Stats
Actions
Tags
From grimoire
Designs behavioral feedback systems (performance management, compliance, software quality loops, habit formation) so consequences are clear, certain, and swift.
How this skill is triggered — by the user, by Claude, or both
Slash command
/grimoire:apply-consequence-clarityThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Design consequences to be clear, certain, and swift — because certainty and timing determine behavior modification more than magnitude, and a small immediate consequence changes behavior more reliably than a large delayed one.
Design consequences to be clear, certain, and swift — because certainty and timing determine behavior modification more than magnitude, and a small immediate consequence changes behavior more reliably than a large delayed one.
Shang Yang's 赏刑篇 (~4th century BC) articulates four required properties of effective rewards and punishments: 明 (clear — people know exactly which behavior leads to which consequence), 必 (certain — consequence reliably follows the behavior), 速 (swift — minimal delay between behavior and consequence), and 平 (proportionate — severity matches the offense). His argument: a legal system with low certainty of enforcement is worse than no system — it teaches people that violations are low-risk. The consequence properties matter more than their magnitude. This principle was the operating foundation of Shang Yang's legal reforms in Qin, which transformed Qin from a weak peripheral state into the power that ultimately unified China.
B.F. Skinner — operant conditioning: The foundational finding in behavioral psychology: reinforcement timing and consistency determine behavior change far more than reward magnitude. A lever-press rewarded immediately produces rapid conditioning; the same reward delivered 30 seconds later produces no conditioning. Fixed-ratio schedules (certain reward every N responses) produce sustained behavior; variable-ratio schedules (uncertain) produce higher frequency but lower stability. This research is the basis of behavioral therapy, organizational behavior management, and behavioral economics globally.
Daniel Kahneman & Amos Tversky — Prospect Theory (1979, Nobel 2002): People discount future consequences exponentially with delay (temporal discounting). A $100 consequence tomorrow is psychologically valued much higher than a $100 consequence in six months, even though the amounts are identical. Consequence timing determines psychological impact, not just nominal value.
Nagin "Deterrence in the Twenty-First Century" (Journal of Crime and Justice, 2013): Systematic meta-analysis across 50+ years of criminology literature. Finding: certainty of punishment reduces crime significantly; severity of punishment has near-zero marginal effect beyond a threshold. Three-strikes laws and mandatory minimums — focused on severity — produce minimal deterrence effects. Swift and certain enforcement programs produce measurable behavioral change. The most-cited contemporary deterrence research; influences criminal justice policy in the US, UK, and EU.
Adobe, Deloitte, Goldman Sachs — real-time feedback movement (2012–2020): All three eliminated annual performance reviews in favor of immediate, continuous feedback:
CI/CD pipelines — software engineering: The shift from weekly/monthly integration builds to immediate automated test feedback fundamentally changed developer behavior: defect introduction rates dropped because developers received feedback on their code within minutes of writing it, while the code was still in working memory. The behavioral impact comes from swiftness, not from the content or magnitude of the test output. Standard engineering practice at Google, Netflix, Stripe, Airbnb, and virtually every modern software organization.
Why best: Delayed consequences break the behavior-consequence association in memory — when the consequence arrives, the behavior is no longer cognitively linked to it, so no learning occurs. Uncertain consequences teach people to model the probability of consequence, and if that probability is below ~30–40%, most people discount it to zero. Neither severity escalation nor rule republication addresses these failure modes. Only clarity, certainty, and swiftness restore the behavior-consequence link.
Distinct from design-feedback-loops: design-feedback-loops addresses information feedback for organizational decision-making — how signals travel through a system so decision-makers can adapt. It explicitly excludes individual behavior modification (the file notes this "responds better to habit design frameworks"). apply-consequence-clarity addresses behavioral consequence design: the properties of consequences that produce or fail to produce behavior change in people.
Distinct from design-incentive-systems: design-incentive-systems addresses structural alignment — designing the system so good outcomes are the rational self-interested choice for any actor. It addresses WHETHER an incentive structure exists. apply-consequence-clarity addresses the delivery properties of consequences — HOW CLEARLY, HOW CERTAINLY, and HOW QUICKLY consequences follow behavior. A structurally aligned incentive system can still fail if consequences are uncertain or delayed.
Distinct from apply-transparent-rules: apply-transparent-rules addresses rule publication — making rules explicit and public before situations arise. It ensures people KNOW the rules. apply-consequence-clarity addresses what happens after a behavior: whether the consequence that follows is clear (unambiguous), certain (reliably occurs), and swift (occurs without significant delay). Both are necessary; neither substitutes for the other.
Adopted by: Adobe, Deloitte, and Goldman Sachs (eliminated annual reviews for immediate continuous feedback); Google, Netflix, Stripe, and Airbnb (CI/CD immediate automated test feedback as standard engineering practice); criminal justice reform programs implementing swift-and-certain enforcement (Nagin meta-analysis basis).
