The honest view from inside technical leadership
Analysis

Leadership

Field notes on how engineering leaders keep control of systems as AI changes how software gets built.

Leadership · Response

Unwritten Promotion Criteria Reward the Engineers Who Describe Their Work Best

Unwritten promotion criteria make self-description a hidden input. Equal performers rate their own work differently, and accountability narrows reward gaps.

Line-art illustration of a career ladder whose upper rungs are drawn as dashed, unwritten lines next to a written checklist, representing unwritten promotion criteria for engineers

Unwritten promotion criteria reward the engineers who are best at describing their own work, and that is a different skill from doing it. Muralidharan Lakshmanan argues in a recent article that managers should write down the career path nobody spells out. This piece follows that one claim into the research on how people rate themselves and how managers turn ratings into rewards.

Lakshmanan argues that engineers arrive carrying a career map from somewhere else

Lakshmanan leads four engineering teams with more than 60 people, a mix of consultants, internal transfers, external hires, engineers from other industries, graduates and interns. His observation is that each person brings a definition of success learned in a previous setting, and that the definition often fails in the new one. He describes his own move from consulting into a full-time role at a Fortune 500 fintech company, where nobody explained that ownership no longer ended with the project, and says it took him years to work that out alone.

His advice follows from that. Managers should connect each team’s work to business priorities so engineers do not have to infer what leadership values, coach each person according to what that person needs, and ask newcomers where they learned their idea of success before explaining how growth works on the team. He lays out four phases of growth, ending in a stepping-up phase where evidence of working at the next level grounds the promotion conversation. He also notes that many engineers believe “their smart work should speak for itself,” while leaders outside the team may never see that work unless someone tells them.

Unwritten promotion criteria turn self-description into part of the evaluation

When the criteria for the next level live in a manager’s head, the evidence that reaches the decision is mostly the evidence the engineer chose to surface. Lakshmanan’s own examples show the mechanism. The engineer who froze in meetings was doing excellent technical work and became visible to the wider company only after his manager built a staged path into presentations, one slide at a time. In another passage, Lakshmanan observes that a minor glitch fix can win executive attention while a far more complex feature gets little. In both cases the outcome moved with how the work was described, and to whom.

That makes self-description a hidden input to every promotion decision. A leader who leaves the criteria unwritten is also choosing, without saying so, to weight that input heavily. The question that follows is whether people who perform equally describe their work equally. The research says they do not.

Equal performers rate their own work differently, even with nothing to gain

Christine Exley and Judd Kessler tested this in The Gender Gap in Self-Promotion, an NBER working paper first released in October 2019. They ran experiments with more than 4,000 participants from online labor markets and more than 10,000 school-aged young people. Women described their ability and performance to potential employers less favorably than men who had performed equally well. The gap remained when every incentive to self-promote was removed, which led the authors to conclude that it reflects a difference in how people evaluate their own performance. It appeared as early as the sixth grade.

The second finding is the one that matters for an engineering manager. If the gap vanished once there was nothing to gain, teaching people to be more strategic about visibility might close it. Because it persists without incentives, any process that depends on self-reported impact will carry the gap into self-reviews, promotion packets and calibration meetings, however good the coaching is. The study measured gender. I expect the same mechanism to show up in the other differences Lakshmanan describes, such as a former consultant used to tightly scoped work or an engineer from another industry used to different norms, though that extension is my inference and the paper does not test it.

Calling a promotion process meritocratic does not protect it from bias

A common response is to state that the organization promotes on merit. Emilio Castilla and Stephen Benard tested what that statement does in The Paradox of Meritocracy in Organizations, published in Administrative Science Quarterly in December 2010. Across three experiments, 445 participants with managerial experience made bonus, promotion and termination recommendations for a set of employee profiles. When the company’s core values were presented as explicitly meritocratic, participants gave a male employee a larger monetary reward than an equally qualified female employee.

A team that says it rewards impact, and leaves impact undefined, has adopted the label the study tested without any structure to counter it. Lakshmanan’s advice to separate personal aspiration from company policy, and to explain what real evidence of readiness looks like for the next level, supplies that structure. Written criteria give a manager something to check a decision against other than a general sense of who seems strong.

Written promotion criteria work when someone has to account for outcomes

Writing the criteria down is necessary, and the evidence suggests it is insufficient by itself. Castilla’s later field study, Accounting for the Gap, published in Organization Science in 2015, followed performance-based reward decisions for almost 9,000 employees at a large private company before and after senior management introduced accountability and transparency procedures. Before the change, women, ethnic minorities and employees born outside the US received lower monetary rewards than US-born white men with the same performance scores, in the same job and work unit, under the same manager. After the procedures were introduced, that gap shrank.

The study covers performance-based pay, and a promotion is a different decision. The mechanism still transfers: reward decisions became visible to people who could ask why. For an engineering organization, the written ladder Lakshmanan describes is the transparency half. The accountability half is a review after each cycle in which someone looks at who was put forward, on what evidence, and asks about the people who were left out.

There is a boundary condition here too. A written ladder can harden into a checklist that rewards whatever activities it lists, and a promotion packet is still written by the engineer it describes. So the manager also has to gather evidence directly, from incident reviews, design documents and peers, so that the packet does not rest on self-description alone. Lakshmanan’s slide-by-slide handover builds the skill of describing work. Direct evidence collection keeps the outcome from depending on that skill.

Why This Matters for Engineering Leadership

Lakshmanan’s teams draw on at least six different routes into engineering, and he reports that the same question about growth needs a different conversation for each. That variety is the condition in which unwritten promotion criteria do the most damage, because each group arrives with its own default for how loudly to describe its work. Three steps follow from the research above. Publish the criteria for each level with examples of acceptable evidence. Have managers collect evidence of impact themselves instead of relying on self-assessment. Review promotion and reward outcomes by group after every cycle, with a named person accountable for explaining any gap. The first step is what Lakshmanan asks managers to do. The other two are what the research suggests it takes for that first step to hold.

FAQ

What are unwritten promotion criteria?

They are the expectations for the next level that live in managers’ heads or team habits and were never documented, such as which kinds of work leadership tends to notice.

Why do unwritten promotion criteria create bias?

They make the decision depend on how engineers describe their own work. Research by Christine Exley and Judd Kessler found that equally performing people describe their performance differently, even with no incentive to do so.

Does declaring a commitment to meritocracy reduce bias?

Not on its own. In Castilla and Benard’s 2010 experiments, managers evaluating a company presented as meritocratic gave a man a larger reward than an equally qualified woman.

What reduces gaps in performance-based rewards?

In Emilio Castilla’s 2015 field study of almost 9,000 employees, adding accountability and transparency procedures to reward decisions reduced pay gaps by gender, race and nationality.

Is coaching engineers to be more visible enough?

No. Coaching builds a real skill, and managers should also gather evidence of impact directly so promotion outcomes do not rest on self-description alone.

Source: How to coach engineers from different backgrounds by Muralidharan Lakshmanan, LeadDev, 19 November 2026.

Read more news & analysis