Which benchmark should a VC fund use?

What a benchmark is supposed to satisfy

Cambridge Associates' framework for benchmarking private investments, published in March 2014, sets three criteria for a benchmark: that it be appropriate, meaning reflective of the manager's investment style and inclusive of a representative universe; unambiguous, meaning the underlying components and any weights are clearly defined; and investable, meaning it represents a viable opportunity set.

The same paper notes that private markets struggle with all three, because of confidentiality and limited access to top-performing funds. That is not a reason to skip benchmarking. It is the reason to state which criterion your chosen benchmark fails, since every private market comparison fails at least one and a manager who names it first is more credible than one who does not.

The seven comparisons, and what each one answers

Each answers a different question, and no single one answers the question an allocator is actually asking.

• Vintage-year benchmark. Compares funds that began investing in the same period, which controls for the entry environment. The most important single control in venture, and the one that makes cross-vintage comparisons meaningless without it.
• Stage benchmark. Compares pre-seed and seed against pre-seed and seed rather than against growth. Loss rates, holding periods, and dispersion differ enough by stage that mixing them produces a comparison of strategies rather than of managers.
• Geography benchmark. Compares within a market, since entry valuations, exit routes, and currency all differ. Cross-border comparison without adjustment is one of the more common quiet errors.
• Strategy benchmark. Compares concentrated against concentrated, and ownership-led against participation-led, because two funds with identical stage and vintage can run opposite constructions.
• Peer cohort. A specific set of named funds an allocator would consider alongside yours. Cambridge Associates' framework treats medians and quartile rankings against same-vintage, same-strategy funds as the best measure of whether a given investment was a good selection.
• Public market equivalent. Compares the fund's cash flows against the same flows invested in a public index, which asks whether the illiquidity was rewarded. Cambridge Associates developed a modified version to address the negative net asset value problem in earlier methodologies, and cautions against comparing an IRR directly to a public time-weighted return.
• Internal decision benchmark. What you hold yourself to: loss rate, ownership at entry, follow-on discipline, pacing. This is the one a first-time manager can actually produce, and it is more useful in a meeting than a quartile claim.

The limitations to state before an allocator states them

Every private benchmark carries at least three of these, and naming them is what separates analysis from marketing.

• Sample limitations. Narrowing by vintage, stage, geography, and strategy at once can leave a cohort of a handful of funds, where a median means very little.
• Survivorship bias. Datasets are built from funds that report, and managers who stop reporting are frequently the ones performing worst.
• Fund-age effects. A young fund's ranking is largely a function of marks rather than realizations, which is the finding above about quartiles moving for years.
• Strategy mismatch. A benchmark drawn from a broader private equity universe is not a venture benchmark, and a private-equity-wide statistic presented as a venture standard is a specific and common error.
• Data availability. Coverage differs sharply by geography, and thin coverage in a market does not mean strong relative performance in it.
• Methodology differences. Two datasets can compute the same measure differently, so a comparison across sources is often not a comparison at all.
• Selection of the comparison itself. The manager chooses the benchmark, which is why allocators discount favourable ones and ask what else you looked at.

Internal analysis against public marketing

These are different activities with different standards, and conflating them is where benchmarking creates regulatory exposure rather than insight.

Internally, use whatever comparison sharpens your own decisions, including rough ones. A cohort of six funds with known limitations is still informative to you.

Publicly, a benchmark comparison is a performance claim. Under the codified United States marketing rule, an advertisement may not contain a statement of material fact the adviser cannot substantiate on demand, may not discuss potential benefits without a fair and balanced treatment of risks, and may not present performance in a way that is misleading through what it includes or excludes. Any gross figure requires net at equal prominence, in a format designed to facilitate comparison, over the same period. A quartile claim, a peer comparison, or a statement of outperformance sits squarely inside that.

Blueprint's page on publishing performance covers what may go on a public website at all, and it is a narrower answer than most managers expect.

What a first-time manager should actually do

With no fund-level record, most of the benchmark conversation is unavailable, and the honest substitutes are stronger than a borrowed comparison.

• Name the cohort you would be judged against, by vintage, stage, geography, and strategy, and say you expect to be measured that way. Knowing your cohort is itself a signal.
• Give the internal benchmarks you hold yourself to, meaning entry ownership, loss assumptions, reserve ratio, and pacing, which are checkable now rather than in six years.
• Where you cite any external data, name the source, the vintage, and the universe, and say what the dataset excludes.
• Do not present a private-equity-wide figure as a venture benchmark, and do not present a benchmark drawn from a different geography as though it applied to yours.
• Where Colibrí Institute research is cited, carry the standard attribution line, and treat its cuts as context for what allocators reward rather than as a grade applied to your fund.

What limited partners are testing

The benchmark question is often a proxy for whether the manager is self-aware about data.

• Does the manager know which cohort they belong to?
• Did they choose the comparison that flatters them, and do they acknowledge choosing it?
• Do they distinguish a young fund's ranking from a settled one?
• Can they name what their benchmark excludes?
• Do they understand that the allocator's real benchmark is often a public index plus an illiquidity premium rather than a peer median?

What a benchmark does not prove

A favourable comparison does not establish skill, particularly early, because a young fund's position moves for years. An unfavourable one does not establish its absence either, for the same reason.

This page states no quartile threshold, no median, and no index return, because those depend on vintage, stage, geography, and dataset, and any figure here would be read as a standard rather than as an example.

This page is educational and general. It is not legal, tax, accounting, or investment advice, and benchmark comparisons used in marketing materials should be reviewed by counsel.