# Incentives

> Predict and shape behavior by understanding the hidden rewards and punishments that drive human action

- Skill: `lev-os/incentives` (Agent Skill)
- Install (CLI): `npx skillmds@latest add lev-os/incentives`
- Raw SKILL.md: https://api.skillmd.com/api/skills/lev-os/incentives/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: lev-os (https://skillmd.com/u/lev-os)
- Updated: 2026-09-10
- Page: https://skillmd.com/skills/lev-os/incentives

---


# Incentives

## Overview

Incentives are the invisible forces that drive human behaviorthe carrots and sticks, rewards and punishments, that shape our choices and actions. Charlie Munger considered incentives one of the most powerful mental models, famously saying: "Never, ever think about something else when you should be thinking about incentives." People respond predictably to incentive structures, often in ways that surprise system designers. Understanding incentives means understanding behavior; shaping incentives means shaping outcomes.

## When to Use

- **Organizational design**: Structuring compensation, promotions, and recognition systems
- **Investment analysis**: Evaluating management decisions by examining what they're incentivized to do
- **Negotiation**: Understanding what the other party truly wants beyond stated positions
- **Culture building**: Aligning team behavior with company values through reinforcement structures
- **Problem diagnosis**: When outcomes don't match intentions, audit the incentive system
- **Product design**: Creating user behavior loops through reward mechanisms

## The Process

### Step 1: Identify Stated vs. Actual Incentives
What are people officially supposed to do? What are they actually rewarded or punished for? The gap reveals the true incentive structure.

**Example**: Company says "we value quality" but promotes managers who ship fastest, regardless of bugs. Actual incentive: Speed over quality.

### Step 2: Map All Stakeholders and Their Incentives
List every party involved and what they optimize for. Don't assume everyone shares your goals.

**Example**: Real estate transaction
- Buyer: Lowest price, best value
- Seller: Highest price, quick close
- Buyer's agent: Commission = 3% of price, wants higher price + fast close
- Seller's agent: Same incentives as buyer's agent
- Lender: Closing loan, doesn't care about price
- Inspector: Finding problems (justifies fee, repeat business from agents)

Surprise: Your agent is incentivized to close quickly at slightly higher price, not negotiate hardest for you.

### Step 3: Predict Behavior Based on Incentives
Given these incentives, how will each party behave? Ignore what they say; predict based on what they're rewarded for.

**Example**: Sales team paid on bookings (signed contracts) not revenue (collected payments). Prediction: They'll discount heavily to close deals, promise features that don't exist, and hand off bad-fit customers to support. Result: High churn, low profitability.

### Step 4: Test for Perverse Incentives
Look for unintended consequences where incentives produce opposite of desired outcome.

**Classic examples**:
- Cobra bounty in colonial India: Paid per dead cobra, people started breeding cobras to kill for bounty
- Wells Fargo quotas: Pressure to open accounts led to millions of fraudulent accounts
- Lines of code metrics: Developers write verbose, inefficient code to hit targets

### Step 5: Redesign Incentives to Align with Goals
Change the structure to reward desired outcomes, not easy-to-game proxies.

**Framework**: Incentivize outcomes, not activities. Measure impact, not motion.

**Example**: Instead of "support team measured by tickets closed per hour" (incentivizes rushing, not helping), use "customer satisfaction scores + first-contact resolution rate" (incentivizes actually solving problems).

### Step 6: Install Countermeasures Against Gaming
People will find creative ways to exploit any incentive system. Build in checks.

**Strategies**:
- Multiple metrics (hard to game all simultaneously)
- Spot audits for quality
- Reputation systems (long-term consequences)
- Skin in the game (share downside risk, not just upside)
- Cultural norms that shame gaming

## Example Application

**Scenario**: SaaS company struggling with customer churn after rapid growth

**Step 1 - Stated vs. actual**:
- Stated: "Customer success is our priority"
- Actual incentives: Sales paid on new bookings, not net revenue. Customer success team measured on renewal rate, but bonuses based on upsells.

**Step 2 - Stakeholder incentives**:
- Sales: Close deals fast, hit quarterly quota, promise anything
- CS: Keep customers from canceling, push expensive upgrades
- Product: Ship features sales promised (reactive), not what customers need
- Executives: Show growth metrics to investors (vanity metrics)

**Step 3 - Predicted behavior**:
- Sales will close bad-fit customers who churn in month 2-3
- CS will discount renewals to prevent churn (hides problem)
- Product will build one-off features for big customers (no leverage)
- Execs will celebrate new logos, ignore unit economics

**Step 4 - Perverse incentives found**:
- Paying sales on bookings incentivizes selling to anyone, even poor fits
- Measuring CS on renewal rate incentivizes discounting, not value delivery
- Product roadmap driven by loudest customer, not strategic vision

**Step 5 - Redesign**:
- Sales: 50% commission on booking, 50% on customer hitting 12-month mark (aligns with customer success)
- CS: Bonuses tied to NPS and expansion revenue from happy customers, not preventing cancellations
- Product: Measured on feature adoption rates and customer-reported value, not shipped features
- Execs: Report on cohort retention curves and LTV:CAC ratio, not just new logos

**Step 6 - Countermeasures**:
- Quarterly "ideal customer profile" fit score for new deals (penalize sales for bad fits)
- CS shares customer feedback in all-hands (social pressure against gaming)
- Product requires business case showing cross-customer value before building features
- Finance tracks true revenue (not bookings) as primary growth metric

**Result**: Churn drops from 8%/month to 3%/month over 6 months. Slower new logo growth, but higher quality customers. LTV increases 3x. Team aligned on sustainable growth.

## Anti-Patterns

**Assuming goodwill trumps incentives**: "Our team is mission-driven, they won't game the system." False. Even well-intentioned people respond to incentives. Design as if people will optimize for personal benefit, because they will.

**Single-metric optimization**: Any single KPI becomes the target and stops being useful (Goodhart's Law). Use balanced scorecards with competing tensions.

**Ignoring status and identity incentives**: Not all incentives are monetary. Recognition, titles, belonging, and autonomy often drive behavior more than cash.

**Complex incentive schemes**: If people can't explain how they're compensated, they can't optimize for it. Complexity creates confusion, not alignment.

**Static incentives in dynamic environments**: What worked last quarter may be wrong this quarter. Markets change, strategies evolveincentives must adapt.

**Incentivizing the wrong people**: Rewarding executives for stock price while individual contributors get pizza parties. Misalignment breeds cynicism.

## Related Frameworks

- **Principal-Agent Problem**: Misaligned incentives between decision-makers (agents) and those affected (principals)
- **Moral Hazard**: When people take risks because they don't bear the full consequences (incentive mismatch)
- **Tragedy of the Commons**: Individual incentives to exploit shared resources destroy collective value
- **Hanlon's Razor**: Before assuming malice, check if bad behavior is simply responding to perverse incentives
- **Second-Order Thinking**: Incentive changes create ripple effectspredict downstream consequences
- **Goodhart's Law**: When a measure becomes a target, it ceases to be a good measure (incentive gaming)
- **Skin in the Game**: Aligning incentives by ensuring decision-makers share in both upside and downside