Private Equity — Article 6 of 12

Portfolio Company Value Creation — Shared Service Centers and CoEs

Private equity operating teams increasingly build shared service centers and Centers of Excellence that serve 8-25 portfolio companies simultaneously, compressing G&A costs by 25-40% and accelerating EBITDA expansion. This deep-dive covers the economics, operating models, technology stack, and governance pitfalls.

9 min read
Private Equity

Large-cap and upper-mid-market sponsors have quietly turned shared service centers (SSCs) and Centers of Excellence (CoEs) into one of the highest-ROI plays in the portfolio operations toolkit. KKR Capstone, Bain Capital Portfolio Group, Blackstone Portfolio Operations, Carlyle's Global Solutions team, and Advent's Portfolio Support Group now operate cross-portfolio platforms that serve 10-30 portfolio companies in finance, HR, IT, procurement, data engineering, and increasingly AI/ML. The economics are straightforward: a $40M revenue add-on company cannot afford a senior FP&A lead, a cybersecurity architect, or a Snowflake engineer. A platform spanning 18 portcos can — and amortize the cost at $80-150K per portco per year instead of $400-600K standalone.

This article walks through how to design, build, and run these platforms. We focus on the unit economics, the captive-vs-BPO-vs-hybrid decision, the technology stack that determines whether the SSC scales past three portcos, and the governance traps that have killed otherwise sound programs. The discussion connects to the 100-day integration playbook and the fractional talent operating model covered elsewhere in this guide.

The Economics: Why Cross-Portfolio Platforms Beat Standalone Functions

In a typical mid-market buyout (EV $250M-$1.5B), G&A runs 8-14% of revenue. Finance, HR, IT, and procurement together account for 60-70% of that G&A spend. When an operating partner consolidates these functions into a sponsor-level shared service — typically structured as a separate legal entity that bills portcos at cost-plus 5-8% — the blended unit economics improve through three levers: labor arbitrage (35-55% savings by relocating to Hyderabad, Krakow, Manila, or Monterrey), span-of-control expansion (one AR analyst now processes 4-6 portcos' invoices versus one), and automation overlay (UiPath bots, Workday, NetSuite, and Coupa eliminate 25-40% of remaining manual touches).

$1.8-3.2MAnnual run-rate G&A savings per portfolio company once a mature SSC absorbs finance, HR ops, IT helpdesk, and procurement. Based on benchmarks across 40+ PE-backed mid-market portcos serviced by Genpact, WNS, EXL, and captive PE platforms in 2023-2025.

At a 6-8x EBITDA exit multiple, $2M of recurring G&A savings translates to $12-16M of enterprise value per portco. Across an 18-portco platform, that is $216-288M of created value before any revenue synergies, cross-selling, or commercial CoE work. This is why sponsors like Vista Equity Partners (through Vista Consulting Group) and Thoma Bravo have built 200+ person operating organizations dedicated to running these platforms across their software portfolios.

Typical G&A Cost Reduction by Function After SSC Migration (% of pre-migration cost base)

Functional Scope: What Belongs in Shared Services vs. CoEs

The cleanest design separates transactional shared services (volume-driven, process-standardizable) from Centers of Excellence (expertise-driven, project-based). Sponsors that conflate the two end up with helpdesks staffed by data scientists — expensive and unhappy. The split should be deliberate.

Shared service centers handle invoice processing (P2P), order-to-cash (O2C), record-to-report (R2R), payroll administration, benefits ops, IT service desk Tier 1/2, master data management, and travel & expense. Volumes drive economics: an SSC absorbing 200K invoices per year across 12 portcos hits the productivity curve. Coupa, Bill.com, Tipalti, and SAP Concur typically anchor the AP stack; ADP Workforce Now or Workday handles payroll across the portfolio.

Centers of Excellence sit one layer up. They house scarce, expensive talent that no single portco can justify: cybersecurity engineers (often supporting the work described in cross-portfolio cyber posture management), Snowflake/Databricks data engineers, AI/ML scientists, Salesforce CPQ specialists, SEO/performance marketing pods, pricing analytics teams, and FP&A modeling experts who parachute in for budgeting season. CoE pricing is typically project-based or capacity-based (e.g., 40 hours/month retainer) rather than per-transaction.

