Complexity Has an Operating Cost.

Every layer, committee, and approval step an organization adds to gain control quietly adds to a bill that never appears on the P&L — until decisions start taking weeks instead of days.

Complexity is usually treated as an unavoidable cost of growth — more products, more markets, more regulation, more people. Some of that is real and can't be designed away. But most organizations respond to real complexity by manufacturing a second, avoidable kind on top of it: more approval layers, more standing committees, more reporting lines, more sign-offs. Each addition is defended, individually, as reasonable risk management. Collectively, they form something nobody actually designed. That second kind has a name, a body of research behind it, and a bill that gets paid in decision speed.

Complexity Is Not the Same Thing as Complicatedness

BCG's Yves Morieux, in the research behind his book Six Simple Rules (with Peter Tollman), draws a sharp line between the two. Complexity is external and largely real — the number of products, customer segments, technologies, and regulatory regimes a company actually operates across. Complicatedness is the internal response: the structures, procedures, coordination bodies, and approval chains organizations build to manage that complexity. Morieux's research, drawn from a survey of more than 100 U.S. and European listed companies, found that external business complexity grew roughly sixfold over six decades — while internal organizational complicatedness grew more than 35 times over in the same period. The response outgrew the problem by an order of magnitude.

That imbalance shows up directly in how managers spend their time. Inside the most complicated organizations Morieux studied, managers reported spending 40% of their time writing reports and 30% in coordination meetings — leaving roughly 30% of the week for the actual work of managing. Over 15 years, the same research found the procedures, layers, interfaces, and approval steps firms use to manage complexity increased anywhere from 50% to 350%.

For a product and technology organization specifically, this complicatedness usually shows up as an approval chain that has quietly grown a step every year: a new security review, a new architecture sign-off, a new cross-functional steering committee added after a past incident, none of them ever removed once the crisis that justified them has passed. Each step was locally reasonable when it was added. The chain they form together was never designed as a whole — and it's the chain, not any single step in it, that decides how long it takes the organization to ship a decision.

Every New Layer Adds a Decision Node

The mechanism is straightforward. Add a product line, a region, a matrixed reporting relationship, and you don't just add one new path through the organization — you multiply the number of places a decision has to pass through to get made. Bain & Company's decade-long research program behind Decide & Deliver treated this as directly measurable rather than a matter of culture: across more than 1,000 companies, they found decision effectiveness correlates with financial performance — revenue growth, return on capital, and total shareholder return — at a 95% confidence level, and that companies in the top quintile on decision effectiveness produce total shareholder returns nearly six percentage points higher than everyone else. Decision speed isn't a soft complaint about bureaucracy. It's a line that shows up in the return.

What Removing the Cost Actually Looks Like

The instinctive response to a complexity problem is to add a new mechanism to manage it — a steering committee, a PMO, another review gate. That's the exact reflex behind BCG's 35x complicatedness against 6x real complexity. Management researchers Gary Hamel and Michele Zanini, documenting Haier's transformation in Humanocracy, describe the opposite move: rather than adding a coordinating layer, Haier restructured itself into thousands of small, self-managing, market-facing units, each with direct accountability to a customer outcome rather than to a layer of management above it. The redesign didn't add oversight. It removed the layers oversight used to travel through, and pushed decision rights down to where the information already was.

That's the pattern worth taking seriously: the fix for a complexity problem is almost never a new structure sitting on top of the old one. It's fewer, clearer decision rights, held closer to the work — a redesign of the operating model, not an addition to it.

There's a reason complicatedness tends to outlast the org chart that produced it. Melvin Conway's 1968 observation that systems mirror the communication structure of the teams that build them applies just as directly to decision-making as it does to software architecture: a company that has quietly tripled in size while keeping the reporting lines and approval habits of a much smaller one will keep making decisions the way it always has, just with three times as many people now expected to weigh in. The structure wasn't redesigned when the business was. It just got more crowded.