Use a competitor scorecard in two layers: first decide whether the candidate is eligible for the role you want to assign; only then use a weighted score to prioritize attention inside that eligible role.
A scorecard is useful because it forces a team to state what “relevant competitor” means. It becomes misleading when every dimension is added into one total and a fatal mismatch is allowed to disappear inside a strong average. A marketplace can influence the buyer, rank highly in search and matter strategically while still not fulfil the target job itself. That makes it relevant, but not a like-for-like direct competitor.
Eligibility is non-compensatory. Ranking is compensatory only after eligibility. A strong score on market visibility or decision-stage influence cannot cancel a failed buyer, offer, market or evidence gate.
Status: Founder reviewed by Pero Baldjiev. Founder reviewed by Pero Baldjiev.
Fast Action: test whether a high score is actually eligible
The browser scorecard below deliberately shows two outputs: a raw weighted profile and the governed role/action. The raw number is visible for teaching; it becomes a ranking score only when the hard direct-competitor gates pass.
Why a competitor scorecard should not start with the total
Most weighted scorecards are compensatory: strong performance on one criterion can offset weak performance on another. That is useful when trade-offs are legitimate. It is dangerous when one dimension is a true eligibility condition. UK Government MCDA guidance explicitly notes that overall numerical ratings permit compensation and that absolute minimum criteria can be used to eliminate options before analysis. MCDA literature also describes veto thresholds that prevent a severe failure on one criterion from being cancelled by strength elsewhere.
| Question | Compensation allowed? | A67 treatment |
|---|---|---|
| Does the same buyer plausibly choose it? | No for DIRECT status | Hard gate |
| Does the entity itself fulfil the same underlying job? | No for DIRECT status | Hard gate |
| Does the target market align? | No for DIRECT status | Hard gate |
| Is evidence strong enough to classify? | No | VERIFY if below threshold |
| How relevant is the decision stage? | Yes after eligibility | Weighted input |
| How strategically material is the candidate? | Yes after eligibility | Weighted input |
The six criteria in the governed relevance policy
A67 uses six dimensions because they answer different questions. The example weights are not universal truths; they are a fictional, versioned teaching policy. In a real decision, the accountable reviewer should define and freeze the weights and thresholds before looking at which candidate benefits.
| Criterion | Example weight | Direct minimum | Purpose |
|---|---|---|---|
| Buyer overlap | 25% | 3/5 | Could the same buyer plausibly choose this candidate? |
| Offer overlap | 30% | 3/5 | Does the entity itself fulfil the same underlying job? |
| Market fit | 15% | 3/5 | Does geography, language and regulation align? |
| Decision-stage relevance | 10% | None | Does it influence the same evaluation moment? |
| Evidence confidence | 10% | 3/5 | Can the role claim be defended from current evidence? |
| Strategic materiality | 10% | None | Would learning from this candidate change action? |
The policy itself is versioned in the workbook. If a reviewer changes a weight or direct threshold without issuing a governed new policy version, downstream candidate roles block rather than silently recalculating under an undocumented standard.
How the scorecard classifies competitor relevance
- Freeze the decision. Define the buyer, job/offer, market, decision stage and excluded decisions.
- Check the entity type and own fulfilment. A marketplace, publisher or community should not be declared as directly fulfilling the target job without resolving the contradiction.
- Attach candidate-owned qualification evidence. The evidence record must belong to the same Candidate ID and Decision ID and use the governed QUALIFICATION role.
- Resolve unknowns before scoring. Unknown inputs and weak evidence produce VERIFY, not a neutral midpoint.
- Apply the direct gates. Entity Type must be
PROVIDER, Own fulfilment must beYES, buyer ≥ 3, offer ≥ 3, market ≥ 3 and evidence confidence ≥ 3. ASTATUS QUO / INTERNALsubstitute cannot become DIRECT merely by scoring highly. - Assign the role. DIRECT only when the provider-specific direct gates pass; CONTEXTUAL when the entity materially influences or substitutes for the decision without like-for-like provider eligibility; EXCLUDED when relevance is too weak; VERIFY when evidence is insufficient.
- Calculate the weighted profile. Keep it visible for transparency, but promote it to a governed ranking score only for DIRECT candidates.
- Freeze the set. Store Decision version, Policy version and every candidate freeze key. New, removed, changed or re-versioned records invalidate the old handoff.
A high-looking score can still belong outside the direct set
The fictional comparison below is designed to make the false-positive problem obvious. It is synthetic training data, not evidence about real companies.
| Candidate | Raw weighted profile | Critical gate | Governed role | Permitted action |
|---|---|---|---|---|
| CoreRoster | 97 | All direct gates pass | DIRECT | INCLUDE DIRECT |
| CompareHub | 76 | Marketplace; own fulfilment NO; offer overlap 1/5 | CONTEXTUAL | INCLUDE CONTEXTUAL |
| AdjacentOps | 68 | Offer overlap 2/5 | CONTEXTUAL | INCLUDE CONTEXTUAL |
| ThinSignal | 76 | Evidence confidence 2/5 | VERIFY | VERIFY BEFORE CLASSIFYING |
| OffMarketCo | 84 | Market fit 1/5 | CONTEXTUAL | INCLUDE CONTEXTUAL |
| UnknownSource | Not calculated | Multiple unresolved inputs | VERIFY | VERIFY BEFORE CLASSIFYING |
CompareHub at 76 is not ranked against CoreRoster at 97. The marketplace is relevant to discovery and evaluation, but it does not fulfil the target job. Role difference outranks the attractive total.