SSC vs. CoE Design Choices
DimensionShared Service CenterCenter of Excellence
Primary driverTransaction volumeSpecialized expertise
Talent profileProcess operators, accountants, L1/L2 ITEngineers, data scientists, domain specialists
Pricing modelPer-transaction or per-FTE supportedProject fee or capacity retainer
Typical locationHyderabad, Manila, Krakow, BogotáBangalore, Warsaw, Austin, Lisbon, Dublin
Tooling spineNetSuite, Workday, Coupa, ServiceNowSnowflake, Databricks, GitHub, Salesforce, Looker
Onboarding a portco6-12 weeks2-4 weeks for first engagement
Margin target (cost-plus)5-8%10-15%

Operating Model: Captive, BPO, or Hybrid

Three structures dominate. The captive model — owned and operated by the sponsor — gives maximum control and captures all the arbitrage but requires real management bandwidth and 18-24 months to reach steady state. KKR Capstone's India captive in Gurgaon and Blackstone's Hyderabad-based BXAccess (built out in 2023-2024 to over 600 FTEs) are large-cap examples. The third-party BPO model, used heavily by mid-market sponsors, contracts with Genpact, WNS, EXL, Infosys BPM, or Accenture Operations to run the shared services on a managed services basis. Setup is faster (90-120 days), but margin leakage to the BPO is 12-20% and process customization is constrained.

The hybrid — a captive shell with BPO-supplied labor ("build-operate-transfer," or BOT) — has become the default for sponsors with $5-25B AUM. The sponsor signs a 3-5 year BOT contract with a partner like WNS or Quatrro, which stands up the center, hires the team, and transitions ownership in years 2-3. The sponsor gets time-to-value of a BPO with end-state economics of a captive. Hahn & Company, Stone Point Capital, and several Nordic sponsors have used BOT structures with success.

We modeled captive versus BPO for 14 months before realizing the right answer was neither alone. We signed a BOT with a tier-one provider, took ownership in month 22, and now run 19 portcos through it at 31% lower cost than the prior fragmented baseline. The transition language in the original MSA was worth more than the day-one rate card.
Operating Partner, $14B AUM mid-market PE firm

The Technology Stack That Makes Multi-Tenant SSCs Work

A SSC fails the moment it cannot cleanly separate one portco's data, workflows, and reporting from another's. Multi-tenancy is a technical problem with a vendor answer. The reference stack we deploy at mid-market sponsors looks like this:

ERP layer: NetSuite OneWorld (for portcos under $500M revenue) or Microsoft Dynamics 365 with multi-entity configuration. Larger portcos stay on their own SAP or Oracle instance, with the SSC operating those systems remotely via privileged access workstations. Workday Financial Management is increasingly chosen for software portcos with $250M+ revenue. HCM: Workday or ADP Workforce Now multi-EIN configurations. Procure-to-pay: Coupa or SAP Ariba for $100M+ revenue portcos; Bill.com or Tipalti below that line. IT service management: ServiceNow with multi-tenant domain separation, or Freshservice for smaller deployments.

Automation overlay: UiPath or Automation Anywhere bots for repetitive tasks (invoice OCR, bank reconciliation, payroll variance checks, AR cash application). A mature SSC runs 80-200 production bots eliminating 18,000-45,000 manual hours per year. AI document extraction from Hyperscience, Rossum, or AWS Textract has cut invoice processing time from 4.2 minutes per invoice (manual) to 22 seconds (straight-through) on AP volumes we benchmarked in 2024-2025.

Data and analytics: A central Snowflake or Databricks lakehouse ingests financial and operational data from every portco nightly. Fivetran or Airbyte handles the pipelines; dbt manages transformations; Looker, Power BI, or ThoughtSpot serves dashboards. This same backbone powers cross-portfolio benchmarking — letting the operating team show portco CFOs that their DSO of 58 days is 14 days worse than the portfolio median, with named comparables. See the data lakehouse architecture reference for design patterns.

⚠️Multi-Tenancy Is Not a Slide
We have seen at least four SSC programs derail because data segregation was treated as a policy concern instead of an architecture decision. If one portco gets sold and the buyer's IT diligence finds that another portco's vendor master, employee PII, or P&L details were accessible to the SSC tenant model, the indemnity exposure can run into eight figures. Row-level security in Snowflake, separate Workday tenants, and per-portco Azure AD tenants with cross-tenant guest access are non-negotiable.

