Treatment before number
Lead with the qualitative competitor treatment, the strongest reason, the evidence ceiling and the next verification task. The user should not have to interpret a 0–100 number before understanding what the result means.
Competitor Fit Challenge · Methodology Beta
Competitor Fit methodology: how scoring, evidence and gates work.
Competitor Fit Challenge calculates Relationship Fit from six weighted lenses totaling 100. The score cannot overrule Answer Coverage, critical gates or Evidence Readiness. Unknown stays unresolved; when missing evidence could change decision use, the correct next state can be MORE EVIDENCE REQUIRED.
Methodology 1.16 weighted lenses67% + critical gates for Strong fit
How Relationship Fit is calculated
Relationship Fit = Σ(weight × controlled answer value) across six lenses totaling 100 points. The current CFC runtime uses Framework Fit as its tool-facing label for this same deterministic six-lens score; it is not probability, confidence or a separate calculation. Stronger classification still requires enough answer coverage and the critical competitive relationships to be supported.
Buyer overlap22 points
Offer overlap20 points
Commercial-demand overlap16 points
Geographic relevance14 points
Customer-comparison likelihood14 points
Decision relevance14 points
Gate principle: high points elsewhere do not erase missing competitive fundamentals. See the six-lens scoring guide for the lens-by-lens interpretation.
Two reading modes, one protected method
Presentation can become simpler without changing the six-lens logic underneath it. Guided mode protects first-time comprehension; Analyst mode exposes the numerical and evidence detail needed for review.
Inspect the mechanics
Expose the weighted score, lens-level answers, answer coverage, critical gates, source records and sensitivity details. The extra detail explains the result; it does not grant additional authority.
Method-stability rule: changing navigation, demoting the exact score or improving evidence presentation does not change the method when the protected dimensions, weights, gates and classification logic remain unchanged.
Controlled answer values
These values are deterministic inputs to provisional Relationship Fit; they are not confidence percentages, probabilities or Evidence Readiness scores.
| Answer | Calculation value |
|---|---|
| Strong evidence | 1.00 |
| Some evidence | 0.67 |
| Little evidence | 0.33 |
| No evidence | 0.00 |
| Unknown | 0.00 toward provisional fit; retained as unresolved in the explanation |
Evidence states stay explicit
A score input and an evidence state answer different questions. Every consequential claim should remain attributable as participant-supplied, observed, inferred or unresolved so the result does not manufacture certainty.
| Evidence state | What it means | Decision boundary |
|---|---|---|
| Participant-supplied | A user or stakeholder supplied the claim or answer. | Useful input, but not independent verification. |
| Observed | A bounded fact was directly found in attributable public or first-party evidence. | Supports only the proposition actually observed. |
| Inferred | A bounded interpretation was derived from supplied or observed evidence. | Must remain labelled as inference; it cannot be promoted to observation. |
| Unresolved | Evidence is missing, contradictory or insufficient for the relationship being tested. | May constrain classification and can require MORE EVIDENCE REQUIRED. |
Four outputs must remain separate
The method intentionally refuses to compress everything into one confidence number.
Relationship Fit
The provisional weighted arithmetic result from the six answered lenses.
Answer coverage
The share of the required evidence that has been answered rather than left Unknown.
Critical-gate status
Whether buyer overlap, realistic substitution and customer-comparison/decision relevance support stronger classification.
Next-best evidence
The specific missing or contradictory evidence most capable of changing permitted decision use.
From weighted evidence to safe decision use
Relationship Fit answers how closely the candidate matches the competitive relationship. Evidence Readiness answers how far that conclusion can safely travel. Critical gates and unresolved evidence can limit use even when the weighted score is high.
Evidence Readiness is separate from Relationship Fit
Fit tells you the relationship. Evidence Readiness tells you how far you can act. A candidate can look commercially plausible while the evidence needed for a consequential decision is still incomplete. Readiness therefore never multiplies or “adjusts” the fit score. See the Evidence Readiness guide for the decision-use boundary.
