Bonuses boost simple tasks; peer review supports quality
Marcus Reed, Small-business automation editor (AI-assisted) · · updated
Research finds that high-stakes bonuses reliably increase output in simple, mechanical work, but actually reduce performance on tasks that require thought or judgment. Meanwhile, peer review is used across many fields to assess quality and improve standards, but it is not a direct substitute for incentive pay. Here’s what the evidence really shows — and what to do differently if you want your team’s incentives to pay off.
Bigger bonuses help only with simple, mechanical tasks
A series of experiments funded by the Federal Reserve Bank and conducted at MIT, Chicago, and Carnegie Mellon showed a clear break: for tasks that are purely mechanical, large performance-related pay (such as big bonuses) increases performance as expected. But as soon as a task requires even basic cognitive skills — memory, reasoning, creativity, or judgment — larger incentives actually make performance worse, not better[2]. This surprising finding held true across different cultures and income levels, including studies in Madurai, India, where the bonuses represented significant sums for participants[2].
In practice, this means that if your team’s work is repetitive or routine (for example, data entry or manual assembly), tying pay closely to output can reliably boost productivity. But if the work demands problem-solving or judgment — like customer service, sales with negotiation, or design — offering larger and larger bonuses is likely to backfire. Employees may feel increased pressure, become anxious, and perform worse than they would with moderate incentives.
Variable pay attracts different people: the sorting effect
Incentive systems don’t just change behavior — they change who sticks around. Plans with variable compensation (such as bonuses, profit-sharing, or performance-based pay bands) tend to attract and retain people with higher skill levels and a stronger sense of personal control over their results (a lower “external locus of control”)[1]. This is known as the sorting effect.
For a business, this means that moving from fixed salaries to variable pay schemes doesn’t just motivate existing staff; it also shapes the team over time. Employees who dislike risk or feel that outcomes are out of their hands may leave, while those confident in their own skills are more likely to join and stay[1]. This can raise the average ability of the workforce, but may also reduce diversity of temperament and approach.
The sorting effect can be a double-edged sword. It works well for roles where measurable, individual results matter. But in jobs where collaboration, stability, or process adherence are key, a workforce selected mainly for self-confidence and competitiveness may create new problems.
Peer review maintains standards, not speed
Peer review — having work evaluated by colleagues with similar expertise — is common in fields where quality, safety, or compliance matter: healthcare, law, academia, engineering, and more[3]. In these settings, peer review helps maintain standards, root out errors, and uphold credibility, rather than directly boosting production or output.
Unlike performance bonuses, peer review is not typically tied to pay or promotion. Instead, it provides professional feedback, validates competence, and can serve as a gatekeeper for advancement or credentialing (as in academia or medicine)[3]. For example, in healthcare, peer review is used to assess whether clinical practice meets standards, and in engineering, technical peer reviews identify defects before products are released[3].
Peer review can improve learning and performance, especially in complex or high-risk work. However, it requires time and trust, and can be undermined by bias or conflicts of interest (role duality)[3]. It is most effective when reviewers are appropriately qualified, the process is transparent, and results are used to support improvement, not just policing.
A worked example: when bonuses backfire
*Illustrative scenario; not a real case or a cited study.*
Suppose a company offers a substantial bonus to call center agents for handling more calls per hour. At first glance, this matches the logic of performance pay for simple, measurable tasks. But actual call center work involves diagnosing problems and resolving customer issues—tasks that draw on judgment, not just speed.
In this illustration, agents focus on call volume to maximize their bonus. Calls per hour climb, but agents rush, leading to more unresolved issues, higher repeat call rates, and an uptick in customer complaints. Customer retention drops, erasing the productivity gains. The bonus, intended to drive output, backfires because the job demands more than mechanical throughput.
This reflects what research confirms: large performance bonuses improve results when the task is purely mechanical and quality is easy to check, but when cognitive effort or judgment is involved, high-stakes incentives can reduce performance instead[2].
Why gaming the metric is the risk in pay-for-performance
Whenever pay is tied to a single, measurable target, staff attention goes to the number—sometimes at the cost of the work quality or customer experience you actually want. For example, if a call center rewards agents for the shortest call times, it is common for agents to focus on ending calls quickly, even if that means unresolved issues or rushed service. Commission-based roles like car sales or assembly line work have long paid this way: the more units sold or produced, the higher the pay[2]. In each case, the measured target becomes the job, even if it stops matching what matters most to the business or customer.
The research summarized in [2] shows that large performance-based bonuses work as intended for simple, mechanical tasks—counting units sold, for example. But as soon as the work asks for judgment or problem-solving, high-stakes rewards can backfire. The attention shifts from doing the whole job well to "hitting the number," and overall performance can drop.
You see risk of metric gaming when:
- The chosen metric covers only one part of the work (solved tickets, not customer satisfaction).
- It's possible to meet the target without improving the outcome (logging a sale that later gets cancelled).
- The job has multiple important elements, but pay focuses on just one.
No source gives a real-world case study of metric gaming in pay-for-performance; what the sources do show is a pattern: when the metric is not the mission, staff will optimize for the number, not the outcome[2].
Questions readers ask
When does peer review work better than bonuses?
Peer review is most effective in roles where quality, safety, or professional standards matter and performance is hard to measure by numbers alone. Bonuses work best for simple, repetitive tasks where output is easy to count and quality is not easily compromised[2, 3].
Can a mix of bonuses and peer review help?
A combination can work: use moderate incentives for measurable, routine work, while relying on peer review for complex tasks and quality assurance. The key is to align each system with the nature of the work, not just apply both everywhere.
How can I avoid gaming of metrics in incentive pay?
Use multiple measures (not just one), review outcomes periodically, and supplement with qualitative feedback or peer review. Make sure the metrics reflect the real goals of the business, not just what’s easy to count[2, 3].