Centers of Excellence Worth Building

Not every CoE pays back. The five that consistently deliver across our engagements:

High-ROI Centers of Excellence in PE Portfolios
Cybersecurity & Compliance CoE
SOC monitoring, vulnerability management, SOC 2/ISO 27001 readiness, M&A cyber DD. Typical staffing: 8-14 engineers serving 12-25 portcos at $40-80K/portco/year. Replaces $250-450K standalone CISO budgets.
Data Engineering CoE
Builds the lakehouse, BI dashboards, and KPI definitions in the first 90-120 days post-close. Standardizes pricing analytics, cohort retention, and unit economics models across the portfolio.
Procurement / Group Purchasing
Negotiates portfolio-wide contracts for AWS/Azure, Microsoft 365, Salesforce, insurance, freight, and benefits. Typical savings: 12-22% versus standalone portco rates. ROI is usually 8-15x program cost.
Revenue Operations CoE
Salesforce admin, HubSpot ops, CPQ, marketing automation, lead scoring. Particularly powerful in software and B2B services portfolios. See connection to add-on identification covered in the AI add-on screening article.
AI / ML CoE
Builds reusable models — churn prediction, dynamic pricing, demand forecasting, document automation — and deploys them across portcos. The 2024-2026 wave: agentic workflows for customer support and back-office, often built on Azure OpenAI or AWS Bedrock with vector stores in Pinecone or pgvector.

The procurement CoE deserves a closer look because it is the fastest payback. A sponsor with 22 portcos collectively spending $180M on cloud, software, and indirect categories can typically negotiate 15-20% off list through consolidated commitments. We routinely see AWS EDP discounts move from 8-12% (standalone portco) to 22-32% (sponsor-aggregated commit), Microsoft 365 E5 from $57/user to $42/user, and Salesforce from $165/user/month to $125 at portfolio-pooled volume. The CoE itself costs $1.2-2.4M/year to run. Net savings of $25-45M against $2M cost makes this the single highest-IRR initiative in the operating partner playbook.

Implementation Sequence: 24-Month Build

Standing Up a PE Shared Services Platform
1
Months 0-3: Strategy & Design

Diagnostic across 5-8 anchor portcos to size addressable spend. Decide captive vs. BPO vs. BOT. Select location (typically India + one nearshore). Lock in ERP and automation tooling. Build the legal entity, transfer pricing model, and inter-company services agreement (ISA) templates.

2
Months 3-9: Pilot with 2-3 Anchor Portcos

Migrate AP, payroll admin, and IT helpdesk for the first 2-3 portcos. Hire core leadership (SSC head, finance ops lead, IT ops lead, transformation lead). Stand up Snowflake/Databricks and the central data layer. Deploy first 20-30 RPA bots.

3
Months 9-18: Scale to 8-12 Portcos

Industrialize onboarding to a 6-week playbook. Add O2C, R2R, procurement ops. Launch the cybersecurity CoE and procurement CoE in parallel. Hit 60-65% of target run-rate savings.

4
Months 18-24: Full Portfolio Coverage + CoE Expansion

Onboard remaining portcos and new acquisitions on Day 30 post-close. Launch AI/ML CoE and RevOps CoE. Begin charging back to portcos at cost-plus. Migration of BOT ownership if applicable.

5
Months 24+: Continuous Optimization

Automation rate above 55% of transaction volume. New portcos onboard in 4-6 weeks. SSC becomes a diligence asset — buyers see lower run-rate cost structures and pay multiple expansion at exit.

Governance: The Things That Kill These Programs

Most SSC failures are not technology failures. They are governance failures. Three patterns recur:

Portco CFOs resist mandatory adoption. If the sponsor's investment committee does not write SSC participation into the management equity plan and the portco CEO's annual scorecard, two of every five portcos will quietly opt out, claiming "unique requirements." By month 12 the platform is sub-scale and unit economics never converge. The fix: the sponsor's operating partner sets adoption as a condition of access to portfolio resources, and the SSC publishes transparent benchmark data showing the portco's cost-per-transaction versus median.

Transfer pricing gets the IRS or HMRC's attention. An SSC billing $4M/year to 15 portcos in 6 jurisdictions is a transfer pricing audit waiting to happen. We have seen sponsors face $8-22M of retroactive assessments where the ISA, cost-plus methodology, and benchmarking studies were not documented to OECD Transfer Pricing Guidelines standards. Engage a Big 4 TP team during design — not after the first audit notice.

