Who owns the bottleneck: decision rights, cycle time, and three published cases

McKinsey & Company surveyed more than 1,200 managers across global companies. Fewer than half of respondents said decisions in their organisation were timely. Sixty-one percent said that at least half of the time they spent making decisions was ineffective. McKinsey estimated the opportunity cost, for a typical Fortune 500 company, at about 530,000 manager-days a year — roughly US$250 million in wages, about AED 918 million at the 3.6725 peg. [1]

Two findings appearing together is the useful part. Decisions are often not timely, and more than half of the hours spent on them are judged ineffective. That combination is what RAPID would predict when a decision has Recommend, Agree and Input roles but no single Decide: the organisation meets, then meets again.

The AED 918 million figure is a Fortune 500 wage equivalent. The transferable point for a smaller operator is the ratio — if 61 percent of decision time is wasted, adding more governance forums increases the numerator.

The same pattern appears when researchers measure who is allowed to decide, rather than how many meetings are held. MIT's Center for Information Systems Research surveyed 342 leaders in 2022 — 61 organisations with revenue of at least US$3 billion (about AED 11 billion), 272 below that threshold, with 57 percent of organisations operating outside North America. On average, respondents said only 47 percent of teams in the part of the organisation they knew could make decentralised operational decisions. [2]

Where most teams held those rights, large organisations needed 244 days on average to sense and seize a business, customer or technology opportunity. Where operational decisions stayed centralised, the same cycle took 566 days.

The gap between 566 days and 244 days is 322 days — a little over ten months. MIT CISR is not measuring meeting quality. It is measuring whether the teams closest to the customer, the process and the data are allowed to decide how to reach a goal leadership has already set.

Centralised firms in the sample sit below industry average on both margin and growth; decentralised firms sit above on both. That is consistent with delay compounding: a late sense-and-seize cycle shows up as missed price, missed capacity and a thinner new-product mix. The sample average of 47 percent of teams holding rights is the expensive middle — some units can act, most still escalate.

A 2025 operations paper on PT XYZ, a corrugated-box manufacturer in Lamongan, East Java, applied Goldratt's Theory of Constraints and Drum-Buffer-Rope scheduling to a delayed production line. [5] Bottlenecks were identified at the converting (SK-2) and stitching (SK-4) workstations. After the five focusing steps of TOC, DBR flow control, and additional hours at the constraint stations, reported throughput rose 32.53 percent.

A second 2025 food-manufacturing TOC study reported a related pattern. Cooling and packaging were the constraint, with cycle times of 21.6 and 24.1 seconds per unit and utilisation above 90 percent. Addressing those stations moved three measures at once. [6]

Neither plant bought a new market or a new strategy. Throughput moved because capacity was added only at the constraint, and non-constraint stations were subordinated to it.

The 32.53 percent and 18.9 percent gains are therefore a test of diagnosis. If a Gulf operator adds people or software across the whole chain and throughput stays flat, either the constraint was not identified or the Decide at that constraint was not named. Idle time falling 45.8 percent in the food case is the signature of a correctly chosen bottleneck: stations previously starved or blocked start flowing once the constraint is managed.

The same ownership gap shows up in regional operating systems, not only in US or East Asian plants.

GCC pharmaceutical supply remains roughly 80 percent import-dependent. [7] Baker McKenzie, writing in May 2026, put the regional pharmaceutical market at about US$23.7 billion — roughly AED 87 billion — with close to 80 percent of volume moving through GCC airspace and/or the Strait of Hormuz. [8]

In an 80-percent-import system the constraint is rarely manufacturing capacity inside the UAE. It is a release, a reroute, a temperature exception or a Tatmeen hold. Those are RAPID decisions. If the Decide is unnamed, wait time follows the GE pattern — years of a known defect in the record, minutes to fix once write-access is forced. The MIT CISR 322-day gap is the strategic version of the same delay.