardware actively selects against good intelligence.
The 12 Questions, in 4 Dimensions
Methodology, manufacturing domain, delivery, contract. Ask all twelve; discount any answer that's a slogan.
🔬 Dimension 1 — Methodology (Q1–Q3)
Q1: Walk me through your sourcing process for a competitor — which channels, in what order? Q2: How do you verify a claim before it reaches my inbox, and what's your confidence standard? Q3: When sources conflict, what's your tie-break rule? A vendor who can't articulate methodology on the spot doesn't have one.
🏭 Dimension 2 — Manufacturing Domain (Q4–Q6)
Q4: Which signals matter in my industry — patents, certifications, tenders, tariffs? Q5: Build me a Key Intelligence Topic (KIT) for my segment right now — what would you monitor first? Q6: Name the manufacturing channels you actually watch: CNIPA, customs, standards bodies, exhibitions? Domain fluency is answerable in seconds — or it isn't real.
📦 Dimension 3 — Delivery and Contract (Q7–Q12)
Q7: What does a weekly deliverable look like — format, length, signal types? Q8: Who is the analyst on my account, and how do they escalate to senior researchers? Q9: When my market shifts, how fast can coverage adapt? Q10: What's the pilot structure — scope, duration, exit without penalty? Q11: Who owns the data and deliverables we co-produce? Q12: What are the termination and IP terms in plain language?
Pilot First, Then Commit
The only verifiable proof of a CI vendor is a pilot on your actual market, with your actual KITs.
🧪 Design a Pilot That Tests the Real Thing
Scope: two to four weeks, two competitors, one KIT each, weekly deliverables, fixed review points. Pay for it — free pilots test sales, paid pilots test delivery. Require sourcing notes alongside every claim: you're buying the audit trail, not the bullet points. A vendor that shines in a paid pilot usually shines in the engagement.
🚩 Three Red Flags That End the Conversation
Red flag one: no methodology, only “our platform” — if the answer to “how do you source?” is a product demo, the product is the pitch. Red flag two: generic deliverables — if the sample could describe any industry, the research is template-driven. Red flag three: no pilot, or a pilot buried inside a long contract. Vendors confident in delivery offer pilots; vendors confident in sales resist them.
✅ The Post-Pilot Decision, With Evidence
After the pilot: did the analyst ask better questions than your team did? Did verification hold when you probed the sourcing notes? Could you reproduce the sourcing path for the key claims? Yes on all three — negotiate the full engagement from the pilot's evidence. No on any — the pilot just saved you a year of bad intelligence. That's the cheapest research you'll ever buy.
Don't buy a pitch. Buy a process.
Ask us the twelve questions — we'll answer them in writing.
Get in touch