Service quality dips during onboarding. Portco controllers measure the SSC on whether last week's payables were processed cleanly, not on the 36-month vision. If onboarding waves cause spikes in error rates or DSO, internal political capital evaporates. Successful programs publish SLAs (e.g., AP cycle time ≤ 4 business days; payroll accuracy ≥ 99.7%; IT ticket resolution P1 ≤ 2 hours) and report on them weekly to portco CFOs and the operating partner.

🎯Treat the SSC as a Carve-Out Asset, Not Just a Cost Center
Several large sponsors — including a top-five global PE firm whose operating platform we have advised — are now exploring whether their internal shared services entity can be carved out and sold as a standalone GBS provider, or used to onboard portcos of other sponsors. At 1,000+ FTE scale, a captive SSC can command $250-600M of enterprise value as a managed services business in its own right. Build the legal structure, the third-party-grade SLAs, and the data segregation from day one to keep that option open.

Measuring Value Creation: What Actually Hits the LP Report

The SSC's value shows up in three places in the LP-facing value creation bridge. First, run-rate EBITDA expansion at each portco — typically 80-180 basis points of margin lift from G&A reduction alone. Second, working capital release — a well-run AR shared service compresses DSO by 6-12 days, freeing $3-9M of cash on a $200M revenue portco. Third, multiple expansion at exit — buyers pay higher multiples for portcos with clean financials, fast monthly close (target: 5 business days), real-time KPI dashboards, and SOC 2 Type II posture, all of which the SSC and CoE backbone delivers.

SSC Value Creation per Portco
VC = (ΔG&A × Exit Multiple) + (ΔWC × 1.0) + (Multiple Lift × Run-Rate EBITDA)
ΔG&A is annual run-rate G&A savings. ΔWC is one-time working capital release from DSO/DPO improvement. Multiple lift is typically 0.3-0.8x turns of EBITDA for portcos with clean books, fast close, and demonstrated cyber/data governance — all enabled by SSC + CoE infrastructure.
Sponsor Readiness Checklist Before Launching an SSC Program

Sponsors that get this right turn a back-office cost-cutting initiative into a structural competitive advantage. The 2024-2026 cohort of PE platforms with mature SSC+CoE infrastructure are now onboarding new acquisitions in 30-45 days, running monthly close in five business days portfolio-wide, and presenting buyers with diligence packs that include three years of clean, comparable, AI-ready data. That alone has moved exit multiples by 0.5-1.0 turns of EBITDA in transactions we have advised on in the last 18 months — value that compounds with the deal sourcing and add-on identification capabilities covered later in this guide.

Frequently Asked Questions

When does a PE firm have enough scale to justify building a shared service center?

The economics typically work above 6-8 portfolio companies with $250M+ combined G&A spend, or 4-5 portcos in the same vertical where process commonality is high. Below that threshold, contracting with an established BPO like Genpact, WNS, or EXL on a managed services basis usually delivers better risk-adjusted returns than building a captive.

Captive, BPO, or BOT — which model wins?

BOT (build-operate-transfer) has become the default for $5-25B AUM sponsors. It delivers 90-120 day time-to-value of a BPO while retaining the end-state economics of a captive after ownership transfers in years 2-3. Pure captive makes sense only for sponsors above $30B AUM with 25+ portcos; pure BPO is fine for smaller platforms or sponsors that do not want operational complexity.

How do you prevent portco CFOs from refusing to use the SSC?

Three mechanisms work: write participation into the portco CEO's annual scorecard and management equity plan, publish transparent benchmarking that shows each portco's cost-per-transaction versus portfolio median, and ensure the SSC's SLAs are genuinely better than what portcos had standalone. Mandate without service quality breeds resentment; service quality without mandate breeds half-adoption.

What is the biggest hidden risk in running a shared service center across portcos?

Data segregation. If one portco is sold and the buyer's IT diligence discovers that another portco's data was accessible inside the SSC's tenant model, the indemnity exposure can be eight figures. Multi-tenancy must be architected — Snowflake row-level security, separate Workday tenants, per-portco Azure AD — not addressed as a policy after the fact.

How quickly does an SSC pay back?

First-portco payback is typically 14-22 months including standup costs. Platform-level payback (the full investment in the SSC entity) is usually 24-30 months. By year 3, mature platforms run at 25-40% lower cost than the fragmented baseline and deliver 80-180 bps of EBITDA margin expansion per portco — translating to 0.5-1.0 turns of multiple uplift at exit.