The Peter Principle: Why People Rise to Their Level of Incompetence

By Tara Minh · Updated Oct 10, 2026

The Peter principle is the idea that people in a hierarchy get promoted for doing their current job well, until they land in a role they can't do well. Laurence J. Peter and Raymond Hull coined it in 1969 as satire. Later research on sales promotions shows a real, measurable version of the problem.

The principle matters because most companies promote the same way: they reward the best performer with the next job up. That works when the two jobs need similar skills. It fails when they don't, and the jump from doing the work to managing the people who do it is the classic example. This article covers where the principle came from, what the data says, how it differs from the Dilbert principle and Parkinson's law, and five ways to prevent it.

Key Facts

Key Facts

• Laurence J. Peter and Raymond Hull published The Peter Principle: Why Things Always Go Wrong in 1969. It was written as satire, and its central claim is that every employee tends to rise to a level of incompetence.
• Benson, Li and Shue's "Promotions and the Peter Principle" (Quarterly Journal of Economics, 2019) used microdata on the performance of sales workers at 131 firms and found evidence consistent with the principle.
• In the earlier NBER working paper, doubling a new manager's pre-promotion sales corresponded to a 7.5 percent decline in each subordinate's sales.
• The same working paper estimated that average managerial quality would be 30 percent higher if firms promoted on predicted managerial potential instead of sales, a figure the authors treat as a lower bound on the incentive benefits that would justify current policy.
• A Gallup analysis found that managers account for at least 70 percent of the variance in employee engagement scores, so a mismatched promotion affects the whole team.

What is the Peter principle?

The Peter principle says that in a hierarchy, employees keep getting promoted as long as they perform well, and they stop being promoted only when they reach a position where they don't. At that point they stay. The result, in Peter's telling, is that over time each post ends up held by someone who can't do it well, while the real work gets done by people who haven't been promoted that far yet. CBC's account of the book describes both ideas.

Two assumptions sit underneath it:

1. Promotion rewards performance in the current role. The best nurse, engineer or sales rep gets the next job up.
2. Skill in one role doesn't guarantee skill in the next. A great individual contributor might have no aptitude for coaching, prioritizing across a team, or handling conflict.

When both are true, the system quietly selects for mismatch. Nobody makes an obviously bad decision. Each promotion looks reasonable on its own, and the damage shows up a year later as a team that's stuck.

Where it came from

Laurence J. Peter was a Canadian educator who began noticing, as a young teacher, that some of the teachers, principals and superintendents around him seemed unaware of their responsibilities. According to the CBC piece, the book grew out of a conversation between Peter and the writer Raymond Hull in a Vancouver theatre lobby in the early to mid 1960s. It was turned down by more than a dozen publishers, then became a best-seller on release in 1969.

Peter and Hull wrote it as a joke with a sting, not as a research paper. They had no dataset. What they had was an observation most people recognize from their own workplaces: the person who was brilliant at the old job and struggling at the new one. The principle stuck because the observation is so common.

What the research says

For decades the Peter principle was folklore. Two kinds of studies have tested it since.

The sales-to-manager evidence

The most cited test is Alan Benson, Danielle Li and Kelly Shue's paper in the Quarterly Journal of Economics. The published abstract says they used microdata on the performance of sales workers at 131 firms and found evidence consistent with the Peter principle: firms put heavy weight on current job performance when deciding who to promote, at the expense of other observable traits that better predict how someone will do as a manager.

Sales is a good laboratory because performance is easy to count. The earlier NBER working paper is the version that reports the underlying numbers: a panel of 53,035 sales workers at 214 US firms from 2005 to 2011, of whom 1,531 were promoted into management. The figures below come from that version, and the firm count differs from the journal abstract.

Read the second row carefully. It doesn't say top sellers become terrible managers. It says that among people who were promoted, the ones with higher sales before promotion tended to have subordinates who sold less afterward. The authors also note that a typical manager in their data supervises about five people, so even a modest percentage drop across a team adds up.

The paper also points to a trait that predicted good management better than raw sales: how many colleagues a person had shared sales credit with. People who had worked with more colleagues tended to make better managers. That's a hint about what to look for. Managing is mostly about working through other people, and a track record of collaborating is closer to that than a track record of closing deals alone.

Why firms do it anyway

The authors are careful not to call this plain stupidity. Promoting top sellers gives everyone else a reason to work hard, and it's easy to defend because the numbers are visible. The working paper says the 30 percent figure should be read as a lower bound on how large those incentive benefits would have to be to justify the current approach. The published abstract adds that firms put less weight on sales when the management role carries more responsibility, and when frontline pay is already tied closely to performance.

In other words, many companies know the trade-off. Shue told CBC that firms are aware that promoting the top sales worker may not produce the best manager, and do it anyway to motivate people.

The simulation evidence

A second line of work is a model, not data. Pluchino, Rapisarda and Garofalo built an agent-based simulation of a hierarchy where promotion rewards the best performers and where competence at one level doesn't carry over to the next. In that setup, they show the Peter principle is unavoidable and cuts the organization's overall efficiency. Their counterintuitive result was that promoting at random, or randomly promoting the best and the worst, did less harm than promoting the best. It's a model with strong assumptions, not a recommendation to draw names from a hat. But it makes the point that promoting on current performance isn't automatically the safe choice.

Peter principle vs Dilbert principle vs Parkinson's law

People often lump these three together. They describe different failures.

