How it varies by context
At facilities of different ages, variation often traces to accumulated informal modification — small process adjustments made over years to work around equipment quirks specific to that site, none individually significant, collectively producing a facility that no longer runs quite the same process as its sister site despite identical original specifications. At facilities with different workforce tenure, variation more often traces to how much of the operation runs on codified procedure versus experienced judgement — a newer, less tenured workforce follows the written procedure more literally, for better and worse, while a longer-tenured workforce has developed adaptations that may outperform the document in some conditions and underperform it in others, in ways the standard reporting cannot distinguish. At facilities running different product mixes even under a shared corporate standard, variation often reflects that the standard itself was calibrated to one product’s typical demands and fits another’s less well, so a facility performing “worse” against the shared metric may actually be managing a harder underlying task the metric was never designed to weight fairly.
A further context is the facility operating multiple shifts. Variation between shifts at the same facility, run by the same management team under the same written standard, is common enough that it is worth treating as its own category rather than folding it into either the age or tenure explanation — a night shift with a smaller supervisory presence and a different mix of experience levels can produce genuinely different outcomes from a day shift at the identical site, which means facility-level averages can mask as much internal variation as they reveal between facilities.
A fifth context, easy to overlook because it sits outside the facility itself, is the local labour market each site draws from. Two facilities running identical processes can face very different levels of staff turnover depending on local employment conditions, and turnover has a direct, measurable relationship to how much of a facility’s operation runs on codified procedure versus on the kind of tacit, experience-built judgement that takes time to accumulate and disappears when someone leaves. A facility in a tight local labour market with high turnover may simply never accumulate the same depth of undocumented expertise that a facility in a more stable labour market builds up over years, regardless of how identical the two sites were on the day they opened. This is worth stating plainly because it is the context most likely to be dismissed as outside a facility manager’s control and therefore left out of the performance conversation entirely, even though it may be doing more to explain the gap than anything within that manager’s actual authority.
What the measurement contributes
V. F. Ridgway’s research on the dysfunctional consequences of performance measurement is directly relevant to why this variation persists rather than being quickly diagnosed. Ridgway examined cases including public employment interviewers evaluated on the number of interviews conducted, who responded by conducting faster interviews and placing fewer applicants — and law-enforcement investigators given a monthly case quota, who responded by selecting easier cases as the month-end deadline approached. In each case, a reasonable-looking measure produced behaviour that undermined the outcome the measure was meant to track. Ridgway’s own summary is direct: “Quantitative measures of performance are tools, and are undoubtedly useful. But research indicates that indiscriminate use and undue confidence and reliance in them result from insufficient knowledge of the full effects and consequences.”
Steven Kerr’s complementary research on reward systems adds a further layer: organisations frequently reward a measurable proxy while hoping for a different, harder-to-measure outcome, and behaviour tends to follow the reward rather than the hope. A manufacturing facility measured primarily on throughput, when the genuinely desired outcome is throughput-with-quality, will tend to produce more of what is measured — which can look, in cross-facility comparison, like one site simply performing better, when it may instead be optimising more precisely for the specific number being compared.
Taken together, Ridgway’s and Kerr’s findings suggest that a portion of any facility-to-facility comparison is, in effect, a comparison of which facility has best learned to satisfy the metric rather than a comparison of which facility is genuinely producing the better underlying outcome — and distinguishing the two requires looking past the shared metric itself toward what each facility may be doing differently to hit it.
This has a practical implication worth stating directly: a facility leadership team eager to close the gap with a better-performing sister site should be cautious about simply copying that site’s visible practices without first checking whether its strong number reflects a genuinely better process or a more favourable version of the same measurement dynamic — adopting another site’s specific tactics for hitting a distorted metric transfers the distortion along with whatever genuine improvement comes with it.
Where this shows up
This perspective is written with manufacturers in mind. It sits within the Asset and Production-Based grouping, and the questions it raises recur in other sectors too.
Questions worth considering
- Which of the current cross-facility comparisons assume the facilities are running an identical process, and is that assumption still accurate?
- Where might a facility be “underperforming” against a shared metric while actually managing a genuinely harder task the metric does not weight?
- What behaviour would change at each facility if the measured proxy and the genuinely desired outcome were pulled further apart?
- How much variation exists between shifts within the same facility, compared with the variation being discussed between facilities?
Your team is facing similar challenges?
Subscribe to The Riddle Ledger
Perspectives are sent as they are published, with no promotional email, and you can unsubscribe at any time.