Designing incentives that enforce themselves: when supervision is not enough

A contractor’s work is slipping. Quality is drifting. The instinctive response is more oversight: extra reporting, a weekly meeting that becomes twice weekly, someone on site to check, an escalation process. Each step is reasonable on its own. Together they create a recurring cost that produces information rather than performance.

Supervision has limits. It grows with the scope of the work while its effect fades with distance. You cannot watch everything, you usually learn about failures after they happen, and the people doing the work always know more about its true state than you can discover from outside. Contracts allocate blame afterwards. Reviews describe what has already happened. Neither changes the odds that the work is done well when nobody is looking.

A different question is available: can the other party’s economics be arranged so that finishing on time and to standard is what they would choose in their own interest? Where the answer is yes, supervision becomes a check rather than the main control, and the cost of assurance falls as performance rises. This article explains how to think about that question, the common structures that create self-enforcing incentives, where they go wrong and how to apply them in a small business.

The information gap

In almost every commercial relationship, the party doing the work knows its real condition and the party paying sees a report. This information asymmetry can be narrowed by inspection and reporting, but only imperfectly and at a cost. The thinner the margin on the work, the stronger the temptation to manage the report rather than the work, because the report is cheaper to improve.

Penalties are weaker than they look

Penalty clauses, such as liquidated damages for late completion, are the usual answer. They are weaker than they appear. A capable bidder prices the expected penalty into its bid, so the buyer effectively pays for insurance. Penalties are invoked late, argued over and often settled at a discount. They transfer consequences after the event. They do not change behaviour during the work, when behaviour can still change.

Common misreadings

• More oversight means more performance. Beyond a point, it produces documentation and damages the relationship that gives early warning. People stop reporting problems early when doing so triggers an audit.
• Penalties are incentives. An incentive changes what someone wants to do. A penalty changes what they pay if caught. Sophisticated parties price the second and manage the first.
• Alignment is a matter of culture. Partnership language and joint workshops are worth having, but they rarely survive the moment one party’s margin depends on the other’s concession.
• Aligning the company aligns the people. Work is done by individuals whose own rewards may point elsewhere. A site manager rewarded on monthly cost behaves differently from a firm rewarded on completion.

Design claims, not controls

A control is something you operate. A claim is something the other party holds and will defend without being asked. The key question becomes: what can this party hold whose value depends on the outcome we want? Not on their effort, which is hard to observe, or on a metric, which can be gamed, but on the outcome itself. When the party with the best information about the work is also the party most exposed to its failure, much of the supervision problem solves itself.

Structures that create self-enforcing incentives

Several familiar arrangements do this, with varying strength:

• Retentions: holding back part of the payment until a defects period has passed, so the contractor has a reason to fix problems promptly.
• Payment on accepted outcomes: linking significant payments to demonstrated performance, such as a machine reaching agreed throughput, rather than to effort or progress.
• Gain-share and pain-share: sharing savings or overruns against a target, so both parties benefit from efficiency.
• Performance-linked fees: part of the fee depends on results the customer can measure.
• Allocation of future work: under ongoing arrangements, giving more work to suppliers who perform well.
• Deferred rewards: for advisers or key contractors, part of the payment deferred and linked to outcomes over time.
• Payment in kind: paying part of the price in a share of the output, such as finished units, production capacity or a revenue share.

Payment in kind

Paying part of the price in a share of the finished output is a particularly strong form. A builder takes some finished units as part of their payment. An equipment supplier takes a share of the revenue from the line it installs. The supplier now holds something whose value falls if the work is late or poor. You have not persuaded them to care about quality. You have made carelessness costly to them at the moment they would otherwise be tempted.

It works only under certain conditions:

• The output must be worth holding: the supplier must actually want it.
• The share must be material compared with their cash margin.
• Their claim must be secure: if your insolvency or a lender’s prior security could defeat their claim, the incentive disappears exactly when it matters most. Security interests can be complex, so take legal advice.
• The supplier must be able to afford it: offering a non-cash payment to a thinly financed subcontractor may simply be a wage cut in disguise.

Common failure modes include doing excellent work on the units they will own and neglecting the rest, disputes about valuation, and adverse selection, where only suppliers who cannot get cash elsewhere accept the arrangement. Fix the valuation basis before work starts, tie the share to the whole outcome, and offer it as an alternative to a fair cash price to see who chooses it.

Measure outcomes fairly

Outcome-linked payments only work if both parties trust the measurement. Agree in advance exactly what will be measured, how and by whom, what the starting baseline is, how long the measurement period runs and what data the supplier can access. Decide how results will be adjusted for factors outside the supplier’s control, such as raw material problems, changes in product mix or the customer’s own staffing. Set out a simple process for resolving disagreements, such as joint review followed by an independent expert. A well-designed incentive with a disputed measure becomes an argument rather than a motivator.

When a project has stalled

Incentive design matters most when work has already gone wrong, because supervision is then least effective and every party is deciding whether to stay. Three choices carry most of the weight:

• One accountable decision-maker. Stalled work often has many people who can say no and nobody who can say yes.
• Finish something small first. Define the smallest complete piece of work that meets an existing obligation and finish it, restoring cash flow and confidence.
• Cash before margin. In recovery, getting paid matters more than maximising profit on the first piece of work.

