The cost base inside the fund.

Most of what gets written for GPs treats operating discipline as a portfolio company problem. The CEO has fifteen workstreams. The operating partner has eight assets. The fund admin charges what they charge. The tech stack is the tech stack. Useful work is happening, but it's all happening downstream.

The harder question, and the one that compounds quietly through fund cycles, is whether the GP's own house is being run on a cost base that still makes sense.

On current evidence, in most private markets funds I look at, the answer is no.

The drift has already happened

Most private markets GPs are running their back office on a cost base shaped by decisions made five or ten years ago.

Fund administration fees set when the AUM was half the size and the workflow was half as complex. Technology stacks that grew by accretion rather than design, with overlapping coverage between the LP portal, the CRM, the data room, and the reporting layer. Distribution operations built jurisdiction by jurisdiction as the fund expanded into AIFMD, then ELTIF, then non-EU placements, without anyone going back to ask whether the resulting structure is the structure you would have built if you had started from a clean sheet today.

The result is a cost base that grew up rather than was designed.

It is rarely catastrophic. The fund still functions. The reporting goes out. The capital calls work. But the line items are 30 to 50 percent above what a fund of comparable size and complexity would pay if it priced the work today, against today's vendor landscape, with today's automation available.

That gap doesn't show up in any single quarter as a problem. It shows up as a slow drag on management fee economics that nobody is paid specifically to fix.

The size of the gap, in the numbers that matter

Take a private markets GP with £1.2 billion AUM across two active funds, running a typical operating shape for a firm of that size. Twenty-two heads, fund administration outsourced, technology stack assembled over six or seven years, distribution into three EU jurisdictions plus the UK. The numbers below are calibrated against a private equity profile, but the same dynamics apply across private credit, infrastructure, real estate, and secondaries.

The cost base for that firm probably sits somewhere around £8 to £10 million in run-rate operating expense, excluding deal team compensation and carry.

In my experience, looking at firms of that profile, between £1.2 million and £2.4 million of that annual run rate is recoverable without service degradation. Sometimes meaningfully more. The recoverable savings sit across four categories.

Fund administration fee renegotiation is the largest single line. Most private markets GPs are paying fee schedules locked in three to five years ago, often based on AUM tiers that no longer reflect their fund size, complexity, or what the market is now charging. Annual saves of £300k to £700k are typical for a £1.2 billion firm that benchmarks honestly.

Technology rationalisation is the second. Overlapping LP portal, CRM, deal pipeline, and document management licences. Tooling bought to solve a problem that was later solved differently. Annual saves of £200k to £500k are usually findable without removing capability.

Distribution operating model is the third. The marginal cost of a new EU jurisdiction is far lower than the cost of running parallel structures across all of them with overlapping local counsel relationships, depositary banks, and reporting chains. Annual saves of £150k to £400k are recoverable through structural simplification under AIFMD II and ELTIF 2.0, which most private markets firms have not yet operationalised.

AI adoption inside the GP itself is the fourth, and currently the largest opportunity in absolute terms. Most private markets GPs are asking their portfolio companies to adopt AI faster than the GP itself is doing internally. The deal team is still doing manual desk research. The IR team is still drafting LP updates from blank pages. The CFO function is still doing variance analysis the way it was done in 2018. The opportunity inside the GP is not exotic, it is the same automation playbook being deployed in the portfolio. Annual savings on FTE redeployment and time recovery commonly run to £500k or more for a firm of this profile, and the strategic effect is larger than the cost line.

Aggregated across these four categories, the recoverable economics for a £1.2 billion private markets GP are typically £1.2 million to £2.4 million per annum. On a five-year fund cycle, that is £6 million to £12 million of management fee margin recovered. For a firm where the partner share of management fees pays the partners' base compensation, that is not a marginal number.

It is rarely framed that crisply because no-one inside the firm is paid to find it.

Why this gap exists, and why it will not close on its own

The reason the gap exists is structural, not analytical.