Up to five useful pages can cover offer, buyer, market, comparison and context roles. This reduces shallow “homepage-only” research but remains public-web evidence.
Customer comparisons, sales calls, CRM, win/loss and account evidence can resolve questions that a competitor website cannot answer about actual buyer behavior.
For consequential use, an accountable reviewer may need to validate the evidence, scope and permitted conclusion independently.
Page-role coverage improves screening without manufacturing certainty
The discovery layer seeks complementary evidence roles. Repeated pages that tell the same story do not automatically improve readiness.
| Page role | What it can support | What it cannot prove alone |
|---|---|---|
| Offer | Displayed product/service scope, package, fulfilment | Actual buyer substitution or win rate |
| Buyer | Stated target segments, use cases, problem framing | Who actually buys or compares in live deals |
| Market | Published geography, service area, jurisdiction | Actual availability in every observed case |
| Comparison | Named alternatives, replacement language, differentiators | Independent buyer preference |
| Context | Documentation, cases and role clarification | Commercial rivalry by itself |
Automated evidence acquisition stays separate from the scoring contract
The free scan can collect bounded first-party public observations and, when an exact entity match is available, separate structured external context. Those sources can strengthen evidence integrity without changing the six dimensions, weights, answer values or critical gates.
PUBLIC COMPANY EVIDENCE
Claims remain attributable to the company
Product, buyer, market, pricing, comparison and other public-site observations are preserved with source URLs, retrieval time and source role. Repetition across a company’s own pages can deepen the evidence but does not make it independent.
INDEPENDENT STRUCTURED CONTEXT
External corroboration is proposition-specific
An independent structured source may corroborate a bounded entity or industry/category proposition after exact domain identity matching. It cannot by itself prove buyer substitution, demand overlap, serviceability, shortlist behaviour or decision impact. If the external adapter is unavailable or has no exact match, the field remains unresolved.
Strong fit has non-compensating conditions
A strong numerical fit is not sufficient. The current public method requires at least 67% answer coverage plus supported buyer overlap, realistic substitution and customer-comparison likelihood or decision relevance.
WHY
Some evidence is structurally important
A candidate cannot become a strong business competitor merely by scoring well on geography and topic similarity if the evidence does not support a shared buyer or realistic substitution.
REFUSAL DISCIPLINE
Do not score around a blocker
When a critical relationship is unresolved, the appropriate output is a limited or provisional classification plus a verification action—not a more precise-looking decimal. If the unresolved relationship could change permitted use, the result should explicitly remain MORE EVIDENCE REQUIRED. See the uncertain-result guide.
Decision Delta is a categorical sensitivity check
The useful question is not “how many score points disappear?” It is “does weakening this evidence change how the candidate may be treated?”
- Change one material condition. Weaken a buyer, offer, commercial-demand or comparison signal that could matter.
- Re-run the treatment. Check whether the candidate remains usable in the same role.
- Name the transition. Preserve the changed treatment and the evidence that caused it.
- Set the next-best-evidence task. Investigate the condition most capable of restoring, weakening or refusing the treatment.
WHY THIS MATTERS
Robustness is about decision stability
A candidate whose permitted use changes when one critical evidence state weakens is decision-sensitive even when its overall numerical fit still looks high. Reopen or rerun only when a named material condition changes; elapsed time alone is not a trigger. See the rerun guide.
Worked arithmetic example
This example demonstrates the calculation, not an authoritative competitor verdict.
Buyer overlap22 × 1.00 = 22.00
Offer overlap20 × 0.67 = 13.40
Commercial-demand overlap16 × 1.00 = 16.00
Geographic relevance14 × 0.67 = 9.38
Customer-comparison likelihoodUnknown = 0.00
Decision relevance14 × 1.00 = 14.00
Illustrative provisional fit74.78 / 100
Classification and archetype answer different questions
Strong fit
High supported relevance with sufficient coverage and critical gates satisfied.
