Most experienced angel investors know the failure rate is high. They've heard the statistics. They've lost a deal or two. They still believe the difference between good angels and bad ones comes down to judgment — picking better early stage startups, reading founders more accurately, avoiding the obvious mistakes.
The data does not support this.
Two decades of research on angel portfolio outcomes points to something more uncomfortable. The typical struggling angel investor doesn't have a judgment problem. They have a structure problem. Their portfolio is too small and too concentrated, built on assumptions about how venture returns distribute that turn out to be wrong in ways that are hard to see until it's too late.
This article works through the evidence — what the studies actually found, what they mean for any angel investor with fewer than twenty or thirty venture capital investments, and what a structurally sound alternative looks like.
The return distribution is more brutal than most angels realize
The foundational study on U.S. angel returns is the Angel Investor Performance Project (AIPP), conducted by Wiltbank and Boeker for the Kauffman Foundation (2007). It covers 3,097 investments made by 539 angels across 86 groups, with 1,137 realized exits between 1990 and 2007.
Subsequent research has not revised its core conclusions. Nearly twenty years later, it remains the most cited dataset in the field.
Fifty-two percent of all exited angel investments returned less than the original capital invested (Wiltbank & Boeker 2007). Most of those were total losses. Only 7% of deals produced returns greater than 10×. And within that already-small slice, the top 5% of investments generated 57% of all cash payouts across the entire dataset (AIPP re-analysis).
That last number deserves a pause. More than half of all the money ever returned to the angels in this sample came from roughly one in twenty deals. Everything else — the careful due diligence, the portfolio companies that "showed promise," the ones that raised their Series B and looked like they were heading somewhere — generated the remaining 43% collectively.
This is not a story about bad investors. The sample includes experienced, group-affiliated angels with decades of practice. It is a story about how venture returns actually distribute.
The more recent data from AngelList reinforces the picture from a different angle. An analysis of more than 10,000 investor portfolios found that investors with three or fewer startups had a negative median portfolio value (Othman & Koh 2020). The typical small-portfolio investor lost money outright — not merely underperformed. Median IRR climbs roughly nine basis points per additional investment.
The implication is significant: the size of the portfolio matters more than almost anything else about how an investor approaches individual deals.
The gap between group-affiliated and independent angels
One piece of data rarely surfaces in the standard angel investing conversation. Wiltbank's earlier individual-level survey of 106 angels (917 investments, 335 exits) found an average IRR of approximately 10% — well below the 25–30% figures from group-based studies (Wiltbank 2005).
The gap is real and worth understanding. Group-affiliated angels get access to better screening, shared due diligence, syndication, and co-investment structures that improve deal quality and expand their options on follow-on rounds. Independent angels — which describes most people who have written a handful of checks through personal networks — don't have most of those advantages.
If you are an independent accredited investor with a few angel positions and no formal group membership, the 25–30% headline IRR figures do not describe your situation. The 10% figure is considerably closer.
Against a thirty-year annualized S&P 500 return of roughly 10%, that puts the median independent angel roughly even with a low-cost index fund — after locking up capital for years in a high risk, illiquid asset class, while spending hundreds of hours sourcing, evaluating, and monitoring deals.
What the research says about portfolio size
The practitioner conversation about portfolio size usually settles around fifteen to twenty deals. Most angel investors borrowed this number from rules of thumb developed in the early days of formal angel investing — not from simulation or empirical modeling.
The first peer-reviewed paper to rigorously model angel portfolio outcomes using the actual AIPP return distribution was Gregson, Bock, and Harrison's "A Review and Simulation of Business Angel Investment Returns," published in Venture Capital (2017). Their conclusion is considerably less comfortable than fifteen to twenty deals.
You need more than fifty investments to significantly reduce the chance of poor returns (Gregson, Bock & Harrison 2017).
Portfolios in the ten-to-twenty-deal range can show impressive average IRRs while carrying very low median IRRs and high downside risk. The distinction matters. A single outlier can lift the average while every other position in the portfolio sits underwater. For most angels with portfolios in that range, the average is not the outcome they are experiencing.
The 2022 follow-up study by the same authors (When to Hold and When to Fold, International Review of Entrepreneurship) added realistic constraints — staggered investment timing, follow-on rounds, tax treatment — and found that median outcomes fell further once those factors entered the model (Bock, Harrison & Gregson 2022). Angels who build portfolios deal by deal, as opportunities appear, face lower median outcomes and more unpredictable results than cleaner simulations suggest.
