Venture capitalists evaluate roughly 200 companies per year and invest in about four. Each deal takes 83 days on average, with 118 hours devoted to due diligence. The filter is severe, and the stakes are high: 90% of VC profits come from just 5% of VC funds. Fewer than 40% of VCs with any investments ever produce a single successful outcome.
For advisors and limited partners trying to assess whether a firm can separate viable startups from doomed ones, two evaluation axes dominate every conversation: the technology and the team. These are the jockey and the horse, the oldest framing in venture investing, and the research on how VCs actually weigh them reveals both a rough consensus and serious contradictions.
The Team Question: Jockeys, Horses, and the People Behind the Product
The “jockey versus horse” debate dates to the earliest days of venture capital. Tom Perkins of Kleiner Perkins championed technology as the primary driver. Don Valentine of Sequoia Capital believed markets were paramount. Arthur Rock of Davis and Rock put his faith in people above all else. The people-first school has since become dominant. According to a Crunchbase analysis, 53% of early-stage VCs rate the team as their most important consideration, compared to 10% for business model and 6% for market opportunity.
But the consensus has cracks. Research tracking startups from inception to IPO found that business quality was more predictive of success than founding team characteristics. The study, cited in the Crunchbase analysis, suggests VCs may consistently make “predictably bad investments” by over-indexing on founder attributes. The problem, according to the same analysis, is that many VCs interpret “founder first” as superficial pattern matching on education, prior employment, or demographics, mistaking correlation for causation.
A 2023 study published in Nature Scientific Reports analyzed 21,187 startups and found that founder personality traits significantly predict venture success. The researchers identified six distinct founder personality types rather than a single “founder type,” and found that larger, personality-diverse teams showed increased likelihood of success. Key traits distinguishing successful founders included openness to adventure, lower modesty, and higher activity levels. The study emphasized personality diversity within teams as a novel dimension of team composition that influences performance.
Coachability has emerged as a specific trait investors increasingly emphasize. A 2025 study in Cogent Business and Management developed an Entrepreneur Coachability Scale based on surveys of investors, coaches, and entrepreneurship professors. The study found that coachability and related competencies, particularly relationship and implementation skills, are vital in investment decisions. Entrepreneurs demonstrating these skills were more likely to secure funding.
Reading the Technology: Defensible Moats Versus Thin Wrappers
Technology assessment is the second pillar of due diligence, and it presents different challenges from team evaluation. The core question is whether the technology creates a defensible advantage or can be easily replicated.
A 2024 article in Fast Company identified “thin wrappers” as a primary red flag for AI-era startups. Founders who place a user interface on top of third-party models, with no proprietary data, workflow integration, or defensible moat, face immediate skepticism. The concern is that switching costs are low and competitors can launch copycats quickly. Investors want to know what remains valuable when the next model release drops. If the moat is “we use GPT too,” the response is skepticism.
The concept of a defensible technology moat extends beyond AI. A startup’s intellectual property, proprietary datasets, network effects, and technical complexity all factor into whether the technology can sustain competitive advantage. The Stanford GSB research on the “venture mindset” emphasizes that VCs run headlong into uncertainty and embrace contrarianism because more than 50 years of VC data show this approach works. The best investors evaluate founders through the lens of their business, looking at the product to understand the humans behind it, rather than pattern matching on superficial attributes.
Y Combinator, which has funded over 4,000 startups, makes investment decisions with 10-minute interviews. Sam Altman, then president of YC, emphasized “clarity of vision” and “the non-obvious brilliance of the idea” as key elements, alongside determination and communication skills. The approach expands the evaluation aperture, creating more data to drive better decisions than any pattern-matching exercise.
Red Flags and Warning Signs
Practitioner literature identifies a consistent set of founder behaviors that signal trouble. A Fast Company survey of VC investors identified claiming to have no competitors as a credibility killer. Building thin wrappers without proprietary technology, overselling prior experience, and being evasive about key metrics all surfaced as common red flags.
