When to Centralize, When to Fragment: The Shared Services Dilemma

In business, scale is a blessing and a burden. On one hand, you gain leverage—more volume, more reach, more negotiating power. On the other, you invite complexity—more teams, more systems, more nuance. For CFOs, nowhere is this tension more evident than in the perennial question: should we centralize or decentralize?

It’s the shared services dilemma. And it’s as old as the first back office.

Every growing enterprise eventually hits a fork in the road. As functions multiply—finance, procurement, HR, IT—the temptation grows to consolidate. Consolidation promises efficiency, consistency, control. But done wrong, it breeds bureaucracy, distance, and bottlenecks. The opposite approach—fragmentation—favors speed, autonomy, and customization. But it can lead to waste, redundancy, and a lack of strategic cohesion.

There is no universal answer. But there is a way to think about it. And the best CFOs treat this decision not as a once-and-done structural choice, but as a dynamic system that adapts as the business evolves.

Why We Centralize

Centralization is often the CFO’s first instinct—and for good reason.

• Cost Savings: The most immediate benefit is operational efficiency. Consolidating AP, payroll, reporting, or procurement under one roof creates economies of scale. Instead of having five teams doing the same job across regions, you have one standardized process and one trained team. You negotiate vendor contracts once, not five times. You integrate systems once, not in patches.
• Control and Compliance: Centralization allows tighter internal controls. When all payments flow through one process, it’s easier to detect fraud, enforce approvals, and manage audit trails. In regulated industries, this can be a lifesaver.
• Data Consistency: In a centralized model, reporting becomes cleaner. You control the definitions. ARR means the same thing in Singapore as in San Francisco. Your dashboards don’t need footnotes or caveats based on business unit quirks.
• Talent Leverage: You can afford to invest in better tools, analytics, and leadership when you’re not duplicating effort. A centralized FP&A team, for example, can recruit top-tier talent that smaller business units couldn’t afford on their own.

Why We Fragment

Despite the allure of efficiency, centralization has a dark side: rigidity. Businesses don’t all move at the same pace. Functions embedded in product, sales, or regional ops often feel that central services are slow, disconnected, or optimized for HQ—not for them.

This is why fragmentation—done intentionally—can be powerful.

• Speed and Autonomy: A product team with its own analyst or finance business partner can make decisions faster. They don’t wait two weeks for a forecast—they build it in the room. When regions run their own purchasing, they can source local suppliers more quickly.
• Customization: Not every part of the business has the same needs. A global SaaS team doesn’t manage contracts the same way a hardware subsidiary does. Fragmentation allows local solutions to thrive.
• Empowerment and Ownership: Localized teams often feel more accountable when they own their numbers. This can drive better performance than a shared team viewed as an external cost center.
• Innovation: Sometimes the best ideas come from the edge. Decentralized functions experiment more. If one region builds a better collections process, it can later be adopted enterprise-wide.

So What’s the Right Answer?

The best organizations don’t pick one. They build a hybrid operating model—one that centralizes functions with high compliance needs or economies of scale, while embedding flexibility where responsiveness matters most.

Here’s a framework that many CFOs use:

Function/Need	Centralize or Fragment?
Accounts Payable	Centralize
Payroll	Centralize
Tax and Compliance	Centralize
FP&A	Hybrid (central + embedded)
Procurement	Hybrid
HR Operations	Centralize
Talent Acquisition	Fragment (or regional)
Revenue Operations	Fragment or Hybrid
IT Infrastructure	Centralize
Business Analytics	Hybrid

The trick isn’t just drawing these lines. It’s building the right interfaces between the center and the edge.

Building a Functional Shared Services Model

If you do centralize, it’s not enough to consolidate. You have to design for service. Too many shared service centers become slow, opaque, or adversarial. That’s not shared services—that’s hostage-taking.

Here’s what distinguishes great shared service functions:

• Clear SLAs and KPIs: You can’t manage what you don’t measure. Set expectations for responsiveness, accuracy, and resolution times. And report them publicly.
• Internal Customer Mindset: Treat your internal teams as clients. Listen to their feedback. Build escalation paths. Appoint relationship managers if necessary.
• Technology Enablement: Centralization works best when powered by automation. A centralized AP team with no workflow tool will fail. Invest in systems that scale.
• Feedback Loops: Shared services must evolve. Build mechanisms to collect complaints, identify process friction, and update policies.
• Financial Transparency: If you’re allocating shared service costs back to BUs, do it openly. Hidden charges breed resentment. Show how costs are calculated and what value is delivered.

The Role of the CFO

As CFO, your job is not to pick centralization or fragmentation—it’s to design the system that serves the business. That means constantly reevaluating:

• Are shared functions enabling growth, or slowing it?
• Is fragmentation driving agility, or creating chaos?
• Where are the hidden costs—process, morale, or reputation?
• Which decisions need consistency, and which need speed?

It also means mediating conflict. Business units will always want more freedom. Corporate functions will always want more control. Your role is to find the equilibrium point where effectiveness and efficiency meet.

Lessons from Real Transitions

Many organizations attempt centralization during downturns or transformations. The impulse is to cut costs by consolidating. But this only works if the operating model already supports standardization. For example, if every region uses a different ERP or invoice coding scheme, centralizing AP will create more confusion, not less.

Conversely, some companies fragment too early—allowing every BU to hire its own finance team, only to find that metrics can’t be consolidated and the company runs on five definitions of gross margin.

The key lesson is timing. Centralize after processes have been standardized. Fragment only when local complexity demands it.

Final Thought

In life and in business, centralization and fragmentation are not ideologies. They are tools. A good craftsman doesn’t swear allegiance to one wrench. They use what fits.

For the finance leader, the shared services dilemma is not about choosing sides. It’s about building a machine that runs well, adapts often, and aligns to strategy—not structure.

Run lean where it matters. Run fast where it counts. And above all, design for clarity—because even the best team can’t execute a blurry playbook.

When in doubt, remember: the goal of finance isn’t to be big. It’s to be right.