When VC Theses Diverge From Portfolios

A venture firm can describe itself as sector-agnostic and operator-focused while repeatedly backing a narrower set of companies.

Beige Media made that point directly in a 2026 analysis, noting that a fund's stated investment thesis and its actual portfolio can diverge substantially. That gap matters because founders, limited partners, and other market participants often treat public language as a proxy for how a firm will deploy capital.

In practice, the portfolio usually gives a cleaner answer. A stated thesis explains how a firm wants to be understood. The deal record shows what it actually funds, at what stage, in which sectors, through which networks, and under what constraints.

That distinction is better described through the finance language of stated and revealed preferences than through the broader idea of preference falsification. The strongest evidence in this area comes from research on venture selection, specialization, matching, and style drift rather than from fund marketing materials.

Key Findings

Public VC theses are often designed for fundraising, LP communication, and deal flow management rather than as strict descriptions of decision rules.
Actual investment behavior reveals repeat patterns in sector, stage, geography, and founder-investor matching that are often broader or narrower than website language suggests.
Survey evidence shows venture firms review about 101 opportunities per investment, while team quality remains the most important selection factor.
Research links investor specialization to stronger outcomes, helping explain why many supposedly broad firms still cluster around familiar sectors.
Style drift describes sustained departures from a fund's mandate, and evidence suggests initial drift can help while follow-on drift can weaken outcomes.
Founders and limited partners can evaluate firms more accurately by studying portfolios, reserves, and repeat deal patterns before relying on stated theses.

Why Public Theses Stay Broad

Public investment theses serve several institutional purposes beyond screening investments.

In a 2023 guide, Carta describes the investment thesis as the document that outlines how a fund plans to use capital to generate returns. It includes target stages, industries, geographies, check sizes, reserves, and portfolio construction.

The same guide presents the thesis as a tool for differentiation, LP communication, and day-to-day fund management. That creates a basic incentive for firms to write theses that are coherent but not overly restrictive.

A broad public thesis can widen inbound from founders, leave room for opportunistic deals, and preserve flexibility as markets change. It can also help a firm present a stable identity to limited partners even when the actual investment process relies on narrower filters.

The selection funnel helps explain why this happens. Greymoor Capital cites the large survey by Paul Gompers, William Gornall, Steven Kaplan, and Ilya Strebulaev, which states that firms review about 101 opportunities for each investment.

In that context, broad language at the top of the funnel can be rational. It keeps access open while harder screens operate later.

That does not make the thesis useless. It remains relevant as a statement of target market, a signal to LPs, and a rough guide to what the partnership wants to emphasize.

But it does mean the thesis should be treated as one input to diligence rather than the final description of strategy.

How Venture Firms Actually Make Decisions

The best-known evidence on venture decision-making comes from the survey published as NBER Working Paper No. 22587 in 2016. The authors surveyed 885 institutional venture capitalists across 681 firms.

They found that, in selecting investments, VCs viewed the management team as more important than business characteristics such as product or technology.

The same body of survey evidence is often summarized through a more specific statistic: the management team was rated important by 95% of VC firms and ranked as the most important factor by 47%.

Those figures are widely cited and align with the paper's broader conclusion that team quality dominates selection decisions. Even without a detailed website thesis, that already implies a narrower practical filter than many public narratives suggest.

Other factors also matter, including market size, product, business model, and fit with the fund.

But the ordering matters. If team judgment, pattern recognition, and prior familiarity dominate actual selection, then the operative strategy will often be more concentrated than a firm's brand language indicates.

This is where revealed preferences become useful. A firm's repeated choices show which sectors it understands best, which founder backgrounds it trusts, and which geographies it can reach efficiently. Those choices are usually more informative than a general statement about being founder-friendly or category-agnostic.

What the Portfolio Reveals

One of the clearest recurring patterns is specialization. In the 2006 paper Specialization and Success: Evidence from Venture Capital, Paul Gompers, Anna Kovner, Josh Lerner, and David Scharfstein found a strong positive relationship between the degree of specialization by individual venture capitalists and investment success.

The paper also found that increasing firm specialization enhanced performance when holding individual investor specialization constant.

That result helps explain a common mismatch between language and behavior. A firm may present itself as broad because broadness increases access and protects optionality.

Yet if specialization improves sourcing, evaluation, and post-investment judgment, the realized portfolio may still cluster around a narrower set of sectors or founder profiles.

Geography often reveals the same pattern. Even when firms describe themselves as open to founders everywhere, venture investing remains heavily shaped by proximity, repeat networks, and local information.

