# Decentralization Is the Constraint
## The org design is not a preference — it is what makes the acquisition rate possible
REQ's three cornerstones are capital allocation, decentralization and people (REQ, Jul 2025, p.72). Of the three, decentralization is the one that is structurally load-bearing:
> [!important] Without it, the M&A engine falters
> It is nearly impossible to sustain a cadence of 5–10 new companies a year — let alone 100 — if integration efforts and micromanaging consume management's time. In the long run it isn't feasible. The organisational design of these acquirers is **a feature, not a bug** (p.76).
Which means the screen runs in reverse too: *if a company has an aggressive acquisition target and no decentralised mindset in place, be sceptical* — REQ says so explicitly.
## What it actually looks like
- **Lean HQs.** Constellation Software: 6 independent operating groups, HQ around 15 people, over 100 acquisitions a year, no central integration (p.20).
- **Small business units by design.** CSI 2012: ~125 business units in ~50 verticals; biggest unit 307 employees, average 44; two-thirds of employees in units under 100. A regression of performance on business-unit size returned **R² < 0.001** — size and performance essentially unrelated. Mark Leonard's read: the advantages of being agile and tight far outweigh economies of scale (p.59).
- **Sister companies, not subsidiaries.** Bufab's "pull integration" — new sisters use best practice where it helps, not because they have to.
- **No cost synergies, only growth synergies.** Or, as Jörgen Wigh of Lagercrantz put it, synergies in the form of **energy and structure** injected into a complacent founder business at a succession moment (p.37).
> [!quote] Ulf Lilius, CEO of Momentum Group
> Decentralization for me is very easy. It's like jeopardy. You ask the question, but they have the answers. That's how you lead. You lead by questions.
## The trust argument
REQ leans on Axelrod's *The Evolution of Cooperation*: head office offers autonomy and trust from the start (tit-for-tat's opening friendly move), subsidiaries sense the freedom and reciprocate with decisions that benefit everyone. And on organisational psychology's **crowding out** — a singular focus on external rewards lowers the intrinsic desire to perform the task. Carrots, not sticks (p.73, p.37).
## The Swedish explanation
Why so many of these are Swedish: a 19th-century industrial heritage; a small isolated country that internationalised early through acquisition; Jan Wallander's decentralised governance model at Handelsbanken in the 1970s spreading through Swedish business; Atlas Copco's 1976 reorganisation cutting group management functions to ~200 people; a dense population of family-owned SMEs; **public annual reports for all companies**, which lowers post-acquisition surprise; and among the highest interpersonal trust levels in the world (p.79–80).
## Why it matters
Decentralization is the reason the seller says yes and the reason the buyer can keep buying. It is simultaneously the moat and the throughput limit.
## Related
- [[Acquisition-Driven Compounders MOC]]
- [[Becoming the Preferred Buyer]]
- [[Scaling M&A and the Runway]]
- [[Why Rollups Deflate]]