Private markets analytics, explained

A public stock has a price every second the market is open. A venture-backed startup, a private-equity portfolio company or an M&A target has none of that — its value is whatever the last funding round, the general partner's own quarterly mark, or a negotiated deal price says it is. Private markets analytics exists because of that gap: there is no ticker to read, so performance has to be reconstructed from the timing of cash moving in and out, and company facts have to be estimated or verified without a regulator-mandated filing to lean on.

The questions people actually ask

A limited partner allocating to a venture or buyout fund wants to know whether the general partner is actually skilled, not just lucky, and whether the number in the pitch deck survives being adjusted for when cash actually moved. A GP running a portfolio company wants to know which one needs intervention now, before the next mark makes the problem visible. A deal team doing diligence on an acquisition target wants to know the target's real revenue, growth rate and customer overlap with competitors — numbers a private company has no obligation to publish. All three are working around the same structural fact: no continuous public price, so every measurement is either a periodic estimate or a reconstruction from indirect evidence.

The data it runs on

Three kinds of data carry most private markets work:

• Fund-level cash flows. Capital calls (money drawn from investors) and distributions (money returned), each dated. Because a GP — not the investor — controls when these happen, private-fund performance has to be measured differently from a public fund an investor can enter or exit at will.
• Periodic valuations (NAV marks). Portfolio companies are valued quarterly, usually by the GP itself under an agreed methodology, then audited annually. These marks lag reality and are typically smoother than a public price would be, since there is no daily market forcing a re-rating.
• Alternative data on private companies. Because private targets rarely disclose financials, diligence increasingly draws on indirect signals — web traffic, app usage, job postings, and consumer transaction data — that approximate revenue and growth without the company's cooperation. This is alternative data doing a job public-market investors use it to supplement, not replace.

Core metrics and how to read them

The internal rate of return is the standard headline number: the discount rate at which the present value of all cash flows — calls as negative, distributions and residual value as positive — nets to zero. It is a money-weighted return, meaning it is sensitive to when cash moved, not just how much came back. That is exactly why it can be gamed: a fund that delays capital calls using a subscription credit line pushes early cash flows later, which mechanically inflates a young fund's reported IRR without changing the underlying investment skill at all.

MOIC (multiple on invested capital) strips out timing entirely: MOIC = (distributions + residual value) / paid-in capital. A 2.5x MOIC means every dollar invested has returned two and a half, full stop, regardless of whether that took three years or nine. Reading IRR and MOIC together — a high IRR with a low MOIC often signals a fast, small win rather than a fund-defining return — is more informative than either alone.

TVPI splits the same idea into realized and unrealized components: total value (distributions already paid out, plus the residual NAV still marked on the books) divided by paid-in capital. The distinction that matters is how much of TVPI is actually distributed cash (sometimes called DPI) versus how much is still an unrealized, GP-marked NAV. A fund reporting an impressive TVPI that is almost entirely unrealized value has shown you a promise, not a return.

The public market equivalent answers a different question: would this money have done better invested in a public index instead? It reinvests the fund's own cash-flow timing into a benchmark index, so the comparison accounts for when capital was actually deployed rather than assuming a lump sum invested on day one — a fairer test than simply quoting "the S&P returned X% over the same years."

Average deal size shows up on the M&A and buyout side as a basic filter on comparable transactions: a $50 million add-on and a $2 billion platform deal are rarely priced or structured the same way, so any comparable-transactions analysis has to control for scale before it means anything.

How the work is done in practice

Private markets analytics splits into two distinct jobs, and the tooling differs for each. Measuring fund and portfolio performance — IRR, MOIC, TVPI, cash-flow modeling — is mostly spreadsheet and fund-administrator work; it is a smaller, more specialized market than the second job. Researching companies for sourcing, diligence and valuation comps is where the general-purpose research platforms sit. S&P Capital IQ is a mainstay here: it is used heavily in private equity and investment banking to pull standardized financials, including for private companies where data is available, and to build comparable-company sets against public peers as a valuation anchor.

Because most private targets disclose nothing a deal team can simply download, diligence leans increasingly on data providers that estimate performance from indirect signals rather than filings. YipitData builds estimates of company sales and growth, for public and private companies alike, from web-scraped and transaction-panel sources — giving an investor a pre-diligence read on a target's trajectory before management ever opens the data room. Earnest Analytics works from consumer card-spend data specifically, estimating a target's revenue, market share and customer overlap with named competitors, which is often exactly the question a corporate development team is trying to answer before deciding whether to bid at all. Neither replaces verified financials once a deal reaches exclusivity — both are what determines which deals get that far.

Common mistakes and misreadings

Comparing IRR across funds of different vintages and ages without adjusting for the J-curve. A young fund's IRR looks artificially poor early on, because capital calls precede any distributions; comparing a two-year-old fund's IRR to a mature fund's is not a fair test of skill.

Treating subscription-line-boosted early IRR as representative. A fund that delays calling capital via a credit facility will show a flattering early IRR that partly reflects financing mechanics, not investment performance — ask what the IRR looks like with the line's cash flows added back.

Reading TVPI as if it were realized profit. A fund can report a strong TVPI that is almost entirely unrealized NAV; until that value is distributed, it is a mark, not cash in hand.

Comparing raw fund returns to a public index return without a proper public market equivalent. "The fund returned 18% and the market returned 12%" ignores when each dollar was actually deployed, which is precisely the distortion PME exists to correct.

Treating alternative-data revenue estimates as precise rather than directional. Web-traffic and card-panel estimates carry real sampling and coverage biases — they are strong signals for prioritizing diligence, not a substitute for a target's actual books.

For the platforms behind this research, see every tool in this category and every alternative data provider in this category.