There's a version of quarter-end that exists in pitch decks and LP presentations: numbers roll up, reports go out, everyone moves on. Then there's the version that actually happens at a fund under $500m, where the entire close runs through one or two people who also have day jobs.
I spent three years inside that second version. Here's what it actually looks like.
Week one: the chase
The quarter closes on a Tuesday. Nothing happens on Wednesday, because nothing can. Every number the fund reports starts life inside a portfolio company, and portfolio companies at this end of the market do not have investor-grade reporting teams. They have a controller, maybe a fractional CFO, and a monthly close of their own that they're behind on.
So week one is email. Polite the first time, firmer the second, a phone call the third. At most funds this size, someone — usually the CFO, sometimes an associate who inherited it — keeps an informal mental tracker of which companies send clean packages, which send a trial balance and a shrug, and which go dark until someone senior calls someone senior.
Call it five to eight portfolio companies, three touches each, spread across two weeks. That chase alone is a part-time job, and it's held entirely in one person's head and inbox.
Week two: reconciliation
The packages arrive. None of them match each other. One company reports EBITDA with addbacks broken out; another buries them. One sends Excel, another sends a PDF export of Excel, which is the same information minus the ability to use it.
Now the real work starts: mapping every company's numbers into the fund's template, quarter after quarter, by hand. Tie the debt schedule to the lender statement. Tie cash to the bank. Chase the one number that moved 40% with no explanation, because that's the number an LP will ask about.
This is where the fund's spreadsheet — and at most funds this size there is one master spreadsheet, whatever the admin relationship looks like on paper — becomes the single point of failure. It has years of history in it. Its formulas were written by someone who may have left. Everyone is afraid of it, and everyone edits it anyway.
Week three: capital accounts and the letter
Valuations get marked. Management fees, carry accruals, and expenses get allocated across LP capital accounts — waterfall math that is genuinely intricate and genuinely unforgiving, because an error here doesn't produce a wrong chart, it produces a wrong statement to a specific investor with their name on it.
Then the LP letter. The partner wants it to sound like the fund. The CFO wants every number in it to tie. Those two goals meet in a document that goes through — realistically — eight to fifteen versions. Q3_Letter_v7_FINAL_partnerEdits_v2.docx is a joke everyone in this industry has stopped laughing at, because last quarter somebody genuinely almost sent v6.
The nights happen here. Not because anyone is slow, but because the letter can't start until the capital accounts are done, the capital accounts can't start until the valuations are marked, and the valuations can't be marked until the last portfolio company sends its package. Every delay upstream lands, compounded, on the person at the end of the line — usually at 11pm, usually in week three.
The structural cause, plainly
None of this is a competence problem. The people running these closes are usually excellent — that's why the fund survives on so few of them.
The cause is structural. The quarter-end process at a small fund is, almost line for line, the process a $2bn fund runs. Same portfolio chase, same reconciliation, same capital account math, same letter. The big fund runs it with a finance team of ten and a fund admin who actually performs. The small fund runs the identical process with one or two people and an admin who mostly re-keys what the fund already prepared.
A process designed for ten, run by two. That's the whole diagnosis. Everything else — the nights, the version chaos, the key-person fragility, the CFO who can't take a vacation in January, April, July, or October — is downstream of that one fact.
It's not an accounting problem. It's a logistics problem.
Here's the reframe, and it's the one idea I'd ask you to keep: the quarter-end is a logistics problem wearing an accounting costume.
Look back at the three weeks. Almost none of that time is spent on accounting judgment. Marking a valuation is judgment. Deciding an addback is legitimate is judgment. Everything else — the chasing, the collecting, the reformatting, the mapping, the tying, the version-tracking, the assembling — is movement of information from where it is to where it needs to be, in the right shape, in the right order, with a record of what happened to it along the way.
That is logistics. And logistics can be systematized in a way judgment can't and shouldn't be.
The judgment in a quarter close fits in a few hours. The logistics eat the month. We automate the logistics and protect the judgment.
That distinction is the entire design principle behind what I build. The system chases the packages, normalizes the formats, maps the numbers, ties them to source, drafts the assembly. The human marks the valuations, rules on the edge cases, and signs. Nothing that spends money or leaves the building moves without a person approving it — that's not a limitation I tolerate, it's a gate I build in on purpose.
Conservative by construction
I'll say the quiet part about automation in fund finance: most of it deserves the CFO's suspicion. A pipeline you can't audit is worse than a spreadsheet you can, because at least the spreadsheet shows its work.
So mine shows its work. Every agent run is logged. Over roughly five months I've logged 920-plus runs at 99.7% success — and the three failures are the part I'd actually show a CFO first, because each one is accounted for: what broke, why, what changed so it can't recur silently. A system that claims it never fails is lying. A system that logs its failures and explains them is one you can build a close on.
What "delivered" means
The end state isn't a dashboard, and it isn't a demo. It's a package on the CFO's desk: capital accounts, portfolio rollup, the letter draft — with every number traceable back through the pipeline to the source document it came from. Click the number, see the lineage. The CFO reviews it, applies the judgment that's actually theirs to apply, and signs.
The quarter-end, delivered. Not the CFO replaced — the CFO returned to the part of the job that needed a CFO in the first place.
The nights were never the price of rigor. They were the price of running a ten-person process with two people. Fix the logistics and the rigor stays; the nights go.
If this described your quarter a little too precisely, the discovery page is where that conversation starts.