The Tragedy of the Engineering Commons
The surest way to ruin an engineering team is to believe it will survive whatever you do to it.
Modern tech management runs on a belief that anything legally permitted is operationally harmless. An engineering team is listed as a resource in the org chart, which is the source of most of the trouble that follows.
A team I worked with shipped the initial version of a platform that unlocked nine figures in annual revenue. The sane response would have been to protect it. Leadership raided it instead, redistributing engineers across other priorities and selling the move as temporary, which is the surest way to make something permanent. The remaining members spent the next five years maintaining a legacy system everyone agreed should be retired, because the resources to retire it were always assigned to the next quarter. The maintenance cost roughly three full-time engineers across half a decade, the same budget that would have replaced the legacy stack and migrated everyone onto the new platform. What presented as savings was a deferred bill that came due with interest.
When the absurdity was pointed out, the response was to contain it. The decision had arrived in a routine team planning meeting through a surprise cameo from leadership, who altered the agenda on the fly and presented a proposal already made elsewhere as if it were still open for debate. One person who objected was socially demoted. Another was moved off the team in a way that was technically voluntary and practically compulsory. The people responsible for the nine-figure platform were quietly recast as problems to be managed rather than people to be promoted, while those who had backed the raid, or had been wise enough to stay silent during it, received a modicum of influence on probation.
Disagreement is encouraged in theory. In practice, the prohibition on disagreeing in public applies in one direction only. Senior people dismiss proposals without scrutiny, and when critique flows upward it becomes an attitude problem. Over time, the upper echelons grow more certain and less aware. Psychological safety and radical candour reveal what they actually mean: speak up when you think our ideas are as brilliant as we do, and otherwise keep it to yourself. Independent thinkers learn to ration their honesty, which is the moment the organization begins selecting for compliance and calls it culture.
The commons mechanism is Hardin’s, transposed onto the org chart: a shared resource without a steward will be harvested until it collapses. Leadership reaps the immediate gain from the shared resource while the costs diffuse across everyone else and arrive later as failed migrations, brittle systems, and people quietly checking out. Influence at the director and VP level is measured less in outcomes than in how many people sit under one’s name on the org chart, and the same pressure travels downwards: ambitious managers a layer below offer up engineers as tribute to whatever their seniors currently count as a priority. The leader gets the appearance of decisiveness, the manager expands their domain, and the cost is borne by the team whose cohesion, context, and capacity get carved up to fund someone else’s ambition.
The whole arrangement rests on the fiction of fungibility, the conviction that engineers are interchangeable units movable without loss. Boris Groysberg’s work on portability established two decades ago that performance is entangled with the environment that sustains it, and that moving people without their team and context drops output sharply. Success is not a suitcase you can wheel between projects. The raid pays for this fiction twice: once in the productivity missing from the receiving team, and again in the productivity lost from the team you raided, which includes the months the displaced engineer spends rebuilding the context they used to have for free.
The deeper damage is to the engineers who watch a team get raided. They learn that loyalty to a team never pays, expertise offers no protection, and visibility is worth more than competence. Some detach and operate as mercenaries, others adapt and become political actors, and a few leave for greener pastures before the ecology has finished collapsing. What remains is an organization that breeds opportunists rather than stewards, which then looks around years later and wonders, with apparent sincerity, why so few people seem willing to own anything.
The pattern spreads. When a single product manager is split across so many teams that leadership begins to treat the role as fractional, the same logic produces the homeopathic PM, a diluted placebo until the product manager manages no products at all. Engineering and product management are only the most visible of the company’s ecologies. Anywhere there is a shared resource without a steward, the same harvesting pattern will eventually appear.
Companies behave predictably within their incentives, raid the teams they need most, and keep expecting the ecology to flourish anyway. Once the commons is barren, nothing grows back, and the people who could have told you why have already left.