A Medium post collecting “deal breakers” from top VCs cataloged additional signals: founders not grounded in reality about revenue projections, treating junior staff poorly, demanding expensive restaurants for meetings, claiming to be “closing next week” to manufacture urgency, refusing to share materials digitally, and being overly dilution-sensitive. Each behavior signals something deeper: poor judgment, weak ethics, or an inability to manage relationships.
The CB Insights Mosaic score, a proprietary measure of private company health scored 0 to 1,000, provides a quantitative warning system. Among companies with full 12-month data that subsequently shut down, 72% saw their Mosaic score decline in the year before death, with scores dropping 15% on average. Partnership activity dropped 44% in the final 12 months compared to the prior year. Two-thirds of companies were shrinking headcount in the six months before death.
What Failure Looks Like: The Post-Mortem Data
Running out of capital topped the list at 70%, but CB Insights noted this is almost always the final cause of death, not the root problem. The more revealing causes were poor product-market fit, bad timing, and unsustainable unit economics. The analysis covered 385 companies for which failure reasons could be identified.
Two-thirds of product-market fit failures were early-stage companies that never found a market. But 20 Series B or later companies also cited poor product-market fit as a primary cause. These later-stage companies had raised on early traction that never expanded into a real market. Zume, which raised $446 million through Series C and pivoted from robot-made pizza to sustainable packaging, is one example.
The median time from last fundraise to shutdown was 22 months. Over half the companies in the dataset died within two years of their last raise. Nearly a quarter had been “walking dead” for over three years since their last raise. CB Insights identified nearly 50,000 VC-backed startups that had not raised funding since the start of 2023.
The Limits of Prediction
The evidence on whether VC due diligence actually predicts success is mixed. An NBER working paper by Gompers, Kovner, Lerner, and Scharfstein argues that a large component of success in entrepreneurship and venture capital can be attributed to skill, pointing out that entrepreneurs with prior success are more likely to succeed again. But the same paper acknowledges the role of luck, and notes that funding by more experienced VC firms enhances success chances only for entrepreneurs without a successful track record.
Andrew Chen, a general partner at Andreessen Horowitz, has written about the difficulty of startup prediction. With only 10 to 15 breakout companies per year, the signal-to-noise ratio is low, and pattern recognition based on past winners may miss the outliers that generate returns. The Holloway Guide to Raising Venture Capital documents how pattern matching produces systemic bias, as VCs replicate past investment patterns that exclude underrepresented founders. Research from Harvard Kennedy School found significant evidence of pervasive gender bias in VC funding decisions.
The Bureau of Labor Statistics reports that approximately 20% of private-sector businesses fail in their first year, rising to about 50% by year five. These figures cover all new businesses, not just VC-backed startups. The BLS data, confirmed by multiple sources including Investopedia and SCORE, provide a baseline for understanding the default risk that due diligence is attempting to overcome.
The Bottom Line
The research paints a picture of an industry that has developed sophisticated evaluation frameworks but still struggles with fundamental prediction problems. The team is the most important stated factor in VC decisions, supported by multiple surveys of practitioners. But the evidence on whether team quality actually predicts success is contested, with at least one major study finding that business quality outperforms team characteristics in predicting IPO outcomes. The technology must work, must be defensible, and must address a real market, yet 43% of VC-backed failures stem from the inability to find that market.
For advisors evaluating VC firms, the questions that matter are not just “what do you look for” but “how do you look for it.” The VCs who rely on superficial pattern matching, who skip reference checks, who accept non-GAAP financials at face value, or who fail to verify that the technology actually works are the ones whose portfolios will concentrate failure.
The ones who evaluate founders through their products, who stress-test unit economics, who run rigorous reference calls, and who understand the sector-specific balance between team and technology weight are the ones positioned to find the outliers that drive returns. The difference between trash and treasure, in the end, is less about the companies being evaluated and more about the rigor of the evaluation itself.