The capability trap: why organizations reward firefighting over prevention

This 2001 California Management Review paper by MIT’s Nelson Repenning and John Sterman explains why process-improvement programs so often fail — and the answer has little to do with the techniques themselves. Managers facing a performance gap can close it two ways: working harder (more hours, more pressure) or working smarter (investing in the capabilities, tooling, and maintenance that prevent problems). Working harder pays off immediately; working smarter pays off after a delay, and often makes things worse before they get better. Under deadline pressure, organizations systematically choose the first option, cannibalizing improvement time to hit short-term targets.

The result is what the authors call the capability trap. Skipped maintenance and deferred improvement erode capability, which lowers performance, which increases pressure, which drives further cuts to improvement work — a self-reinforcing spiral into permanent firefighting. Worse, managers tend to misattribute the decline to worker attitudes or lack of discipline rather than to the system they created, a self-confirming error that justifies even more pressure. Meanwhile the engineer who prevents a crisis produces no visible event to reward, while the one who heroically fixes a visible failure gets promoted — hence the title.

Escaping the trap requires leadership willing to absorb the worse-before-better period: protecting improvement investment even as short-term metrics dip, as illustrated by case studies including Du Pont’s preventive-maintenance program and Harley-Davidson’s product development. Though written about manufacturing, the paper has become a touchstone in software circles, where it maps cleanly onto technical debt, SRE toil, and why reliability work is chronically underfunded until the outage happens.