# After the promotion: who keeps the manager competent?

> The Peter Principle explains how promotions misplace people. What comes next: can the organisation correct it, or does the team prop the manager up?

Source: https://sheng.page/en/posts/after-promotion-who-keeps-the-manager-competent/ · Published: 2026-11-17 · Author: Sheng

The Peter Principle describes how people can be promoted beyond their competence, and this essay asks whether the organisation can correct the mismatch once it has happened. Without a way to change roles, a team can end up maintaining the appearance that its manager still fits the job, while the problem holding up the work gets harder to fix.

Correction (17 November 2026, 23:42 Taipei): the closing section now says engineers' concerns have a chance of influencing the plan once someone owns the decision, matching the Chinese version. Suggested by Codex (gpt-6-astra), checked by Claude.

The starting point was Mandy Brown's "Petered" and the Hacker News discussion. My argument is that when a title exceeds someone's ability to do the job, appearances can fill the gap at the expense of the people reporting to them.

## A meeting that solves nothing

Imagine an engineering team's weekly meeting. The status report gets more complete every week, with a Gantt chart and a filled-in risk column, and the meeting moves from weekly to twice a week, along with a cross-team sync. The project has stalled on one architecture decision, whether to break up the old billing module, which comes up in the last five minutes every time and is held over to the next week.

Engineers spend half a day before each meeting pulling figures together, then carry the same open question back to their desks and code around it. Nobody is slacking, and their time goes into showing progress.

A meeting is useful when the decision gets made and someone takes responsibility for it, which the amount of material prepared does not by itself show.

## Promotion puts two decisions in one

A promotion rewards past performance and picks someone for a different job at the same time. Judging only the current job makes it easy to miss what the next one needs. Laurence J. Peter and Raymond Hull named the tendency in their 1969 book *The Peter Principle*, which was written as satire.

Alan Benson, Danielle Li and Kelly Shue studied promotions among salespeople. In their 2018 NBER working paper and its summary, covering 214 firms, doubling sales was associated with a 14.3% higher probability of promotion relative to the baseline, and doubling a new manager's pre-promotion sales was associated with 7.5% lower sales performance for each subordinate. People with prior collaboration experience performed better as managers on average. These are sales data, so the figures don't carry straight over to engineering teams.

Writing code doesn't automatically prepare someone to coordinate work or give feedback. In "Becoming the Boss" (2007), Linda A. Hill reviews her initial study following 19 new managers for a year and her later research, and notes that the great majority of new managers learn to function in the role, though the study was not specifically of engineering managers. New managers need time to learn, and if the gap persists a role change is worth considering.

## From solving problems to proving competence

When a role asks for more than the person can do, and there's no room to learn or adjust, effort can drift towards results that are easier to show: more slide decks, or a larger rewrite.

Zach Lloyd describes how, when he led Google Sheets, small bug fixes lost out to projects that looked better to the promotion committee. Matheus Lima recounts adding abstractions to avoid repeating a few lines, leaving the next engineer half a day's work just to understand them. Both are about promotion incentives. They don't show that a mismatch between a manager and their role caused these problems.

Suppose a rewrite earns its initiator credit in a promotion case, then stops halfway, leaving the original team running two systems.

Security reviews, audits, tests and necessary coordination have a purpose, judged by the risks a process covers and the work it saves.

## What the team picks up

Back at that weekly meeting, engineers keep documenting the same technical options while waiting for a decision. They prepare the coordination notes the manager should have written and deal with the unresolved problems. If a concern is only taken seriously once it's presented as the manager's idea, people may eventually stop raising it.

## Knowing it doesn't fit, and why there's no way back

A title comes with pay and future opportunities. If returning to engineering means a pay cut or leaving the company, the manager may stay put, and the team absorbs the cost.

In "The Engineer/Manager Pendulum" (2017), Charity Majors argued that management is a change of profession. She quoted Sarah Mei describing a lateral move onto a parallel track. In 2020, Majors acknowledged that her slogan "management is not a promotion" was more aspirational than accurate.

GitLab's engineering handbook says someone who tries management and finds it unsuitable can return to the IC track. It treats Staff Engineer and Engineering Manager as equivalent for base pay and prestige. Its transfer guidelines, marked as under revision, say pay may fall when moving to a lower grade. Dropbox's promotion guidelines and Buffer's 2019 framework describe parallel IC and manager paths. Those pages don't specify arrangements for returning to IC work.

