Mechanism design is the reverse of game theory: instead of analysing the outcome of given rules, you choose the rules so that self-interested participants are led to an outcome you want. You design the game — the mechanism — so that rational players, acting on their private information and incentives, produce the desired result. The designer does not need to know private information; the mechanism elicits it or aligns behaviour despite it.
The key idea is incentive compatibility. A mechanism is incentive-compatible when truthful participation (or best-response behaviour) yields the outcome the designer wants. Classic applications: auctions (design rules so bidders reveal true values and the best bidder wins), matching (design so participants report preferences honestly and the match is stable), and contracts (design payoffs so the agent takes the action the principal wants). The revelation principle says that for any mechanism there is an equivalent direct mechanism where everyone reports their type truthfully — so the designer can focus on direct, truthful mechanisms.
In organisations, mechanism design appears in compensation (align pay with desired effort and risk-taking), in allocation (who gets what project, promotion, or resource), and in governance (voting rules, decision rights). The founder or leader is the designer; employees, partners, and counterparties are the players. The mistake is assuming that stating a goal is enough; the mechanism — the actual rules, payoffs, and information structure — determines what happens. Get the mechanism wrong and rational agents will game it or disengage.
The revelation principle is a practical shortcut: you can usually restrict to mechanisms where participants are asked to report their type (information) and truth-telling is optimal. So the designer's job often reduces to designing the right "form" — what to ask, how to map reports to outcomes, what payoffs to attach — so that honest participation is the best response. That's why good incentive design feels like "making it in people's interest to tell the truth and do the right thing" rather than "hoping they will."
Section 2
How to See It
Look for situations where an outcome depends on the choices of multiple parties who have private information or conflicting incentives. When you want a specific result (e.g. truth, effort, efficient allocation), ask: what rules would make it in each party's interest to behave in a way that produces that result?
Business
You're seeing Mechanism Design when a company ties sales compensation to long-term customer outcomes (e.g. retention, NPS) so that reps are incentivised to sell to the right customers, not just to close any deal. The mechanism aligns private incentives with the firm's goal.
Technology
You're seeing Mechanism Design when a platform sets reputation scores, dispute resolution, or pricing rules so that buyers and sellers behave honestly. The design of the marketplace — not just the UI — determines whether participants game or support the system.
Investing
You're seeing Mechanism Design when a fund structures carry, vesting, and governance so that GPs and LPs have aligned incentives over time. The legal and economic structure is the mechanism; it determines whether the GP will maximise fund return or their own short-term payoff.
Markets
You're seeing Mechanism Design when an auction format (e.g. second-price, Vickrey) is chosen so that bidders have a dominant strategy to bid truthfully. The auction is the mechanism; the designer (seller, regulator) chooses it to achieve revenue or efficiency.
Section 3
How to Use It
Decision filter
"When you want others to behave in a certain way, don't just state the goal — design the mechanism. Ask: what do they know that I don't? What do they want? What rules (payoffs, information flow, sequence) would make the desired behaviour their best response? Then implement that mechanism and expect them to respond to incentives, not to exhortation."
As a founder
Design compensation, promotion, and decision rights so that doing the right thing is the rational choice. If you want long-term thinking, tie rewards to long-term outcomes. If you want collaboration, avoid zero-sum metrics. If you want truth-telling, use mechanisms that make lying costly or unrewarding (e.g. anonymous feedback, sealed bids). Test the mechanism: what would a rational, self-interested person do? If that isn't what you want, redesign.
As an investor
Structure terms and governance so that founder and investor incentives stay aligned. Vesting, board rights, and information rights are part of the mechanism. Design for the case where the other party is rational and self-interested; if the mechanism still produces good behaviour, it's robust. If it relies on goodwill, it's fragile.
As a decision-maker
In negotiations and partnerships, the "mechanism" is the contract, the process, and the default outcomes. Design the process so that the other side's best response is to cooperate or reveal information you need. Don't assume they will do the right thing; make the right thing payoff-maximising for them.
Common misapplication: Assuming that announcing a goal changes behaviour. Goals are cheap; mechanisms are what bind. If the incentive structure rewards the wrong behaviour (e.g. short-term revenue over long-term value), people will optimise for the wrong thing. Redesign the mechanism, not the memo.
Second misapplication: Designing for the ideal participant. Mechanism design assumes participants are rational and self-interested. If your mechanism only works when people are altruistic or forget to game it, it will fail. Stress-test against strategic behaviour.
Netflix's "freedom and responsibility" and high salary (no bonus) model is a mechanism: remove the incentive to optimise for bonus metrics and align behaviour with long-term value. The design choice — pay well, don't tie to short-term KPIs — is meant to make the right behaviour (judgment, collaboration) the natural response.
Bill CampbellCoach to Jobs, Schmidt, Pichai; Director, Apple & Google
Campbell emphasised that culture and incentives had to be designed, not hoped for. "You get what you reward" is mechanism design in plain language. He pushed leaders to look at the actual incentives (promotion, compensation, recognition) and fix them when they drove the wrong behaviour.