The Dilbert principle assumes managers were weak before they were promoted. The Peter principle assumes they were strong. That difference matters, because the fixes are opposite: you can't solve a Peter problem by promoting better performers, since better performance was the cause.

Why it happens

Three forces feed the pattern.

The skills change at each level. Robert Katz's skills theory of leadership describes technical, human and conceptual skills, and says the mix shifts as you move up. Technical skill carries most of the weight at the frontline, while human and conceptual skills carry more at the top. A promotion based on technical results rewards the skill that matters less in the new job.

The incentives point at the past. Promotion decisions are built on what's measurable, and what's measurable is what someone has already done. Potential for a different job is harder to see, so it gets less weight. This is the pattern the sales study documents.

Nobody retrains the person after the move. Managers who got the job because they were excellent at the old one are often handed a team and no preparation. The work of a manager differs from the work of the person managing, a distinction covered in leadership vs management, and the gap rarely closes on its own. Management skills like delegating, giving feedback and running a meeting are learned, not inherited from being good at the old job.

Signs the Peter principle is at work in your team

You rarely see it announced. You see symptoms:

• A newly promoted manager keeps doing the old job and takes over their team's best tasks. This often turns into micromanagement.
• Their best people leave or stop volunteering ideas within a year.
• The manager avoids hard conversations and delays decisions.
• Results are fine as long as the manager personally does the work, and fall when they can't.
• People mention that the company "lost its best engineer and gained an average manager."

One or two of these doesn't prove anything. A pattern across several promotions in the same function does.

How to prevent it: a step-by-step approach

None of these steps is expensive. Most are changes to how you decide and what you offer.

Step 1: Separate the question "did they perform?" from "can they do the next job?"

Write down both. Performance in the current role earns recognition, a raise or a bonus. The next-role question needs its own evidence. If you can't answer it, you have a recognition problem, not a promotion decision.

Step 2: Create a second ladder

A dual career ladder lets people advance in seniority, pay and title without taking on direct reports. Shue described this to CBC as a "dual track" in which a person can become a senior scientist and be recognized through title and compensation without being moved into a management function. Make the top of the individual-contributor track pay as well as the entry level of management, or people will still choose management for the money.

Step 3: Run a trial assignment

Before the promotion, give the person a defined stretch: lead a project with three or four people for a quarter, cover for a manager on leave, or mentor two new hires. You learn how they handle other people's work. They learn whether they like it. Both are reasons to do it.

Step 4: Promote on management potential, with guardrails

Look for the traits that predict good managing: collaboration, coaching behavior, comfort with ambiguity, interest in other people's growth. But be careful with "potential" ratings. In an example Shue gave CBC, a company using a nine-box grid showed women getting slightly higher performance ratings and sharply lower potential ratings. Potential is a judgment, and judgments carry bias. Require evidence for each rating, and compare ratings across groups.

Step 5: Make going back possible, and normal

The last link in the chain is what happens when the promotion doesn't work. Peter's original point was that people stay put at their level of incompetence. A company that allows a manager to return to an individual-contributor role without penalty, and says so up front, avoids that trap. Frame the trial period as a standard part of the move, agree on the return path before the promotion, and keep the person's pay conversation separate from the title.

A promotion-readiness scorecard you can copy

Use this table for every promotion into management. Fill it in with evidence, not impressions, and score each row from 1 to 5.

Anyone scoring low on the middle rows needs a trial assignment or a development plan before the promotion, not after.

Examples by business function

These are illustrations of the pattern, not case studies.

• Sales. A rep who closes 140 percent of quota becomes sales manager. She's used to winning deals by working harder, and now has to coach reps who need a different approach. This is the case the sales research above examines.
• Engineering. A staff engineer with deep knowledge of the codebase becomes engineering manager. He keeps reviewing every pull request himself, which slows the team. A dual track would have kept him building and let a different person lead.
• Customer support. The agent with the best satisfaction scores becomes team lead. Her skill was calm one-on-one empathy. The new job is scheduling, quality review and handling escalations from people she used to work beside.
• Finance. The most accurate analyst becomes controller. Accuracy still matters, but now the job is deadlines, negotiating with other departments, and managing people who are less detail-oriented than she is.
• Operations. A shift supervisor who keeps the line running becomes plant manager. The conceptual work of planning capacity and budgets is a different skill from fixing problems in the moment.

In each case the original success was real, and it's the reason the promotion looked safe.

Do's and don'ts

Do:

• Promote into management only people who've shown some interest in developing others.
• Offer a senior individual-contributor path with real status and pay.
• Give new managers training in the first 90 days, not the first 12 months.
• Agree on a review point and a return path before the move.
• Hold a succession conversation with potential candidates early, using the approach in succession planning.

Don't:

• Treat a promotion as the only reward you have.
• Use "she's our best rep" as the sole reason.
• Move someone into management as a retention fix without checking they want it.
• Treat a manager's struggle as a character flaw. It's usually a fit and support problem.
• Punish a person who steps back.

Limits of the Peter principle

The principle is useful, but not universal.

• It was satire. Treat it as a lens, not a law. Many people do succeed after promotion, and most organizations also learn and adjust.
• The data are about sales. The strongest evidence comes from sales teams, where performance is cleanly measured. Other jobs may behave differently.
• Incentives are a real counterweight. Promotion is a motivator. A policy of never promoting top performers could hurt effort, which is why the researchers don't say firms are simply making mistakes.
• Potential is hard to measure. The cure of promoting on potential can introduce bias, as the nine-box example shows.
• Not everyone stalls. Some people grow into the job, and a good development plan helps more of them do it.