The COO of a private markets GP is running fund admin, technology, compliance, regulatory reporting, LP servicing, and increasingly cyber and data infrastructure. They have a small team. They are not under-resourced because the firm is poorly run. They are under-resourced because the role has expanded faster than the headcount.

Renegotiating fund admin fees properly is a six-week exercise. Rationalising the tech stack is a three-month diagnostic. Restructuring distribution operations is a fund-cycle programme. The COO does not have a six-week window. They have today's reporting, this week's audit, this month's regulatory deadline, and this quarter's LP requests.

The Managing Partner sees the cost base as a fact rather than a target. That is not negligence. It is a signal that the cost base is being read as fixed when it is, in reality, structurally negotiable. The signal is wrong, but the position is rational given the available bandwidth.

External advisors who could close the gap are typically paid as a percentage of savings, which creates a different problem. The advisor's incentive aligns with finding savings, not with respecting the operating reality of the firm. Vendors get strong-armed in ways that strain relationships the firm needs to maintain. The savings land but the cost shows up in service quality six months later. Most private markets GPs have been through this once and don't want to do it again.

What is missing is not analysis. It is a senior advisor who can do the work alongside the COO at a steady senior pace, with the right relationships into the vendor base to renegotiate without creating relational damage, and without an incentive structure that pulls them toward over-cutting.

That role does not exist in most private markets firms. The result is a cost base that drifts year by year and gets re-examined only when something breaks.

What changes if you treat the GP house as a target operating model

Three things follow from treating the GP's own operating shape as a target rather than an inheritance.

The work becomes calendar-driven, not crisis-driven. The fund admin fee renegotiation happens on a cycle, not when the contract surfaces a problem. The tech stack review happens annually, not when something falls over. The distribution operating model gets revisited every time a new fund vehicle opens, not three years later when the structural cost has compounded.

The economics get tracked. Most private markets GPs do not have a single number in front of the Managing Partners showing what the firm spends on fund admin per million of AUM, what the technology cost is per head, what the distribution operating cost is per LP commitment by jurisdiction. Without those numbers the cost base cannot be managed, only paid. With them, the conversation about where to invest and where to recover changes from intuition to structured trade-off.

AI adoption inside the GP becomes a credibility position with portfolio companies. A GP that has rebuilt its own deal-team workflow, its own IR function, its own CFO process around AI tooling can speak to its portfolio CEOs about AI adoption with operational authority. A GP that hasn't, can't. The portfolio companies notice. So do the LPs in the next fundraise.

The uncomfortable conclusion

The private markets GPs that will quietly outperform on management fee margin over the next two fund cycles will not be those that found the next great deal or hired the strongest deal team. The deal layer is competitive but not differentiating at the level of fund-level economics.

They will be the ones that treated their own operating shape as a target operating model from the day the fund closed, and who put the right person in the right role to hold that line over the cycle against the daily pressure of running a firm.

That is an operating discipline, not a vendor management discipline. It does not require a procurement function, more headcount, or larger consulting budgets. It requires a single, senior, accountable person whose mandate is the cost base, the technology stack, the distribution operations, and the AI adoption inside the firm, working alongside the COO rather than instead of them.

If you are a private markets GP, three questions are worth sitting with before the next operating committee.

What did the firm pay in fund admin fees this year, and when was the last time those fees were benchmarked against the current vendor market rather than against last year's number? If the answer is "more than three years ago," the gap is structural.

What is the firm's current AI adoption rate inside the deal team, the IR function, and the CFO function, and how does it compare to what the firm is asking its portfolio CEOs to deliver? If the GP is running behind the portfolio, the credibility cost is already showing up in board conversations.

Who, specifically, in the firm is accountable for the cost base as a target rather than as a fact? If the name is not obvious, the cost base is drifting.

The portfolio layer is where most operating attention currently sits. The cost base inside the fund is where most of the recoverable economics actually live.

That is the conversation worth having before the next fund cycle starts shaping next year's run rate.