Impact: Adobe reported a 30% reduction in voluntary turnover after implementing its Check-in immediate feedback system (2012); Nagin's meta-analysis of 50+ years of criminology literature found that certainty of punishment reduces crime significantly while severity of punishment has near-zero marginal effect beyond a threshold.
Map the behavior-consequence chain for each target behavior. For each behavior you are trying to increase or decrease, identify: what is the current consequence? When does it occur? What is the probability that it occurs given the behavior? Is the connection between behavior and consequence visible to the person exhibiting the behavior? Most consequence failures are visible at this mapping step.
Test clarity: can the person predict the consequence before acting? A consequence is clear when a person can state, before acting, exactly what consequence will follow. Ask the people whose behavior you're trying to shape: "If you do X, what happens to you?" If they cannot answer precisely — or give inconsistent answers — the consequence is unclear. Fix: state the consequence explicitly, in writing, before the situation arises (with apply-transparent-rules as the prerequisite).
Test certainty: what fraction of violations or achievements receive a consequence? Measure the actual probability of consequence given behavior over the last 90 days. A consequence that occurs in fewer than 70–80% of qualifying events is probabilistic enough that actors will model and discount it. Fix: design structural mechanisms that make consequence delivery automatic rather than discretionary. Automated tests, system-enforced approval workflows, and mandatory documented reviews all increase certainty without requiring discretionary action.
Test swiftness: how long is the behavior-consequence delay? Measure the actual elapsed time between the behavior and the consequence. Plot the distribution. Behavioral research is consistent: consequences delayed by more than 24–48 hours have sharply reduced behavioral impact; annual or quarterly consequences have near-zero impact on day-to-day behavior. Fix: redesign the feedback loop to close the gap. If the consequence cannot be made swift, design a proximate immediate signal (a fast visible indicator) that predicts the eventual consequence.
Test proportionality: does the consequence magnitude match the behavior significance? Disproportionate consequences (severe consequences for minor behaviors, trivial consequences for major ones) teach people that the consequence system is not a reliable signal of the system's actual values. Both over-punishment and under-punishment damage behavioral alignment. Fix: calibrate consequence magnitude to behavior significance; document the calibration explicitly.
Prioritize certainty over severity when designing new consequence systems. When building a performance management system, compliance mechanism, or behavioral incentive: invest resources in consequence certainty (making it near-certain that the consequence follows) before investing in consequence severity (making it larger). A small certain consequence is more effective than a large uncertain one.
Separate the design of the consequence system from the application of individual consequences. Consequence clarity requires that the properties (clear, certain, swift, proportionate) are built into the SYSTEM, not into each individual application. When a manager must exercise discretion to decide whether to apply a consequence in each case, certainty collapses to the base rate of discretionary enforcement, which is typically far below the 70–80% threshold required for behavioral impact.
Performance management: Team consistently misses documentation requirements. Current consequence: documentation quality mentioned in annual review. Certainty: ~30% (many violations go unmentioned). Delay: up to 12 months. Redesign: automated check in CI/CD pipeline flags undocumented functions immediately on commit; PR cannot merge until documentation added. New consequence: certain (100% — automated), swift (minutes), clear (specific error message), proportionate (blocked PR, not career consequence). Documentation quality improves within two sprints.
Compliance enforcement: Company policy requires expense reports submitted within 30 days. Current enforcement: finance team follows up occasionally; late reports rarely result in any consequence. Behavioral impact: ~40% of reports are late. Redesign: system automatically withholds reimbursement for reports submitted after 30 days (funds released when report filed). New consequence: certain (automated), swift (no reimbursement until filed), clear (policy stated at submission), proportionate (delay in reimbursement, not discipline). Lateness drops to <5%.
Software quality: Team has high defect rate despite code review requirement. Reviews happen but are superficial. Current consequence for poor code quality: production incident weeks or months later. Delay breaks behavior-consequence link. Redesign: mandatory automated static analysis + unit tests in CI; build fails immediately on quality issues; merge blocked until fixed. New consequence: certain (100% — automated), swift (minutes after commit), clear (specific error messages), proportionate (blocked merge). Defect rate drops within first sprint.
Sales behavior: Sales team advised to complete CRM records after each customer call. Compliance is ~50%. Current consequence: vague feedback in quarterly review. Redesign: pipeline reports (which all sales managers review weekly) are automatically generated from CRM data; incomplete CRM records produce blank pipeline entries, making the salesperson's pipeline appear empty to leadership. New consequence: certain (automatic), swift (next weekly review), clear (visible as blank pipeline), proportionate (reputational, not disciplinary). CRM completion reaches >90%.
npx claudepluginhub jeffreytse/grimoire --plugin grimoireAudits governance and organizational processes to replace virtue-reliance with incentive structures that make good behavior the rational self-interested choice.
Enforces structured troubleshooting and evidence-first delivery habits for repeated failures, passive behavior, or quality issues. Uses big-tech performance culture rhetoric to drive proactive, exhaustive problem-solving.
Maps actual incentives driving behavior, distinguishing stated motivations from real reward structures. Useful for explaining and changing seemingly irrational actions in organizations or systems.