Walk into the back office of almost any fund under half a billion in assets and you will find the same person. Different name, different title — controller, fund accountant, sometimes "director of finance" — but the same job. One person running payroll. The same person paying invoices. The same person chasing receivables. The same person closing the portfolio financials and assembling the quarter-end package.
Nobody designed this. It is what an org chart looks like when AUM grows faster than the back office and the headcount math never quite justifies the second hire.
I call it the one-accountant problem, and I think it is the most under-discussed structural risk in lower-middle-market private equity.
How funds get here
The path is almost always the same shape.
At fund one, an outsourced admin plus a part-time bookkeeper covers it. By fund two, there is a full-time hire — one good accountant who can genuinely do everything. Then the fund does what funds do: adds a portfolio company, adds a lender with its own covenant package, adds a co-invest vehicle, adds a management company with its own books.
Each new obligation is individually too small to justify a hire. So each one lands on the same desk.
Run the arithmetic the way a managing partner runs it. A second accountant costs, call it, $130k loaded. The overflow work is maybe fifteen hours a week — real, but not obviously $130k of real. So the hire gets deferred a quarter. Then a year. Then indefinitely. Meanwhile the one accountant's job has quietly tripled, fifteen hours at a time, and nobody ever made the decision on purpose.
That is the whole mechanism. Not negligence. Just headcount math that never clears the bar in any single quarter, compounding across twenty of them.
What it actually costs
The costs are real, and most of them are invisible until the worst possible moment.
- Key-person risk. If this person is out for two weeks in the wrong two weeks, quarter-end does not happen. The institutional knowledge — which entity pays which invoice, why that one portfolio company reports on a lag, where the workpaper for the waterfall actually lives — is in one head and a folder of spreadsheets only she can navigate.
- The quarter-end crunch. Everything converges at once: portfolio financials, the capital account statements, the LP package, the management company close. Two brutal weeks a quarter, four times a year, forever.
- Deferred quality. The reporting LPs would actually love — attribution detail, portfolio KPI trends, anything analytical — never gets built, because there is no slack. When one person is the assembly line, the only output is what the deadline demands.
- Burnout, then turnover. And the fund discovers the true cost of the role only when it has to replace it — usually mid-quarter, usually badly.
None of this shows up on a budget line. All of it shows up eventually.
Why the usual fixes disappoint
Two standard answers exist, and both tend to underdeliver for the same reason.
Outsourced fund administration. The pitch is that the work leaves the building. In practice, the work gets a commute. The admin needs inputs, so the one accountant now prepares packets for the admin. The admin produces output, so the one accountant now reviews it and reconciles their NAV against her own books — because she signs, and she knows she signs. Add the latency of every handoff, and quarter-end is now a negotiation with a service desk. The desk did not get lighter. It got a second inbox.
New software. The pitch is efficiency. But every system needs feeding. Who does the implementation? The data migration? The monthly upkeep, the chart-of-accounts mapping, the exception queue? The one accountant. You have handed the assembly-line worker a new machine and called it relief.
Both fixes share the same flaw: they add work around the accountant instead of removing work from her. The desk stays the bottleneck; only the traffic pattern changes.
Promote the accountant
Here is the reframe I would offer any fund living this, and it is the entire basis of how I build.
Do not give the one accountant another tool to operate. Deliver finished work product. The draft financials, done. The reconciliations run, with exceptions flagged rather than buried. The quarter-end package assembled, schedules and support attached — arriving on her desk the way work arrives on a partner's desk: complete, and waiting for judgment.
The one accountant should be the controls layer, not the assembly line.
That is the promotion. Not a new title — a new position in the process. She stops producing the work and starts reviewing it. She reads, she challenges, she signs. The judgment was always the valuable part of the job; assembly was just what the role defaulted to when there was no one else to do it.
This is what agent systems are actually good for. Not chat. Not "insights." Production — the repeatable, rule-bound, evidence-backed assembly of the work a fund's back office turns out every quarter, structured so that a human can inspect it and put their name on it.
I spent three years inside a boutique private equity firm building exactly this: pipelines that produce the operational work, and humans who approve it.
Conservative by design
If the one accountant is going to be the controls layer, the machinery underneath her has to be legible. A reviewer can only sign what she can see.
So my systems are built the way a CFO would want them built. Every run is logged — over the last five months or so, 920-plus logged agent runs at 99.7%. The three failures in that log are not rounding error I wave away; each one is accounted for. I can tell you which run, what went wrong, and what changed afterward. And anything that spends money or publishes externally stops and waits for a human signature. No exceptions, no clever workarounds.
That structure exists because of the one-accountant problem, not despite it. When one person carries the controls function for an entire fund, the worst thing you can hand them is a black box. The right thing to hand them is finished work with its evidence attached — so that review is fast, sign-off is defensible, and the audit trail writes itself.
Notice what this does to the role. It does not shrink it. The accountant who reviews and signs is more valuable to the fund than the accountant who keys and assembles — closer to the numbers that matter, further from the ones that don't, and no longer the single thread the whole quarter hangs by. The key-person risk doesn't vanish, but it moves to where it belongs: judgment, which you want concentrated, instead of throughput, which you don't.
The test
If you run a fund this size, the diagnostic takes one question: when your accountant takes two weeks off, what happens to quarter-end?
If the honest answer is "it waits," you have the one-accountant problem. The fix is not a second inbox and it is not another system to feed. It is moving the work off that desk and leaving the signature on it.
If this sounds like your back office, the discovery page is where that conversation starts — I keep time open for exactly this.