Conduct Quotient
Conduct Quotient (CQ) is one private score for one named context. It is produced from separately governed human-experience and verified-customer-record inputs only after both qualify.
PLAIN-LANGUAGE SUMMARY
What this page establishes
The working CQ display runs from 1.00 to 5.00 and always shows two decimal places when a score exists. The familiar format is meant to make the result easy to read, not to claim more certainty than the evidence supports.
Net Conduct Score (NCS) and Verified Conduct Index (VCI) remain independently valid, inspectable, and challengeable beneath CQ. A material problem in one cannot be averaged away by strength in the other. When the evidence is incomplete, disputed, invalid, or irrelevant, the responsible result is a named non-score state.
Definition and role
CQ MUST name its scope. A company-specific CQ, an industry CQ, and an overall CQ are different representations and MUST NOT be substituted for one another. The initial live sequence begins with company-specific CQ. Overall CQ remains outside the first live pilot even if its shadow research advances.
A displayed CQ MUST retain an auditable route to both protected inputs, the active evidence, the named scope, and the calculation version.
One score, two protected inputs
The private consumer surface leads with one contextual CQ. The consumer can open How your CQ was formed to inspect the human-experience and verified-customer-record inputs. Workers and ordinary receiving businesses do not receive those values.
Human experience
Approved records
ELIGIBILITY RESULT
FIGURE 7.2-A The 4.81 value is a synthetic interface example. The score is visible to the consumer; a receiving business ordinarily receives only the approved result needed for one benefit.
PROVISIONAL RESEARCH FAMILY
CQinternal = min((N + V) / 2, min(N, V) + κ)
N and V are separately valid contextual inputs on an internal 0–100 scale. Suite K used κ = 10 as a research setting. The public display candidate maps the result to 1.00–5.00 and rounds to the nearest hundredth, half-to-even.
K3 capped arithmetic is the provisional structural lead. K6 weakest-link anchored synthesis, a two-dimensional explanation, and proof-only output remain research controls. No coefficient, mapping, or score is authorized for production.
Evidence gates and non-score states
Each required input MUST separately pass its construct, provenance, attribution, sampling, sufficiency, independence, fairness, rights, and dispute gates. Confidence controls whether a score is available; it does not compensate for weak standing.
| Condition | Consumer result | Status |
|---|---|---|
| Eligible evidence has not reached the display threshold | Building history | REQUIRED |
| A material input or calculation is being challenged | Under review | REQUIRED |
| The result cannot currently be produced safely | Temporarily unavailable | REQUIRED |
| Evidence does not support the requested scope | Not available in this context | REQUIRED |
| The consumer has not enrolled or has withdrawn | Not participating | REQUIRED |
These states are not low scores and MUST NOT reduce baseline service. Missing, invalid, disputed, corrected, or prohibited evidence MUST NOT be converted into adverse standing.
Consumer display and business disclosure
A consumer may see an available CQ, its scope, evidence state, last material update, input explanations, active records, and challenge route. The first-skim view MUST NOT lead with two competing input scores or a raw operational fraction.
A receiving business learns only whether the consumer qualifies for one approved, additional benefit. It does not receive the CQ, NCS, VCI, source companies, or underlying events.
The Conduct layer governs eligibility and the limited proof. A separate loyalty, promotion, commerce, or customer-experience system may deliver the approved benefit, but it MUST NOT recalculate CQ, retain the result as a general customer segment, infer a failed result, or make baseline service depend on participation or eligibility.
Incentive-orchestration platforms are therefore a potential implementation-partner category rather than part of the Conduct standard itself. Talon.Oneis one current market example. Open Conduct has no partnership with or endorsement from Talon.One, and no provider is required by this proposal.
Broader score sharing may be researched only when the consumer initiates it and the exact use independently passes necessity, comprehension, privacy, fairness, legal, and baseline-service review. Conduct does not create public profiles, leaderboards, searchable people ratings, or a way to deny ordinary service.
Open questions
OQ-001Do people understand the 1.00–5.00 CQ, its scope, evidence state, and limitations?+
The familiar format could improve comprehension while also importing false assumptions about precision, rank, and moral judgment. The display, persistence, and explanations require human testing.
Respond to this questionOQ-113Which coefficients, calibration, thresholds, and context boundaries should instantiate K3 and L4?+
Suites K and L select structural families, not production parameters. Real-data replay, fairness analysis, human research, and independent review must determine whether any implementation advances.
Respond to this questionOQ-125Could an overall CQ ever earn release beyond shadow research?+
An overall score is part of the long-term product ambition but excluded from the initial live pilot. It would need to preserve context differences, prevent dominance, improve understanding, and pass fairness, necessity, proportionality, and legal review.
Respond to this questionOQ-133Does K3 with the ten-point research cap outperform K6 and alternative cap values?+
The comparison must examine real validity, fairness, stability, manipulation resistance, proxy artifacts, and explanation. The current cap is a research setting, not a production coefficient.
Respond to this questionOQ-134Which interface makes CQ understandable without inviting anxiety, constant monitoring, or moral grading?+
The score, scope selector, input explanation, non-score states, and change history need comparative UX, language, accessibility, cross-cultural, and longitudinal research.
Respond to this question