Partial fit
Meaningful overlap exists, but the evidence or competitive relationship is incomplete.
Weak fit
Some similarity exists without enough supported competitive relevance.
Exclude or investigate
Evidence points outside the set or remains too unresolved for safe inclusion.
Market Twin
High similarity in commercial market role.
Audience Overlapper
Meaningful shared audience but incomplete substitution.
Category Neighbor
Adjacent category or offer with bounded overlap.
Search Shadow
Visible in search or AI discovery without enough evidence of buyer substitution or business rivalry.
Unverified Rival
Plausible candidate whose material evidence remains unresolved.
What can change the method version?
The current public contract is Version 1.0. A new version should be created when decision logic changes materially—not for ordinary copy, accessibility or layout maintenance.
| Change | Governance treatment |
|---|---|
| Copy, spelling, metadata or accessibility repair | Page maintenance; no method-version claim needed unless the change alters decision meaning. |
| Clarifying an existing rule without changing its effect | Document the clarification; review whether a version increment is necessary. |
| Changing a dimension, weight, answer value, coverage threshold or critical gate | Material methodology change; create a new method identity and retest fixtures/parity. |
| Changing archetype meaning or classification logic | Material methodology change; document migration and decision impact. |
Method limits
Not market truth
The method structures evidence supplied to it. It cannot independently observe private customer behaviour, pipeline or commercial substitution.
Not a threat score
Fit does not estimate the magnitude of commercial threat, market share or future performance.
Not budget authorization
A fit classification may inform later analysis; it does not approve spend or competitor-response action.
Not permanent
A named material change in buyer, offer, geography, substitution, comparison evidence or evidence quality can justify a rerun. Passage of time alone does not.
Competitor Fit methodology: quick answers
These are the shortest safe answers to the questions most likely to be extracted out of context.
How is Competitor Fit calculated?
Relationship Fit is the sum of six weighted lens values: Buyer overlap 22, Offer overlap 20, Commercial-demand overlap 16, Geographic relevance 14, Customer-comparison likelihood 14 and Decision relevance 14. Strong fit additionally requires at least 67% answer coverage and the critical competitive relationships to be supported.
Does Unknown mean “no competitor fit”?
No. Unknown contributes 0.00 toward provisional Relationship Fit but remains unresolved. It is not evidence of absence and cannot be used as a hidden negative.
Can search visibility prove a business competitor?
No. Search or AI visibility can identify candidates, but business rivalry still requires evidence of commercial relevance such as shared buyers, realistic substitution, customer comparison or decision impact.
When should the challenge be rerun?
Rerun only when a named material condition changes—such as buyer, offer, geography, substitution evidence, comparison evidence or evidence quality. Do not rerun merely because time passed.
Evidence and source roles
External sources benchmark evidence and competitor-analysis practice. The six-dimension weighting, coverage rule, classifications and archetypes are SERPIFAI methodology and are not supplied or certified by those sources.
- CEBMa — Evidence-Based Decision Making — evidence quality, multiple source types and decision use.
- Google Search Central — Guidance on third-party SEO tools, services and advice
- Ahrefs — Competitive Analysis — direct competitors, keyword overlap and commercial relevance.
- Semrush — Competitive Analysis — current multi-surface competitor analysis practice.
- Semrush — SEO competitor analysis — explicit separation between direct business rivals and organic/search competitors.
- Ahrefs — SEO competitor analysis — search competitors may differ from direct business competitors.
- Klue — Buyer/deal evidence in competitive enablement
- Crayon — Win/loss analysis
- Wikidata Query Service — independently attributable structured entity context
Use the method on one candidate
Start with evidence you can defend. Choose Unknown rather than inventing precision.
OTHER RELEVANT COMMERCIAL PATH
Run the challenge →Run the Competitor Fit Challenge
The result should preserve Answer → Proof → Next: Relationship Fit, Evidence Readiness, critical gates and the next-best-evidence task remain separate rather than collapsing into one confidence number.