A portfolio concentrated in a few cities may say more about a firm's true sourcing network than any statement about being national or global.

The strongest practical signal is not one isolated deal but repetition. If a fund's first checks repeatedly go to enterprise software founders from similar technical backgrounds, that is a better guide to future fit.

The same is true for companies within a small set of regional hubs.

Affinity, Heuristics, and Hidden Filters

Some of the most important filters do not appear in public theses at all.

A 2025 Journal of Financial and Quantitative Analysis summary of the paper Alumni Networks in Venture Capital Financing reported that one-third of VC deals involved a founder and investor from the same university.

It also stated that venture capitalists were more likely to invest in, and place larger bets on, startups with founders from their alma mater.

That finding is important because it captures a real source of selection that sits outside standard thesis language. Very few firms advertise university affinity as a core investment rule.

Yet the deal record may show that educational ties shape sourcing, trust, diligence speed, and conviction.

A related paper published online in 2026 by Strategy Science, A Familiar Face: Measuring Visual Similarity in Venture Capital by Jane Wu, found that facial similarity between entrepreneur and investor was associated with a higher likelihood of investment.

The paper also states that the relationship was more pronounced when uncertainty was higher. It notes that visually similar entrepreneur-investor investments underperformed in terms of successful exits.

That result should be read carefully. It does not suggest that venture decisions are reducible to appearance. It does show that in early-stage settings, soft heuristics can influence allocation.

Those heuristics are rarely visible in a thesis memo, but they can still shape who gets funded.

Taken together, these studies suggest that portfolios capture not only explicit strategy but also embedded behavior. Some of that behavior reflects expertise. Some reflects network structure. Some reflects bias or shortcut judgment under uncertainty.

All of it is more visible in applied investment patterns than in public prose.

When Divergence Becomes Style Drift

Finance has a more formal term for a pronounced gap between stated strategy and realized behavior: style drift.

In venture and private equity, style drift generally refers to investments that fall outside a fund's stated mandate on dimensions such as stage, industry, or geography.

The 2024 ECGI working paper The Performance Puzzle in Venture Capital and Private Equity Style Drifts finds that initial style drift investments can improve outcomes relative to non-drift investments in some settings.

In its baseline regressions, the paper reports that an initial style drift investment was associated with a 2.22 to 2.43 percentage point increase in the probability of an IPO exit.

The same paper draws a more cautious picture for follow-on behavior. It argues that follow-on style drift investments can produce weaker outcomes than initial drift investments and discusses escalation of commitment as one reason.

In other words, moving outside the mandate once may reflect opportunistic judgment. Repeating drift through later rounds can reflect incentive problems or attachment to earlier decisions.

That distinction matters for both founders and LPs. A single exception to a public thesis does not automatically show incoherence. A repeated pattern of off-mandate investing can indicate the stated thesis no longer describes the real strategy.

This is also where incentives enter the picture. Fund managers need to preserve reputation, maintain access to future fundraising, and defend prior decisions. Those pressures can make divergence from the stated thesis endogenous to the business model.

How to Read a Venture Firm More Accurately

For founders, the practical lesson is simple. Use the thesis as a first screen, then test it against the portfolio.

The useful questions are empirical: What sectors dominate the first checks? Which stages recur most often? In which geographies does the firm repeatedly lead or co-lead? What founder backgrounds appear across the portfolio? Which partner biographies line up with the actual companies funded?

Those questions matter because founders are not only choosing capital. They are choosing attention, pattern recognition, and future support.

A firm whose public language sounds broad but whose deal history is narrow may still be a strong fit if the startup matches that narrower pattern. It may be a weak fit if it does not.

For limited partners, the stakes are larger and the audit should be deeper. A fund's thesis should be compared with deployment history, reserve strategy, concentration, partner sourcing patterns, and the share of deals that sit outside the stated mandate.

Carta's framework is useful here because it explicitly ties the thesis to portfolio construction, check size, geography, reserve policy, and differentiation. Those are all variables that can be checked against actual behavior.

This is especially important when a fund markets itself as generalist while individual partners appear highly specialized. The gap is also revealing when a stated geographic range masks concentration in a small network of cities, universities, or founder archetypes.

In those cases, the gap between thesis and portfolio is not just descriptive. It changes how a founder should pitch and how an LP should underwrite the manager.

The central point is not that public theses are false. It is that they are incomplete. They describe aspiration, positioning, and institutional self-presentation. Portfolios describe execution.

When the two align, the thesis helps explain the record. When they diverge, the record usually carries more analytical weight.

In venture capital, the public story remains useful, but the deal list is still the closest thing to audited behavior.