What's the Point of Benchmarking VC Performance?

LPs are frequently presented with funds claiming to be “top-decile” or “top-quartile”. While experienced investors tend to discount such marketing language, many still place significant trust in VC benchmarks and the perceived objectivity they offer.

What’s surprising is how rarely these benchmarks are questioned. Their reliability, accuracy, and practical usefulness are often taken for granted, even though they play a significant role in shaping investment decisions.

In reality, the KPIs we track and the datasets we benchmark against are far less reliable than most assume and flexible enough for almost any manager to find a way to appear “top-decile”.

This blog post takes a critical look at the most common VC performance metrics and benchmarks, explaining why they often measure less than we think and how relying on them can distort, rather than clarify, the true picture of fund performance.

Which KPIs Are You Benchmarking?

TVPI (Total Value to Paid-In Capital)

TVPI and MOIC are widely used performance metrics and, in the short term, serve as early indicators of potential value creation. However, their reliability depends heavily on the valuation methodology applied by VC funds, which can vary significantly across the industry.

The most common approach is to value positions at the most recent priced round. More conservative investors may choose to only mark up positions after a round led by an external investor (as opposed to an internal round). At the same time, more permissive VCs may even recognise mark-ups from SAFEs or convertibles.

Mark-downs and write-offs are even less standardised. Unless a company is liquidated or sold, applying markdowns is usually done voluntarily. Some investors are quite ruthless in marking down underperforming companies. In contrast, others may strategically delay write-offs until a valuation increase elsewhere can offset any negative impact on the fund’s TVPI. It is not uncommon for LPs with exposure to multiple funds invested in the same company to see it reported at different valuations.

Over time, as more investments are realised, valuation differences fade and TVPI becomes more reliable, gradually converging toward DPI.

IRR (Internal Rate of Return)

IRR can be a useful performance metric when comparing returns across different asset classes over a longer period of time, but due to excessive financial engineering, especially in the private equity industry, it has steadily been losing its utility for investors.

The main weakness of IRR is its time-sensitivity: early cash flows have a disproportionately high impact on the performance metric, which means that two funds with the same money multiple can show vastly different IRRs depending on the timing of cash flows.

While venture capital is less exposed to cash flow structuring, it faces a similar distortion risk through valuation practices. The same methods that inflate TVPIs can also inflate IRRs. In the early years of a VC fund, managers may delay write-offs and aggressively mark up portfolio companies to boost reported IRRs. In later years, significant “paper gains” can push IRRs even higher, only to collapse if exits fail to materialise.

DPI (Distributions to Paid-In Capital)

DPI is the ultimate performance metric for investors. Unlike TVPI, it actually measures realised returns, and unlike IRR, it leaves little room for manipulation.

Its main limitation, however, is timing. DPI only becomes relevant in the later years of a fund’s life, since distributions usually take a long time to materialise. Early DPI can be misleading, as it often reflects small exits (1x–5x) or partial returns that, unless recycled, don’t meaningfully impact total fund returns.

In fact, significant early DPI is rarely a positive signal. VCs only have a handful of potential “fund returners” in their portfolio; therefore, an early exit in a larger company may suggest that the fund sold too soon or that the company, while successful, fell short of the scale needed to return the fund on its own.

Which Data Are You Benchmarking Against?

Data Service Providers

The primary sources of VC performance benchmarks are commercial data providers such as PitchBook and Cambridge Associates. Although investors often pay considerable amounts for access to their “proprietary” datasets, the underlying benchmarks remain incomplete and prone to bias.

According to David Clark (CIO of VenCap), PitchBook, the largest VC data provider, only has return data of about 5% of all sub $100 million funds that were raised from 2010 to 2020. Similarly, Cambridge Associates has less than 4%. Such small samples are not statistically meaningful for drawing industry-wide conclusions.

And that’s just the quantity problem. The quality of data is equally concerning. Some inputs may come from public institutions like CalPERS, which must disclose performance and are reasonably accurate. However, much of the data on these platforms is voluntarily self-reported by funds, meaning that strong performers (or those wishing to appear as such) are far more likely to submit their results, which skews the dataset upward. Worse, self-reported numbers are difficult to actually verify.

Fund Admins

Fund administrators like AngelList and Carta can generate benchmarks from anonymised performance data of the VC funds they service, in addition to public sources. This gives them an edge in accuracy since the data isn’t self-reported, but also means that their benchmarks are limited both by scope and by the characteristics of their customer base.

Both Carta and AngelList have significant U.S. exposure, and while they cover a relatively large universe of funds, their datasets still represent only a fraction of the total market, omitting a substantial portion of U.S. funds and most of the non-U.S. fund universe. AngelList’s data, in particular, also tends to emphasize micro-VCs and solo GPs over more traditional funds. These benchmarks can therefore be insightful but are far from representative of the entire VC industry.

Government LPs

Government LPs, particularly in Europe, are the most active participants in the market. According to Sifted, in 2023 government-backed LPs were the largest LP group, accounting for up to 37% of all VC funding raised. Their scale gives them a broad view of the market, but also introduces significant bias.

National fund-of-funds programs typically focus narrowly on their own geographies, while Pan-European initiatives like the NATO Innovation Fund (NIF) or the European Investment Fund (EIF) are driven by political mandates, often with specific industry, technology, or geographical priorities.

These LPs also tend to disproportionately favor large, generalist funds in mature ecosystems, ignoring micro-VCs, solo GPs, or emerging ecosystems. As a result, the benchmarks they are able to produce paint a rather skewed picture of the market.

So What’s the Point?

​​If KPIs can be inflated, misleading, or manipulated, and data sources are incomplete, biased, or statistically irrelevant, what is the point of benchmarking VC performance at all?

The short answer is: there is no better alternative. In a notoriously opaque asset class like VC, benchmarks, however flawed, remain one of the few tools investors can use to gain a directional sense of performance.

The problem arises when investors place too much weight on benchmarks and develop unrealistic expectations of the asset class as a result. Benchmarks are a largely backward-looking tool, useful for context but limited in their ability to inform forward-looking decisions. They should not be the basis for LP re-up decisions or part of GP fundraising pitches.

Rather than relying on any single benchmark in isolation, investors should view them collectively and in the context of broader market dynamics. At best, benchmarks provide a rough industry overview and should be treated as a compass rather than as GPS.

Thank you for taking the time to read our blogpost! For more information about Multiple Capital, please visit our website or reach out to our team directly.