Cost Allocation for a Shared Services Center

A cost allocation shared services center divides the cost of centralized functions—IT, human resources, finance, and procurement—among the business units that benefit from them. Companies choose between chargeback (actual billing) and showback (informational accounting) models, each creating different incentives and political friction.

Why companies need shared services allocation

Finance, IT, and HR naturally consolidate into central functions: a single payroll system, one data center, one audit and compliance team. These services span all business units, but their costs must eventually reduce unit-level profitability somehow. Without allocation, the parent’s financial statements show the expense, but business units—which are often measured by segment profit—cannot fairly compare themselves to external competitors or to each other.

A manufacturing business with three divisions cannot know whether Division B’s lower return on assets reflects weaker management or merely heavier use of IT infrastructure. Shared services allocation answers that: it tells each unit what it actually costs to run.

Chargeback vs. showback models

The two dominant approaches differ in one critical way: whether business units actually pay.

Chargeback treats the shared services center as a profit center or cost-recovery unit. The IT department, for example, bills each business unit monthly for the services it consumed. The unit pays the invoice from its budget; the central function’s revenue equals its cost. This creates a real P&L consequence: if a unit refuses to use IT’s cloud platform and builds its own, it saves the charge. If IT’s costs rise unchecked, the units feel it directly and may push for efficiency or threaten to outsource.

Showback is purely informational. A business unit sees a line item on its management accounts showing its share of central costs, but no cash flows between the unit and the center. Both the unit and the center remain cost centers. Showback is gentler politically—units do not feel as if they are being double-charged—but it also mutes the price signal. A unit manager has no incentive to avoid IT services because the charge is not real money leaving their budget.

Most large corporations run showback in their management reporting (for internal accountability) while using chargeback at transfer-price level (for tax and regulatory purposes), especially in multi-country structures.

Allocation bases and their disputes

The formula for splitting a shared cost across units hinges on what “consumption” means. Common bases include:

• Headcount: Simple. Assumes each employee costs the same to process payroll and HR. Works for HR but penalizes units with older or higher-wage workers who actually consume more compliance time.
• Usage-based: A company meters IT services (CPU cycles, storage, API calls) or counting transactions (purchase orders processed, expense reports, benefit claims). Fair in theory, expensive to track; units may game the system by batching transactions off-peak.
• Fixed allocation: Units pay a percentage of the total regardless of use—essentially a rent. Creates stability but kills incentive to conserve.
• Time tracking: Finance and audit staff log hours by unit. Accurate but labor-intensive and subject to error or self-interest in how analysts code their time.

Most corporations use a hybrid: a fixed base charge (say, 30% of IT costs, split evenly by headcount) plus a usage charge (70%, split by actual transaction volume or project hours). This balances stability with accountability.

The biggest dispute in cost allocation is the fixed-vs.-variable split. If IT infrastructure costs $50M annually and 80% of that is fixed (the data center lease, licenses, core staff), is that baseline cost driven equally by all units or only by the units that actually drove the infrastructure investment? A shrinking division may argue it should pay only for services it consumes, not for the sunk infrastructure, while the IT director argues the infrastructure exists for all units and the costs won’t actually fall if one unit shrinks.

Transfer pricing and tax implications

For multi-country corporations, shared services allocation is also a transfer-pricing matter. When a U.S. parent’s IT center provides services to its Indian subsidiary, the subsidiary must pay something—otherwise the parent bears all the cost and the subsidiary’s earnings look artificially high, inviting tax authority challenges.

IFRS and U.S. GAAP do not prescribe a single allocation method. The test is reasonableness: can you defend the allocation to an auditor? The IRS, especially, wants to see that inter-company charges reflect arm’s-length pricing—what an unrelated party would charge. A shared services center billing at cost recovery (no markup) is common; billing at a markup (to create a profit center) is defensible if the center is truly efficient.

When shared services become a profit center

Some mature organizations treat shared services as a profit center and charge a markup above cost recovery. Finance Center of Excellence bills business units 15% above cost; IT charges a percentage fee on the server budget it manages. This creates internal competition: units can theoretically outsource if external vendors undercut the internal charge.

In practice, units rarely do. Switching costs are high, and internal shared services providers have information advantages. The real discipline comes from external benchmarking: if IT’s cost per transaction is 20% above SaaS alternatives, the CFO knows it and can apply pressure.

The danger is that a shared services center, insulated from customer defection, becomes a bureaucratic cost center. Executives hedge this by treating the relationship as a contract: the IT director negotiates a three-year service-level agreement with the business units, and cost is locked in.

Common allocation disputes and how to resolve them

New hires and headcount changes: A unit that grows headcount quickly will see shared services costs rise. Is the increase legitimate or should it have been anticipated in the annual plan? Most companies set allocations once a year based on expected headcount, then true up annually. Mid-year headcount spikes are absorbed until the next cycle.

Temporary projects: A unit sponsoring a major systems migration uses hundreds of IT hours. Should the cost be allocated to that unit alone (because it chose the project) or spread across all units (because all benefit from the new system)? Governance here is critical. Most companies pre-approve and budget major shared services commitments, allocating costs based on the business case, not after-the-fact time tracking.

Services not used: If a unit doesn’t use the central cloud platform, should it still pay for capacity reserved for it? Chargeback philosophies differ. Some charge only for actual use (usage-sensitive); others charge a minimum even if unused (committed capacity). The latter creates incentive to actually use the service or opt out.

Quality disagreements: A shared services center has a service-level agreement (SLA) promising 99% uptime or 2-day turnaround on expense reports. If it misses, does the unit get a credit? Mature organizations include SLA-based rebates in the contract, ranging from 5% to 20% of the monthly charge for breaches.

Designing an allocation that sticks

Successful shared services allocation requires clarity on three points:

1. Allocation base is defensible: It correlates with benefit received, not arbitrary. Headcount per unit is defensible for HR; CPU consumption is defensible for IT infrastructure.

2. The model is disclosed and accepted: Business units see the formula in writing. Annual reviews with business leaders build buy-in, even when the news is that costs are rising.

3. Governance has teeth: A shared services steering committee, with representatives from major units, reviews performance, audits allocation accuracy, and approves material changes. Without this, the center loses credibility and units suspect they are being overcharged.

The allocation itself is less important than the ritual of transparency. A unit charged on a less-than-perfect formula but in a process it trusts will cooperate; a unit charged on a formula it doesn’t understand or trust will fight every invoice and lobby for outsourcing, even if the internal cost is cheaper.

See also