Where a party wants to leave a stalled project, staged exit terms linked to completion keep their interest attached to finishing rather than to withdrawing immediately.

Restraints are weaker than incentives

Businesses sometimes try to protect themselves by binding advisers, key staff or partners with long restraints on working elsewhere in the industry. In Australia, restraint-of-trade clauses must be reasonable to be enforceable, and long or broad restraints are often of doubtful effect. The law in this area is also subject to change, so take legal advice for your circumstances. Confidentiality obligations are narrower and usually more defensible, because they address the real concern: misuse of specific information.

The incentive alternative is often more durable. An adviser engaged on a clear mandate, with agreed objectives, regular review and part of their reward deferred and linked to outcomes, has a continuing reason to stay useful and a real cost in walking away. A restraint that cannot be enforced is an illusion of protection. An incentive that never needs enforcing is an asset.

A six-question test

Before signing a significant arrangement, ask:

If you cannot answer “what do they lose, automatically, if this fails?” in one sentence, you have designed supervision and called it alignment. And if the answer depends on going to court, treat it as a penalty, not an incentive.

A worked example

This is an illustration. A small food manufacturer engages an automation integrator to upgrade a packing line for $220,000. The first contract has a fixed price, monthly progress payments, weekly reporting meetings and delay damages of $300 a day. The integrator focuses on reaching billable milestones. At handover, the line works but runs well below its design speed, and the integrator’s best engineer has already moved to another job.

For the next project, the owner redesigns the payoff:

• Retention: 10% of the price, $22,000, held until the line has run stably for 90 days.
• Performance payment: 20% of the price, $44,000, paid when the line achieves the agreed throughput for four consecutive weeks.
• Gain-share: the integrator receives 20% of the extra contribution from throughput above the target for six months, capped at $15,000.
• Individual alignment: the integrator agrees that its lead engineer’s bonus will be linked to the same performance milestone.
• Less reporting: weekly meetings become fortnightly, focused on exceptions.

This time, the integrator stays on site to tune the line, because its own payment depends on it. The line reaches target throughput in the sixth week. The owner spends less time supervising and more time preparing the operators.

How this applies to a small Australian business

Small businesses often lack the resources for heavy supervision, which makes self-enforcing arrangements especially useful. Practical steps:

• Map what each key supplier or contractor actually gains and loses if the outcome fails.
• Link significant payments to accepted outcomes, not just progress.
• Use retentions and defects periods appropriately.
• Consider gain-share where both parties can influence a measurable result.
• Align the people doing the work, not just the companies.
• Use future work allocation as an incentive under ongoing arrangements.
• Take legal advice on securities, restraints and contract terms. Security of payment laws in the construction industry also affect how payments and retentions work.

The articles on inspection and test plans for supplier work and standing offers for repeat work cover related practices.

Signals worth watching

• Assurance costs rising as a share of contract value.
• Defects found late, at handover or in use.
• Penalty clauses priced into bids.
• Suppliers’ key people moving on before the work is proven.
• Restraint clauses relied on without checking their enforceability.
• Recovery situations with no single accountable decision-maker.

Common mistakes

• Adding oversight as the only response to poor performance.
• Treating penalties as incentives.
• Linking payment to effort or progress rather than outcomes.
• Aligning the company but not the individuals.
• Offering non-cash payment to parties who cannot afford it.
• Relying on unenforceable restraints.

Frequently asked questions

Will suppliers accept outcome-linked payments? Good suppliers often welcome them when the outcome is fair, measurable and partly within their control, especially with a share of upside. Expect a higher price if they carry more risk, and judge whether the improved incentive is worth it.

How large should an outcome-linked payment be? Large enough to matter to the supplier relative to its margin on the job, and small enough that it does not threaten its viability if something outside its control goes wrong. Many arrangements put a meaningful minority of the price at risk.

What if the outcome depends partly on us? Define clearly what each party must do, such as providing materials or trained operators on time, and adjust payments if your side does not deliver.

Can incentives replace inspection entirely? No. Incentives change the odds of good work, but inspection still matters for safety, compliance and critical quality points. The aim is to rely less on supervision as the main control, not to abandon checking.

Is this only for large contracts? No. Even small jobs can use retentions, payment on completion and future work allocation.

Questions to ask

• For our three largest supplier relationships, what does the supplier lose automatically if the outcome fails?
• How much are we spending on supervision, and what would we spend if incentives carried more of the load?
• Have we checked whether our restraint clauses are enforceable?
• Who has single authority in our current problem projects?
• If a supplier treated our incentive structure as a puzzle, where would they solve it?

Bringing it together

Supervision detects failure after it happens. Incentive design changes the odds of it happening. Map what the other party actually holds and loses, link significant payments to outcomes, use retentions, gain-share, future work and, where suitable, payment in kind, align the individuals as well as the organisations, and check that stakes are secure and material. Choose good counterparties first, and keep looking, because a capable counterparty will test any structure. The best arrangement is one both parties would follow even if nobody were watching.

---

Source: KEVOS notes. Examples and figures in this article are illustrations. This article is general information, not legal or financial advice.