Skip to content

Cross-Functional Incentive Alignment

Definition

A shared customer-facing metric fails to change behavior wherever it crosses into a function that is bonused on its own local numbers instead — because that function's specialists have good, defensible reasons (fraud losses avoided, cost per call reduced) to keep doing what is hurting the shared outcome. The fix the book identifies is not persuasion or education alone but restructuring incentive compensation so every functional leader's bonus depends in part on the shared metric, converting a rational local optimum into a rational shared one.

In the Book

Chapter 9's central case is an anonymized U.S. bank ("CEO James Smith") where frontline customer-service teams improved transactional NPS for a year or two, then progress stalled because other functional executives weren't on board — "it was just too easy for execs in other functions to ignore our old satisfaction scores... it seemed theoretical." Two concrete conflicts illustrate why: the fraud-control group, part of risk management and bonused on fraud losses prevented, kept subjecting the bank's most profitable customers to embarrassing in-store identity checks — their own analysis showed the fraud-prevention policy was profitable on average even counting the customer frustration it caused, so they "dug in their heels." Separately, a cost-cutting move to route payment-dispute calls overseas reduced handling cost but created more detractors, since overseas reps lacked the cultural context to empathize or jog customers' memories. Smith's fix was to change the bonus formula: customer metrics, including NPS, had previously been "shareholder metrics in disguise" (new-customer growth, revenue per customer) buried inside a balanced scorecard; he increased the weight on true customer metrics so that functional executives' incentive pay depended on hitting NPS goals — after which cooperation across functional lines to fix root causes increased. Virgin Media's Neil Berkett independently reports the identical bottleneck: the top of the organization and the front line embrace NPS quickly, but "the place that takes a lot of effort is in the middle... functional executives" who "knew how to control the historic metrics to make their bonus."

Why It Matters

It explains a specific, recurring failure pattern in any cross-functional change effort: individually rational local optimization (a function correctly executing against its own scorecard) can be collectively harmful to a shared outcome nobody's bonus reflects, and the people defending the harmful behavior are not being obstinate — they are being correctly incentivized by a different measure. The generalizable move is diagnostic and structural, not motivational: find where a shared goal crosses a function measured on something else, and change what that function is paid on, rather than simply asking harder for cooperation. The same pattern recurs wherever a security team is measured on incidents blocked rather than user friction caused, or a support team is measured on call-handling time rather than resolution quality.