Nobody gets credit for fixing problems that never happened: why improvement programmes fail

Why do most process improvement programmes fail even when the tools behind them are sound? In a 2001 article in California Management Review, MIT researchers Nelson P. Repenning and John D. Sterman argued that the problem lies not in the tool but in the system around it — the interaction of tools, equipment, workers and managers. Their paper is titled “Nobody Ever Gets Credit for Fixing Problems that Never Happened: Creating and Sustaining Process Improvement”.

The improvement paradox

The authors open with a puzzle. U.S. companies spent more than $100 billion on management consultants and training in 1997, yet few improvement efforts produced significant results. Total Quality Management was the clearest case: careful studies showed that firms seriously committed to TQM outperformed their competitors, but fewer than 10% of the Fortune 1000 had well-developed TQM programmes, and its use fell from third most popular business tool in 1993 to fourteenth in 1999. “You can't buy a turnkey six-sigma quality program,” the authors write. “It must be developed from within.”

The capability trap

Their model treats performance as the product of time spent working and the capability of the process. Capability is a stock: it accumulates through improvement work, which pays off only after a delay, and it erodes continuously if nobody maintains it. Because improvement is slower and less visible than simply working harder, organisations drift into a vicious cycle the authors call the capability trap: pressure to hit targets cuts the time available for improvement, capability declines, defects and rework grow, and firefighting consumes the hours that could have gone into eliminating root causes.

Why the trap persists

The trap persists because the lessons people draw from it are wrong. Managers attribute poor performance to unmotivated workers, engineers hide defects to avoid blame — in one firm the weekly progress review was known as the “liars' club” — and management answers with more monitoring. Organisations end up rewarding those who rescue crises and overlooking those who prevent them. As an engineer at an auto company told the researchers: “Nobody ever gets credit for fixing problems that never happened.”

What worked

The paper's cases show a way out. At Du Pont, a maintenance programme built around a role-playing “Manufacturing Game” taught teams to expect conditions to get worse before they got better: at plants that adopted it by the end of 1993, pump reliability rose 12% for every doubling of cumulative experience and direct maintenance costs fell 20%, while at 23 comparable plants that did not adopt it costs rose 7%. At BP's Lima refinery, maintenance costs ballooned 30% in the first six months of the programme, after which pump mean time between failures climbed from 12 to 58 months; the authors put the new value created at $43 million a year against $320,000 in costs.