Most “bad” supplier selections aren’t caused by missing data. They happen when the team sees a slick demo or a low price and then reverse-engineers the criteria to justify the gut feel. If you want a process you can defend to Finance, Legal, and an unhappy stakeholder six months later, the discipline is simple: define the deal-breakers and the weighted scorecard before any supplier responds.
Step 1: Align the business requirement (and write it down like you mean it)
Start by forcing clarity on what “good” looks like in the business, not in procurement. The output of this step isn’t a wish list—it’s a requirements brief that can survive handoffs and staff changes. If stakeholders can’t agree on outcomes, they’ll fight about vendors later.
Output: 1–2 page requirements brief (problem statement, scope, in/out, target go-live date, success measures).
Include constraints that are real: integration environment, data residency, union rules, site access windows, budget ceiling if it exists.
Write 3–5 measurable success measures (e.g., “reduce invoice exceptions by X” is fine; “modernize AP” is not).
Name the decision owner and the evaluation team (who scores what).
Step 2: Set pass/fail gates and deal-breakers (before you talk to suppliers)
Pass/fail gates prevent the most common time-waster: scoring a supplier that can’t legally, technically, or operationally do the job. Deal-breakers also protect you from “but they’re great in every other way” arguments when something is non-negotiable.
Output: a gate checklist with evidence required (yes/no, with proof).
Typical gates: mandatory certifications, ability to sign your core contract terms (or defined redlines), financial viability checks, minimum service coverage, required integrations, data residency requirements, background checks for on-site work.
Define what counts as evidence (e.g., SOC report, insurance certificate, sample contract markup, architecture diagram, customer references in your region).
Decide now what happens if a vendor fails a gate: immediate disqualification or “cure period” with a deadline.
Step 3: Build the weighted scorecard (still before any vendor responds)
If you wait until proposals arrive, the scorecard becomes a storytelling tool, not a decision tool. Agree the criteria, weights, and scoring definitions while the team is still thinking about needs rather than personalities.
Output: a scorecard with criteria, weights totaling 100%, scoring scale, and anchor definitions for each score.
Use fewer criteria than you think. Ten well-defined criteria beat thirty vague ones.
Separate “can they do it?” from “how well?” Gates handle the first. The scorecard handles the second.
Define anchors: what does a 1, 3, and 5 look like for each criterion? Without anchors, every score is just a mood.
Assign scoring owners by expertise (IT scores integration, Finance scores pricing model risk, Ops scores service delivery).
A practical example: if “Implementation approach” is 15%, define a 5 as “named delivery team, credible plan with milestones, risk register, and customer examples on similar scope,” and a 1 as “generic methodology slides with no resourcing or timeline.”
Step 4: Longlist the market (fast) and create a three-to-four supplier shortlist (deliberate)
Longlisting is for coverage: you’re confirming who exists and who is plausibly relevant. Shortlisting is where you protect your team’s time. Three or four suppliers is usually the sweet spot—competitive tension without turning evaluation into a second job.
Output: longlist (8–15) with a one-line rationale each; shortlist (3–4) approved by the decision owner.
Use a lightweight RFI or structured discovery call script tied to your gates and top criteria.
Apply gates early (e.g., “must support EU data residency” or “must integrate with our ERP version”).
Document why each supplier is included or excluded. This is what you’ll want when someone asks, “Why didn’t we invite Vendor X?”
Step 5: Run the RFP/RFQ and keep responses comparable
Comparability beats volume. Suppliers will happily give you 80-page proposals that answer questions you didn’t ask. Your job is to force consistent inputs so scoring is about evidence, not who wrote the best marketing copy.
Output: an RFP pack with instructions, response templates, pricing sheets, and the evaluation model (at least the criteria categories and process).
Require pricing in a single template: unit rates, assumptions, volume bands, one-time vs recurring, indexation rules, implementation fees, and optional items clearly separated.
Ask for proof artifacts where it matters: sample reports, SLA definitions, escalation paths, implementation plan, security responses, and reference contacts.
Set rules: Q&A window, single point of contact, no side-channel selling to stakeholders, and a deadline with consequences.
Step 6: Evaluate against the scorecard (and stop pretending subjective scoring is objective)
This is where teams cut corners: they score loosely, discuss endlessly, and then “average it out,” which hides disagreement rather than resolving it. If you want defensibility, capture evidence, score independently first, then calibrate.
Output: completed scorecards, an evidence log per supplier, and a calibration record showing final agreed scores and why.
Run evaluation in two passes: (1) individual scoring based on written responses and demos; (2) calibration meeting to align on evidence and resolve gaps.
Tie every non-price score to a citation: proposal section, demo script item, reference feedback, or artifact.
Use a scripted demo tied to scenarios you care about (e.g., “raise a PO exception and resolve it end-to-end”), not a supplier’s standard deck.
Keep gates alive: if a supplier fails a mandatory security requirement during clarification, they don’t get “points off”—they fail.
The common shortcut to avoid: changing weights after seeing bids (“Price matters more than we thought”) or scoring based on confidence (“they seem solid”) instead of evidence. If you genuinely need to change weights, log the reason, get decision-owner approval, and re-score everyone using the same revised model. Anything else is theatre.
Step 7: Award with a clear rationale (and lock down the commercial reality)
Award is not the same as picking the top score. Sometimes the top-scoring supplier carries contract risk, implementation capacity issues, or a pricing structure that will punish you at scale. The point is to make the trade-off explicit and recorded.
Output: award recommendation memo (winner, runner-up, rationale, key risks, mitigations, total cost view, approvals) and a negotiation plan.
Validate total cost, not just the first-year number: ramp fees, change request rates, minimum volumes, travel policies, indexation, and termination charges.
Confirm delivery capacity and named resources for the start window (especially for professional services-heavy awards).
Document concessions and open items; convert “we’ll figure it out later” into contract language, SLAs, or a signed statement of work.
Notify unsuccessful bidders with a brief, factual debrief tied to criteria—no editorializing.
A practical control: your “defensibility file”
If you ever face an audit, a protest, or an internal post-mortem, you’ll be glad you kept a clean record. It also stops the team from rewriting history after the fact.
Requirements brief and stakeholder sign-off
Pass/fail gates with evidence received
Pre-bid scorecard with weights and scoring anchors (dated)
Longlist/shortlist rationale
RFP pack, Q&A log, and clarifications
Individual scorecards, calibration notes, and final scoring summary
Reference check notes and risk register
Award memo, approvals, and negotiated contract/SOW