A venture LP I had dinner with last week nearly missed a tender offer at more than two times NAV, because the position was sitting on his books at last year's mark and the offer didn't look interesting at first glance. He has invested in venture funds for decades, across dozens of them, and his record is the kind that makes other allocators ask where he's putting money. He still can't see his own portfolio.
His marks get updated once a year instead of quarterly. The ingestion is manual, and doing it more often was, in his words, a ridiculous amount of energy. Then he asked me the question I've been repeating to people ever since. "It's a database. Why can't I look at it?"
What the stale mark nearly cost him
The position was carried on his books at around $25,000 and was worth a large multiple of that. The tender notice had come months earlier and nobody caught it, because the mark on his statement was so far behind the real value that the offer didn't register. The context he needed was in a fund manager's email thread, and it wasn't on his own ledger, so he came within a few days of missing a liquidity event over a stale number.
That part of the LP experience doesn't show up in a pitch deck. The deck shows IRR and MOIC and TVPI. It doesn't show the LP opening his portfolio file, seeing a position carried at last year's number, and having no way to know whether anything behind it has moved, and it doesn't show the back-channel he keeps with each GP, the calendar reminders he sets to chase quarterly updates that arrive annually anyway, or the weekend in March he spends working out what happened the previous October. Every LP I've talked to tells me some version of this, at different funds and at different check sizes.
For two decades the structural answer has been quarterly reports and an annual audit. Real-time clarity would have required an army of analysts working every position in every fund, with the primary documents flowing through them constantly, normalizing every cap table and every SAFE and every term sheet by hand. The economics didn't work and the technology didn't exist, so everyone in the chain accepted the friction because the alternative was too expensive to consider.
I accepted it too. At my last company I made big decisions on numbers I knew were stale, and I treated that as the cost of operating in private markets.
What changed
That cost is collapsing. Software can now read a primary document, whether it's a cap table, a financing statement, a board deck, or a tender notice, and pull the relevant number in seconds with the source page attached, so the army of analysts turns into a small team checking the edge cases and what used to take a quarter to assemble gets assembled as the documents come in. For this LP, that means the database he asked to look at can exist, with a current mark in it instead of last year's, and a tender notice that shows up as a change in a number rather than as a line in somebody else's email.
LPs aren't about to demand a new reporting cadence, but they are, slowly, comparing their experience across the firms they back. A GP who can produce a defensible current mark on five days' notice gets credit for it, and a GP who still answers "we'll get you that after the quarter closes" loses a little ground with that LP every fund cycle without ever hearing about it, because the LP won't send a complaint email.
What the GP gets from fixing it is concrete. The marks get better, tenders get noticed while the window is still open, follow-on decisions happen on current numbers, and the DPI conversation at year end has fewer surprises in it. By my back-of-envelope math from his story alone, one captured tender at the right size pays for a decade of better infrastructure.
If you are an LP, you don't have to accept the friction as the cost of doing private capital anymore. You can ask which of your GPs can show you the document trail behind their mark and how long it takes them to produce it, and the answers will sort your GPs into two groups whose difference will show up on your books over the next three vintages.
The LP I had dinner with didn't need to be sold on any of this. He'd been living it for thirty years, and he wanted to know when the database would exist. We built GoodStream to be that database, with every number in it tied to the document it came from.
