Nidec Replaces Its CEO and Tells Managers They Can Stop Shipments
On September 29 Nidec’s board replaced president and chief executive Mitsuya Kishida. Nidec is a Japanese manufacturer of electric motors for hard drives, home appliances, cars and industrial machinery.
According to the remedial measures set out this month, the company will revise its global quality assurance regulations “to inform all concerned” that quality assurance managers at each business unit have the authority to suspend shipment. The measures came after an outside committee confirmed 844 cases of quality misconduct across the group.
What makes this noteworthy: managers already had this authority for years, and it went largely unused. The committee’s own report states that the quality assurance function is independent of the business departments in principle, and responsible for suspending manufacturing and shipping as necessary.
Of the 844 cases, 805 were changes to a material, a machine, a process or a method, made without the contractually required customer approval. A forensic review of email from current and former officers found no instruction from above to make any of those changes. These were local calls, made to keep things moving.
As a driving factor for this behavior, the committee found that business departments set their target figures based on their markets and their own capacity, while the group’s need for certain profit levels produced separate guidelines. The business departments were then “expected to bridge the gap between these figures.” When they said the targets were difficult to meet, such feedback was not adequately reflected.
Departments expected to fall short were managed through frequent meetings that sometimes ran late into the night. In the committee’s survey of employees, internal pressure on deadlines, sales, profit and cost came back as a cause of the misconduct.
Even though quality assurance managers theoretically had the authority to stop shipments, they sometimes reported to the site manager, who would risk missing their quota if a shipment stopped. The report lists what followed: raising a problem meant halted production, scrapped stock, a call to the customer, a loss on the books and severe reprimand from the executives running performance management. Whoever raised the problem was also expected to resolve it.
When the Written Rule Stops Matching the Work
In 1994 American fighter jets over northern Iraq mistook two American helicopters for enemy aircraft and shot them down, killing everyone aboard. Scott Snook, who studied that accident, attributed some of the failure to what he calls “practical drift”: the slow, steady uncoupling of local practice from written procedure.
Where a procedure does not fit the local task, people adapt around it. Repetition then turns the workaround into “how work is done around here.” The procedure stays on the shelf, correct and unread, and the practice that replaced it becomes a weak link in the organization’s decision system.
Nidec’s target-setting shaped the numbers in both directions. Bad news that costs the bearer tends to arrive late, in a smaller version, or not at all, and reporting a shortfall carries a cost. The head office therefore kept receiving inflated numbers, and set the following year’s targets from them.
Steps to Remediate Practical Drift
So what can you do to minimize drift in your own organization’s decision system?
Take three policies your organization leans on (for example authority to delay the release date of a project). What would enforcing them cost the person responsible: what is the downstream impact, who do they report to, and who is left owning the fix? If the same person turns up for all three policies, you are at risk of practical drift.
Codify the pass/fail criteria. Move the call away from the individual’s judgment to an objective score (for example a measured tolerance a component either meets or does not meet). With a written criterion, raising a failure means reporting a measurement instead of challenging a target. Keep the score out of performance records. Give the remediation task to somebody other than the person who raises the problem, so that raising one does not add the work of fixing it.
Audit usage. An authority nobody uses is either unnecessary or unaffordable. For each stop authority, ask how often it was exercised over the last two years, then read that against how often the condition it exists for has plausibly occurred (for example the number of lots that failed an incoming inspection). Where the gap is wide, the people who hold the authority can explain it, and that explanation is worth weighing against what a stoppage would have cost them.
Assess reliability of performance targets. Where the people who create delivery targets are then measured against them, trace the numbers back to their source. If you find that the only source is a report from the unit being measured (for example the team’s own forecast of next quarter’s yield), treat those numbers as a best case rather than a probable scenario. The team that supplies the forecast is also the team held to it, so the number arrives with that pressure already in it.
Decouple use of authority from consequences. Start with the most load-bearing policy (for example the one that could delay a customer-bound shipment), then find the person who used it most recently, or who considered using it but didn’t. Ask what doing so did or would have cost them, and check whether their own incentives could skew that answer.
How to Optimize Your Organization’s Decision System
The policies, incentives and structures that form your organization’s decision system shape your bottom line alongside the people in it. A capable manager usually loses to a weak system, which is why the same people can decide well in one organization and poorly in another. That is also why the pattern tends to outlast any one person.
To learn more about how your organization’s decision system performs, and how to improve it, visit decision-maturity.com.