Operationalize the result as separate competitor sets
The current competitive-analysis market commonly distinguishes direct and indirect competitors because the categories answer different questions. Competitive Intelligence Alliance also describes weighted competitor scorecards as a way to prioritize CI attention. A67 combines those ideas but keeps the roles separate: scoring helps order the direct set; it does not collapse providers, marketplaces, publishers and contextual references into one league table.
Use the result as a role-specific research portfolio
Eligible for like-for-like comparison. A governed ranking score can prioritize research attention inside this set.
Influences demand, discovery, evaluation or adjacent alternatives. Analyze separately; do not rank against DIRECT providers.
Not materially relevant to this frozen decision. Preserve the reason so exclusion can be reopened later.
Evidence or controlled inputs are insufficient. Do not convert unknown into an average score.
Internal/manual substitutes may be highly relevant and may strongly fulfil the job, but under this model they remain contextual alternatives rather than DIRECT providers.
Evidence, policy and candidate versions must travel together
The workbook treats competitor relevance as a governed classification, not a disposable worksheet. Qualification evidence is owned by the same Candidate ID. The relevance policy has its own version. Candidate records have As-of dates and versioned content snapshots. Each candidate freeze identity carries the bound Qualification Evidence ID and the Evidence record's own freeze key, so a legitimate Evidence V2 invalidates an older frozen handoff even when the candidate's role and rank do not change.
- Change candidate wording without updating its version → VERSION UPDATE REQUIRED.
- Change qualification evidence without updating its version → candidate binding fails closed.
- Change a weight or threshold without a new policy version → all candidate classification blocks.
- Issue a valid Policy V2 after a frozen snapshot → the old handoff becomes DRIFTED.
- Add a new governed candidate after freeze → NEW RECORD - DRIFTED.
When to reopen competitor relevance
- Sales or win/loss evidence changes buyer substitutability.
- The candidate launches or removes the target offer.
- Market availability, geography or regulation changes.
- A marketplace or publisher begins direct fulfilment, or a provider becomes only an intermediary.
- Evidence confidence changes materially.
- The decision stage changes.
- The team intentionally changes relevance weights or thresholds under a new policy version.
How to adapt the scorecard without copying the fictional weights
Do not copy the example weights simply because they add to 100. Start from the real decision. Decide which criteria are true eligibility conditions and which are trade-offs. If a criterion cannot legitimately be compensated by another, make it a gate or minimum threshold rather than giving it a larger weight and hoping the arithmetic behaves.
| Carry forward | Replace with your own decision evidence |
|---|---|
| Role Before Rank | Your definition of DIRECT, CONTEXTUAL and EXCLUDED |
| Candidate-owned qualification evidence | Your attributable sources and as-of dates |
| Hard eligibility gates | Your buyer/offer/market/evidence minimums |
| Versioned relevance policy | Your approved weights and threshold rationale |
| Separate direct ranking | Your research-attention priorities inside the eligible direct set |
Download the Competitor Relevance Scorecard
The Excel workbook is the deep-work surface: Decision Contract, versioned relevance policy, candidate-owned qualification evidence, gate-first scorecard, governed portfolio and drift-sensitive handoff. The PDF is the compact field guide for review meetings.
Questions about competitor relevance scorecards
Should I rank every discovered competitor by one score?
No. First decide whether each entity belongs in the same comparison role. Direct providers, marketplaces, publishers, substitutes and contextual references should not automatically share one league table.
Can a high total still be a false positive?
Yes. If buyer, offer, market or evidence is a hard eligibility condition, strength elsewhere cannot compensate for that failure. The raw profile can stay visible, but it does not grant DIRECT status.
What should happen when evidence is weak?
Use VERIFY BEFORE CLASSIFYING. Do not substitute a midpoint, zero or guessed value for an unknown evidence state.
How should weights be chosen?
Weights should reflect the frozen decision and be agreed before inspecting which candidate benefits. Criteria that are true vetoes should be gates or minimum thresholds, not merely high weights.
When is a contextual competitor useful?
When the entity materially influences discovery, evaluation, budget, substitutes or adjacent demand without being eligible for like-for-like direct comparison.
Does READY TO SHARE mean the set is approved?
No. It means the frozen fictional packet is internally consistent. A real competitor classification still belongs to the accountable reviewer and its evidence.
Where A67 fits in the Competitive Decision Toolkit
Sources, comparison context and method limits
- UK Government Analysis Function — Introductory Guide to MCDA: compensatory scoring and minimum-criteria context.
- Cinelli et al. — MCDA taxonomy review: thresholds and veto/non-compensatory modelling context.
- Competitive Intelligence Alliance — competitor scorecards: current weighted-scorecard comparison context.
- Shopify — Competitive Analysis Guide 2026: direct/indirect competitor categorization context.
- Outmano — Direct and Indirect Competitors 2026: same-buyer/similar-product role context.
- Google Search Central — people-first content guidance: publication/search context only.
These sources provide transferable context. They do not validate the proprietary SERPIFAI thresholds, fictional candidate roles or any unstated client outcome.