Shared services across portfolio companies, and the part that actually repeats

Shared services in private equity usually means the fund's own back office. Fund accounting, investor reporting, capital calls, the machinery of the firm itself. That market is well served and it is not the subject here. This is about the other back office, the one inside the portfolio companies, where the month closes, the lender package goes out, and the operating work of an actual business gets done.

At that level the question arrives one company at a time. A management team runs short of capacity in finance or operations, somebody proposes a fix, and the fix gets designed from scratch by people who have never designed one before. Then the same question arrives at the next company, in a different industry, on a different system, and gets designed from scratch a second time. The work of answering it repeats. The answer rarely does.

Should a PE firm build shared services across its portfolio companies?

Standardizing how each company builds capacity tends to survive contact with a portfolio. Pooling the work into one center tends not to.

The instinct to pool is a rational response to fixed cost. A traditional capability center carries a foreign entity, a lease, a compliance function, and a site leader, and none of that shrinks with the size of the team inside it. Amortizing those costs across several companies is the only arithmetic under which they make sense. Take the fixed cost away and the reason to pool goes with it.

What remains are the frictions. Portfolio companies run different systems, different close calendars, and different industry vocabularies, so a pooled center's first job is translation, and translation is overhead that grows with every company added. Hold periods differ too. A center built to serve six companies is serving four of them a year later, and the allocation conversation reopens every time the composition changes.

The sharper version shows up at exit. A buyer buys the company. Capacity that lives inside the company transfers with it. Capacity the company was drawing from a center owned by the fund does not, so a diligence question opens on how the business runs the day after close, and the answer involves standing something up on a deadline set by someone else.

None of that argues against standardization. It argues against the one form of standardization that puts the capacity somewhere other than the company using it.

What operating capacity can be standardized across portfolio companies

The design standardizes. The team does not.

What travels between companies is a method. Which streams of work belong inside the company and which are better bought as a service, which is what the Own vs. Rent rule settles in a sentence. How a seat gets defined before anyone sits in it. What the first ninety days look like. Who holds the standard when the work starts to drift. Those are portable because they are structural, and structure is the part of an operating model that does not care what the company sells.

Everything downstream of that stays local. The people, the systems, the reporting calendar, the person each seat answers to, and which stream gets covered first, which depends entirely on where a given company is straining this year. A method that travels and a team that stays is the arrangement a portfolio can actually hold.

There is one exception worth naming, and it runs the other way. The reporting a sponsor asks of its companies looks similar in every company it owns, because the sponsor wrote it. That stream is the most standardizable work anywhere in the portfolio, and it is usually the least standardized, because each company assembles its own version of it by hand every month.

How operating partners replicate an operating model across a portfolio

The same way any operating standard replicates. One company installs it because that company wanted it, the installation works, and what gets written down afterward is the design rather than the outcome.

Operating partners already move playbooks this way. A pricing approach, a reporting standard, a systems selection. The pattern holds across all of them. A mandate lands as compliance and a working example lands as interest, and the second management team adopts because a peer it respects is already running the thing.

What actually travels is a small amount of written material. The seat definitions that held. The hiring standard that produced them. The cadence that kept the work from climbing back onto senior calendars. And the failure modes, which are the most valuable of the four and the only part that has to be paid for once.

What travels between portfolio companies is the design. The team stays where the work is.

What changes on company two

Most of the cost of a first installation is definition, and definition is the part that does not get paid twice.

The first company spends its time deciding what the work actually is. Which streams recur, what done means for each of them, what a seat owns and what it escalates. That work is slow in a way that has nothing to do with hiring. Even choosing which stream goes first is its own decision, and a company making it for the first time makes it carefully.

By the second company those definitions exist as drafts rather than as open questions. The interview standard exists. The onboarding sequence exists. The list of what went wrong the first time exists, which is why a second installation is usually less a design exercise than a fit exercise. By the third, the question has changed shape again. It is no longer whether to build capacity. It is which company gets it next, and when. The clock gets shorter still when a company inside the portfolio closes an add-on, where the work arrives on the day the deal does.

A management team that did not ask for the capacity will not run it well because a peer company did. That is the limit on all of this. Replication works company by company, at the pace the management teams actually want it, which is slower than a mandate and considerably more durable.

Where Kayana fits

Kayana builds this capacity inside one company at a time. The operators come from the capability centers the Fortune 500 built. They work inside a single portfolio company's cadence and tools, and the environment around every seat is designed and run by us rather than left to the company to invent. Grown seat by seat, that layer is the Mini-GCC model, and it takes no entity, no real estate, and no multi-year build. Nothing is pooled, because there is no fixed cost asking to be amortized.

The version built at one $75M PE-backed company was three people, at roughly $125K a year. Six months in, two of its senior leaders had recovered enough time to stand up two new product lines, and the same structure has since carried the team past 10 people. Voluntary turnover across the teams we build runs 3–5% in year one, which is the condition a design has to meet before it can travel at all. A design nobody stayed to run is a document rather than an operating model.

Quick answers

The rules that make an operating layer compound rather than reset are the subject of the book. Operational Alpha →