Right Side Capital, using distributions calibrated to AIPP, put the minimum at 100 seed investments to adequately reduce single-deal risk. That figure comes from a firm with an obvious interest in larger deal flow, but the direction is consistent with peer-reviewed work.
The practical implication: a portfolio of five to fifteen angel investments — which describes a large share of active accredited investors — sits well below any threshold at which the statistical properties of venture funding work in the investor's favor.
The power law problem, stated plainly
Venture returns do not follow a normal distribution. A small number of outcomes generate the overwhelming majority of value. Most outcomes cluster near zero. This is not controversial; it is the working assumption of every experienced general partner at every serious VC fund.
The consequences for individual angels follow directly. If 57% of all returns come from 5% of deals, and you hold twenty investments, the chance that your specific twenty includes the deal responsible for fund-level returns is low. This is not a judgment call about which early stage companies you backed. It is a math problem applied to a small sample drawn from a skewed distribution.
"How do I pick better?" assumes the problem is selection. For most angels, the more accurate question is: "How do I hold enough positions that the statistical properties of this asset class work in my favor rather than against me?"
Those are different questions. They have different answers.
The first leads to more time spent on due diligence per deal. The second leads to a portfolio construction conversation.
The time cost most angels don't account for
Wiltbank and Boeker's AIPP data found a clear relationship between due diligence effort and outcomes. Deals with twenty or more hours of due diligence were substantially more likely to produce a positive return than those with less (Wiltbank & Boeker 2007). A follow-on study by Forrester et al. (2019) put the mean diligence time at 60.6 hours per deal across 277 angels, with a more conservative mean of 34 hours after adjusting for outliers.
A reasonable working range: 20–60 hours per deal.
Building a portfolio of fifty venture capital investments at that pace requires 1,000–3,000 hours of front-end work — sourcing, screening, due diligence, documentation — not counting the time spent on deals reviewed and passed. Maintaining that portfolio over a decade adds another 100–150 hours per year for monitoring, governance, and follow-on decisions (Wiltbank 2005; Hustle Fund 2025).
A properly diversified angel strategy consumes several thousand hours of work across ten or more years.
Hustle Fund's 2025 analysis of their investor programs estimated that median angel portfolios return 1.0–1.5× over ten years — 0–4% annualized. Against that backdrop, they calculated an effective hourly rate for time invested and found that the purely financial compensation per hour, at median outcomes, is negligible.
Direct angel investing makes sense for learning, for building founder relationships, for the genuine satisfaction many investors get from working closely with early stage companies. As a financial decision for investors who cannot reach the scale and time commitment the data requires, the math is harder to justify.
How long the money actually stays locked up
The Securities and Exchange Commission defines accredited investors, in part, by net worth and income thresholds. What the accreditation standard does not capture is the time horizon required to participate meaningfully in venture funding.
Angels typically enter a company one to three rounds before institutional venture capital firms. Their effective holding period runs longer than the VC's by at least one to two years on any given deal.
The median age of companies at IPO was 13.5 years in 2024, down from a high of 15 years in 2022 but far above the five-to-nine-year medians common in the 1980s and 1990s (Ritter dataset, via MicroVentures 2025; J.P. Morgan Asset Management 2026). The average time to exit via M&A for VC-backed startups has extended to 6.3 years, up from 4.6 years in 2005.
An angel who invests in a pre-seed company today should plan for a ten-to-fourteen-year hold before any realistic path to liquidity in a positive scenario.
In a negative scenario — where the company does not fail cleanly but also does not exit — the position may persist for years with no resolution. Practitioners call these "zombie companies": economically self-sustaining, returning no capital, offering no tax write-off.
The long-term illiquidity premium implied by these timelines should be substantial. Whether it is, for a typical angel portfolio, comes back again to the portfolio size question.
What angels misdiagnose about their own performance
Angels who talk publicly about poor outcomes most often conclude the problem was selection. They backed the wrong companies, chose founders who didn't execute, or entered markets before the timing was right.
There is something true in this. Selection does matter. Experience, pattern recognition, and strong sourcing networks all influence outcomes. But the research is clear that these factors are secondary to portfolio construction — particularly for investors operating below the diversification thresholds the simulations identify.
Three documented cases of experienced angel investors who reconsidered their approach illustrate the pattern.
Halle Tecco, founder of Rock Health and a faculty member at Harvard Medical School, invested in 54 companies over fifteen years. Realized returns were deeply negative on the portfolio as a whole. She ultimately shifted to LP positions in VC firms.
Tucker Max allocated $1.2M across roughly eighty companies. His LP investments through ATX Seed Fund and Evolve VC tracked toward substantially better outcomes than his direct angel bets. His conclusion: being an LP in a VC fund is a better structure than direct angel investing for most accredited investors.