Section 6
Visual Explanation
Mechanism Design — The designer chooses rules (left). Participants have private information and incentives (centre). The mechanism maps their choices to outcomes (right). Design so that rational play produces the desired outcome.
Section 7
Connected Models
Mechanism design builds on game theory and incentives and connects to adverse selection, signalling, and principal-agent problems. The models below either reinforce it (game theory, incentives), create tension (principal-agent, adverse selection), or extend to applications (Nash equilibrium, signalling).
Reinforces
Game Theory
Game theory studies how players behave under given rules. Mechanism design inverts the question: what rules produce the desired behaviour? The reinforcement: game theory gives you the tools (Nash equilibrium, dominant strategies); mechanism design uses them to choose the game. You need both — analyse the game you're in, or design the game you want.
Reinforces
Incentives
Incentives are what drive behaviour; mechanism design is the discipline of structuring incentives so that the desired behaviour emerges. The reinforcement: "incentives matter" is the premise; mechanism design is the engineering. When you align payoffs with the outcome you want, you're doing mechanism design.
Tension
The Agency Problem
The agency problem is misalignment between principal and agent. Mechanism design is one response: design the contract and information structure so that the agent's best response serves the principal. The tension: agency problems persist when the mechanism is poorly designed (e.g. short-term metrics, incomplete contracts). Fix the mechanism to fix the agency problem.
Tension
Adverse Selection
Adverse selection arises when private information leads to bad sorting (e.g. only risky buyers seek insurance). Mechanism design can mitigate it by designing contracts or procedures that elicit types or screen (e.g. deductibles, warranties). The tension: if you don't design for it, adverse selection will distort outcomes; mechanism design is the tool to counteract it.
Section 8
One Key Quote
"Mechanism design asks: given a social goal, can we design a game so that the equilibrium outcome coincides with that goal? The revelation principle tells us we can restrict attention to direct mechanisms where agents report their types and truth-telling is optimal."
— Eric Maskin (Nobel Lecture, 2007)
The designer doesn't need to observe private information; the right mechanism makes it optimal for participants to reveal it or to act as if they had. The practical lesson: don't rely on asking nicely or on trust when incentives are misaligned. Design the rules so that the right behaviour is the best response.
Section 9
Analyst's Take
Faster Than Normal — Editorial View
You get what you reward. If the mechanism rewards short-term revenue, you get short-term behaviour. If it rewards long-term value, you get long-term behaviour. Before blaming people, look at the mechanism — the actual payoffs, metrics, and decision rights. Change those first.
Design for the strategic player. Assume everyone will optimise within the rules. If the mechanism only works when people don't game it, it will fail. Stress-test: what would a rational, self-interested person do? If that's bad, redesign.
Information and incentives are the two levers. The designer controls what information is revealed, when, and to whom — and what payoffs follow from actions and reports. Use both. Sometimes the fix is better information (e.g. transparent metrics); sometimes it's different incentives (e.g. change the bonus formula). Often it's both.
Contracts and process are mechanisms. In partnerships, investing, and employment, the contract and the process define the game. Negotiate and draft with mechanism design in mind: what behaviour does this structure induce? Align the structure with the outcome you want.
Section 10
Test Yourself
Is this mental model at work here?
Scenario 1
A company switches sales comp from 'revenue closed' to 'revenue retained after 24 months.' Rep behaviour shifts toward qualifying leads and avoiding churn.
Scenario 2
A board sets a CEO's bonus on annual earnings. The CEO focuses on short-term earnings and defers investment. The board is surprised.
Scenario 3
A platform introduces a reputation score that affects visibility. Sellers start gaming the score (fake reviews, incentives) instead of improving service.
Scenario 4
A partnership contract ties payment to verifiable milestones and includes clawbacks if the partner underdelivers. Both sides behave cooperatively.
Section 11
Summary & Further Reading
Summary: Mechanism design is choosing the rules of the game so that self-interested participants, acting on private information, produce a desired outcome. It is the reverse of game theory: you design the mechanism (rules, payoffs, information) so that equilibrium behaviour is what you want. Incentive compatibility — making the desired behaviour a best response — is the core. Use it when leading and negotiating: design compensation, contracts, and processes so that the right behaviour is rational for the other party. Don't rely on goals or exhortation; design the mechanism and expect people to respond to incentives.
Auctions as designed mechanisms. Practical and theoretical treatment of auction design for revenue and efficiency.
Leads-to
Nash Equilibrium
Nash equilibrium is the outcome when no player wants to deviate. In mechanism design, the goal is to choose a mechanism whose Nash equilibrium (or dominant-strategy equilibrium) is the desired outcome. The link: you design the game so that the equilibrium of the game is the outcome you want.
Leads-to
Signalling & Countersignalling
Signalling is costly action that reveals type. In mechanism design, the designer may use signals (e.g. qualifications, collateral) as part of the mechanism to screen or elicit truth. The connection: mechanism design often involves designing what signals are required or rewarded so that participants reveal information or sort themselves.
