Perspectives · Fund Operations
You Can't Spreadsheet a Track Record LPs Will Trust
Three IRR conventions, two waterfall structures and one LP question that exposes them all. Why a spreadsheet can't hold a private markets track record.
· 5 min read
Part of the guide Private Markets Fund Data: Common Problems and How to Fix Them
If you run finance or investor relations at a private markets firm between $500 million and $10 billion AUM, you probably don't have a dedicated performance team. You have a controller, a few spreadsheets, and an LP who just asked a question that should take ten minutes. This piece covers the one problem that surfaces only when a question spans funds: your private equity track record doesn't hold together.
The question that takes three teams a week
An LP asks for net returns with and without one deal. Simple request. At a growing GP with three funds, it takes three teams a week.
Here's why. Fund I was built by the founder in a spreadsheet with unwritten conventions. Fund II's controller adapted the model to a new administrator's data feeds. Fund III came from an acquired team or a rebuilt tracker. Each spreadsheet foots on its own, so nothing flags as an error. The problem only appears when a question crosses fund boundaries and forces you to reconcile three sets of assumptions that were never written down.
Your administrator can't bail you out. Their remit is fund-level NAV and capital balances, not cross-fund bespoke cuts. When you ask for net IRR excluding one deal across three funds, you're asking for something nobody built. And when one restated performance figure surfaces, the LP has to re-verify every other number in the deck. Trust erodes fast.
There are now three IRR conventions, two waterfall structures, and no single place where a cross-fund question gets answered.
The conventions that actually diverge
When you dig into the numbers, five net IRR conventions diverge across your funds:
Cash-flow timing. Fund I uses mid-period dating. Fund III uses exact dates with XIRR. The same cash flow lands in different periods and produces a different IRR.
What "net" nets out. One fund nets out management fees, carry, and fund expenses. Another nets only fees and carry. "Net" is not a universal term.
Subscription-line treatment. Whether borrowed capital calls are netted out changes IRR materially. MSCI found in 2024 that the median subscription line inflated IRRs by roughly 100 basis points for recent buyout and real estate fund vintages.
Waterfall type. Fund I uses American deal-by-deal carry. Fund II uses European whole-fund. The same gross return produces different net cash flows to LPs.
Currency translation. Spot rates, period-average rates, or transaction-date rates. Pick one per fund and you get three different answers for any deal with cross-border cash flows.
What an auditable track record looks like
Document the conventions. Write down every assumption per fund: cash-flow timing, what "net" includes, subscription-line treatment, waterfall type, and currency method. A one-page memo per fund is enough to start. The goal is to make your IRR calculation methodology explicit before someone leaves or a new fund launches.
Centralize the inputs. Land administrator data and internal accounting data in one place, structured the same way across funds. This doesn't require a platform on day one. A single cross-fund data tab with consistent column headers beats three separate workbooks that each foot independently.
Make every number traceable. Each figure should link back through its conventions to the originating NAV pack or ledger entry. If an LP asks how you got a number, you should be able to point to the source document and the method applied. Version-stamp your snapshots so you can reconstruct what the deck said on the date you sent it.
Weigh auditability above calculation sophistication. An LP trusts a simple, fully traceable number over an opaque, sophisticated one. The ILPA Performance Template, released in 2025, requires net IRR and net TVPI both with and without subscription lines. That requirement is itself an admission that one number has no honest answer without its conventions attached.
Where to start
If you don't have a performance team, start with what you can control this month. Draft a conventions memo for each fund. Build a single cross-fund data tab that pulls from the same administrator reports. Version-stamp every snapshot you send to an LP. These makeshift steps won't solve the structural problem, but they will buy you time and surface where the real gaps are.
The structural fix is to stop computing firm-level numbers in spreadsheets that were never designed to reconcile across funds. Aquata's Performance Track Record use case lands administrator data and internal accounting data in one governed layer, structured the same way, before anyone computes a firm-level number. Documented conventions are applied consistently across funds, so IRR, TWR, MOIC, TVPI, and DPI come from the same inputs at fund, deal, and investor level. A cut such as net returns excluding one deal becomes a parameterized query rather than a week of reconciliation.
Every figure is traceable back through its conventions to the originating NAV pack or ledger entry, and bi-temporal records let you reconstruct any number as it stood on a past date. No performance team or data engineer assumed. For a firm looking to turn fragmented fund data into a track record LPs will trust, that is the difference between answering a question in ten minutes and answering it in a week.
