Three years ago an LP on a quarterly call would nod at a 1.8x TVPI and move on. This year the same LP shows up with a model he built from public filings, a back-channel quote from a secondaries broker, and a question about why the firm's mark on Company X sits 30 percent above the last comp print. Disputes happened before, but they happened in a conference room, and by the time a number reached the letter it was settled. Now the CFO can spend most of a call defending one number, and the CFO owns LP trust, which I think makes it the hardest job in the firm.
Founders reforecast with AI. Comps are distorted by the hyperscalers. De-SPACs get re-marked, secondaries put a price on positions nobody planned to sell, the ARR-versus-CARR argument never ends, and LPs run their own diligence in the background. Every number in the letter is contested before the letter lands, and the CFO is the one who has to make sense of all of it. She translates founder reality into LP reality and back again, she defends marks on calls where the LP has brought conflicting numbers from three different sources, and she keeps an audit trail that has to survive a fund audit and an LPAC challenge in the same week. When an LP question goes sideways, the number on the letter is hers, and so is the defense of it.
Where the hours go
The teams I talk to spend their days re-keying cap tables out of PDFs into spreadsheets, digging through email threads for the most recent SAFE, and reconciling the same revenue number across three decks because the founder updated one and forgot the other two. The whole layer the LP letter rests on gets built by hand, every quarter, against a deadline that gets tighter every year. That work has to get done, and every CFO I have talked to knows it is not where she earns her keep.
I don't think this is a cycle. By 2027 the bar is going to be simple to state: any reported number, traced back to its primary document, in seconds, with the lineage attached so the LP can see where it came from before they think to ask.
The operational layer of the firm gets smaller when that happens. The hours that go to chasing documents and reconciling become background work that runs all the time instead of in a panic every ninety days, the people who were re-keying cap tables move to checking the edge cases the software cannot settle on its own, and the team ends up with a different mix of people rather than fewer of them.
The CFO's own time moves to portfolio construction, reserve strategy, and the LP conversations only she can have, where the question is what the number means for the next fund. That is the work that gets crowded out first when the team spends Friday night reconciling a deck.
What the LP is diligencing now
The data infrastructure becomes part of the track record. LPs will diligence the data layer the way they diligence the IC process, because one feeds the other. If the CFO cannot show how a number was produced, the number does not get believed, and if she can, each quarter's marks make the next quarter's easier to believe and the next raise gets easier with them.
A CFO who fixes the plumbing fixes the LP experience downstream from it. The marks get easier to defend, the next raise starts with a data room that is already in production shape, and the quarterly call stops being an hour of explaining. Fix the plumbing once and every fund the firm raises after that runs on it.
The firms that get there will have set the data up for this from the start, with every document carrying its lineage and every number carrying its source, so an LP question does not need an associate in a spreadsheet on Wednesday night. Supplying that layer is what we do at GoodStream, and the CFO is the person at the firm who decides whether it runs on one.
A CFO who builds that in 2026 will still have LPs who give her the benefit of the doubt in 2028, and at the firms that don't build it the quarterly calls will keep going the way they go now, one contested number at a time, with an associate looking for the document while the LP waits.