In the Hacker News thread, [donatj](https://news.ycombinator.com/item?id=50413611) reports leaving after about a year as a manager to return to development, calling it a demotion. [dfxm12](https://news.ycombinator.com/item?id=50413883) reports similar pay for tech managers and people managers in their organisation, though some senior tech managers also have direct reports. Both are commenters' self-reports.

Can an engineer who doesn't manage keep progressing in level and pay? If someone tries managing for six months and wants to return, they need to know the terms and who will take over their projects. When the person carries the whole cost of returning, a dual ladder can still leave them in a role that doesn't suit them.

## Back to the meeting

Back at the meeting, engineers' concerns have a chance of influencing the plan once someone takes responsibility for the billing module decision. A manager better suited to writing code also needs a route back, so the team doesn't have to maintain the appearance that the role still fits.

Are we helping people do the work well, or making the work prove that someone deserves the position?

## Notes

- The Peter Principle describes the mismatch; the question here is whether an organisation can correct it once it has happened.
- Benson, Li and Shue's 214 firms, 14.3% and 7.5% come from the 2018 NBER working paper and are about salespeople.
- Hill draws on an initial study of 19 new managers and subsequent research to describe how most learn the role.
- GitLab describes a return to the IC track, though its transfer guidelines allow lower pay at a lower grade; the Dropbox and Buffer pages describe parallel paths without specifying return arrangements.

## Further reading

- Wikipedia, "[Peter principle](https://en.wikipedia.org/wiki/Peter_principle)": encyclopaedia entry on the origin and background of the idea.
- Mandy Brown, "[Petered](https://aworkinglibrary.com/writing/petered)" (3 November 2026): the starting point for this post, a commentary on a personal blog.
- Hacker News, "[Petered](https://news.ycombinator.com/item?id=50412152)" (8 November 2026): a public forum thread, where comments are personal views and self-reports.
- Benson, Li and Shue, "[Promotions and the Peter Principle](https://www.nber.org/papers/w24343)" (NBER Working Paper 24343, February 2018): academic research. The version in *The Quarterly Journal of Economics* (2019, vol. 134, no. 4) reports 131 firms; the figures here are from the working paper.
- NBER Digest, "[The Peter Principle Isn't Just Real, It's Costly](https://www.nber.org/digest/may18/peter-principle-isnt-just-real-its-costly)" (April 2018): NBER's research summary, with the 14.3% and 7.5% figures and the sample of more than 1,500 promoted managers.
- Linda A. Hill, "[Becoming the Boss](https://hbr.org/2007/01/becoming-the-boss)" (*Harvard Business Review*, January 2007): a researcher's review of her work on first-time managers.
- Zach Lloyd, "[The Google Incentive Mismatch: Problems with Promotion-Oriented Cultures](https://www.warp.dev/blog/problems-with-promotion-oriented-cultures)" (2022): a named author's first-hand account of promotion incentives on Google Sheets.
- Matheus Lima, "[Nobody Gets Promoted for Simplicity](https://terriblesoftware.org/2026/03/03/nobody-gets-promoted-for-simplicity/)" (2026): a personal blog post that includes the author's own over-abstraction.
- Charity Majors, "[The Engineer/Manager Pendulum](https://charity.wtf/p/the-engineer-manager-pendulum)" (2017): personal career commentary; the parallel-track line is Sarah Mei's.
- Charity Majors, "[If Management Isn't A Promotion, Then Engineering Isn't A Demotion](https://charity.wtf/2020/09/06/if-management-isnt-a-promotion-then-engineering-isnt-a-demotion/)" (6 September 2020): personal blog post where she calls her own slogan more aspirational than accurate.
- GitLab, "[Engineering Career Development](https://handbook.gitlab.com/handbook/engineering/careers/)": the company's official handbook, covering the return to the IC track and Staff versus Engineering Manager pay.
- GitLab, "[Promotions and Transfers](https://handbook.gitlab.com/handbook/people-group/promotions-transfers/#transfer-compensation-guidelines)": the company's official transfer guidelines, marked as under revision, including pay on moves to a lower grade.
- Dropbox, "[Promotion guidelines](https://dropbox.github.io/dbx-career-framework/promotion_guidelines.html)": the company's published career framework, with separate IC and manager columns in its level table.
- Buffer, "[Career framework](https://buffer.com/resources/career-framework/)" (27 January 2019): a company blog post describing its framework at the time, with separate paths for makers and managers.

> Ideas and technical judgement by Sheng; drafted with Claude · Sources checked on 10 November 2026 (Taipei time).