Sam Dogen invested $60,000 in a single early stage company and received $91,902 back after nearly a decade — a 5.1% annualized return. His takeaway aligned with the other two: investing in venture capital firms through a managed fund structure produces better long-term results than picking individual deals.
None of these investors lacked judgment. All of them ran into the structural properties of the asset class in exactly the way the data predicts.
What a structurally sound venture allocation looks like
Research on diversified, multi-manager fund-of-funds structures shows a meaningfully different risk profile than individual angel portfolios.
Diversification benefits in venture portfolios level off at approximately 20–25 underlying VC funds (Dompé 2019; Gredil, Liu & Sensoy 2024). Diversified fund-of-funds strategies cut the chance of capital loss to approximately 8%, compared to approximately 20% for concentrated approaches, net of fees (Vanguard 2025).
That comparison — 8% vs 20% chance of capital loss, net of fees — is the clearest summary of what structure delivers in this asset class.
Venture remains a long-term, high risk, illiquid investment. A fund-of-funds structure does not change that. What it does is absorb the single-deal risk that individual portfolio concentration creates, and replace it with broad exposure to the asset class itself.
The top-quartile to bottom-quartile IRR spread in venture capital remains 53 percentage points (Harris, Jenkinson, Kaplan & Stucke). Picking the right general partner matters enormously. A multi-manager structure distributes exposure across VC firms within a defined ecosystem rather than concentrating on a single fund or a handful of direct bets — addressing both manager selection risk and diversification within a single allocation.
Whether this structure fits a given investor depends on their capital, time horizon, and what they are actually trying to accomplish with their venture allocation. For investors whose primary goal is financial participation in the asset class — rather than deal involvement, founder relationships, or the learning that comes from direct investing — the structural question is worth asking honestly before the next angel check gets written.
A note on the data
The research cited throughout this article has known limitations. Angel return studies face survivorship bias: successful angels respond to surveys more readily than unsuccessful ones. Wiltbank and Boeker tested for this within the AIPP sample and found no significant difference in returns between high- and low-response groups, suggesting the bias exists but does not drive the results (Wiltbank & Boeker 2007). DeGennaro and Dwyer's Federal Reserve working paper (2010) reached the same conclusion.
Group-based headline IRRs — in the 25–30% range — likely overstate what most angels actually experience. The individual-level figure of approximately 10% from Wiltbank's 2005 study is probably closer to the ground truth for investors working outside formal group structures. Both figures are useful. The gap between them is itself informative.
Esinli Capital is a venture capital fund-of-funds for accredited investors. Each fund invests across multiple VC managers and hundreds of underlying portfolio companies within a single innovation ecosystem. The Tel Aviv Fund is open for reservations ahead of Q2 2026 first close.
Learn more at esinli.com — no commitment, no obligation, two minutes.
Sources
Wiltbank, R. & Boeker, W. (2007). Returns to Angel Investors in Groups. Kauffman Foundation / Angel Capital Education Foundation.
Wiltbank, R. (2005). Startup Angels. EquityNet.
Wiltbank, R. (2009). Siding with the Angels. NESTA.
Othman, A. & Koh, J. (2020). AngelList portfolio analysis, 10,000+ investor portfolios.
Gregson, G., Bock, A. & Harrison, R. (2017). A Review and Simulation of Business Angel Investment Returns. Venture Capital, 19(4).
Bock, A., Harrison, R. & Gregson, G. (2022). When to Hold and When to Fold: Simulating Portfolio Returns to Angel Investing in Early Stage Ventures. International Review of Entrepreneurship, 20(1).
Forrester, R. et al. (2019). Angel investor due diligence practices. Survey of 277 angels.
DeGennaro, R. & Dwyer, G. (2010). Expected Returns to Stock Investments by Angel Investors in Groups. Federal Reserve Bank of Atlanta.
Harris, R., Jenkinson, T., Kaplan, S. & Stucke, R. (2017). Financial Intermediation in Private Equity. NBER.
Dompé, A.C. (2019). Diversification Study. CAIA.
Gredil, O., Liu, Y. & Sensoy, B. (2024). Diversifying Private Equity. SSRN.
Vanguard (2025). Benefits of a Fund-of-Funds Strategy in Private Equity.
Ritter, J. (2024 dataset). IPO data, University of Florida.
Hustle Fund (2025). Is Angel Investing Worth It?
J.P. Morgan Asset Management (2026). Private Markets Outlook.
