The Founder Index: What if we used data in early-stage Investing?
- ehargrove5
- 16 hours ago
- 7 min read
Introduction: From Subjectivity to Data
Having reviewed thousands of startup pitches, both as an angel investor and as Managing Director of a prominent startup accelerator, I realized how subjective the investment process can be. While instincts and gut feelings have their place, personal biases inevitably creep in. In an industry where success often hinges on intangible factors like grit and timing, great opportunities can be missed or overestimated based purely on intuition.
This insight inspired a more data-driven approach. One where we measure founder traits and experiences that have historically correlated with entrepreneurial success. By mitigating personal bias and systematically identifying promising founders, we hoped to avoid letting hidden gems slip through the cracks.
We call our data approach, the Founder Index.
The Challenge of Data in Early-Stage Investing
Later-stage startups offer abundant data through revenue, users, and other concrete metrics. By contrast, seed or pre-seed investments lack clear, standardized measures. Investors typically consider market potential, pitch decks, references, and a founder’s résumé - factors that, while valuable, also open the door to industry biases, educational biases, or even superficial judgments like presentation style.
Our approach attempts to fill this gap by going directly to the source: the founders themselves. By analyzing how specific background factors, early company choices, and personality traits correlate with eventual outcomes, we hope to gauge a founder’s likelihood of success or failure, even without conventional financial metrics.
Nature vs. Nurture: What the Data Suggests
A core takeaway from our research is that nature (innate qualities) and nurture (accumulated experiences) both shape a founder’s trajectory:
Nature tends to be crucial in getting a startup off the ground, avoiding total failure, securing initial funding, and building early momentum.
Nurture influences how large that success becomes, determining whether a founder’s exit is modest or transformational. Prior startup experience, years of problem-solving, and familiarity with scaling teams all correlate strongly with bigger outcomes.
This duality explains why a “one-size-fits-all” model struggles to differentiate founders who achieve modest exits (6- or 7-figure) from those who hit larger scale (8-, 9-, or 10-figure exits). Our two-stage modeling acknowledges this nature/nurture distinction, offering more targeted and accurate insights.
Our Data: 60+ Yes/No Questions
To ground our research, we developed a 60+ question survey exploring diverse founder attributes, everything from upbringing to work preferences:
Background & Upbringing:
Did you immigrate from another country?
Did you have siblings growing up?
Did your family struggle financially?
Personality & Work Style:
Do you like to take risks?
Do you set unreasonably high standards for yourself and others?
Do you like to be in control?
Experience:
Have you ever worked at a startup?
Have you started other companies before this one?
Have you managed a team in the past?
Team & Relationships:
Do you have co-founders?
Have you known your co-founder or team for more than three years?
Has your co-founder or team ever let you down?
Why Yes/No?
We intentionally kept questions simple (mostly yes/no) and relatively “immutable”, less likely to change over time. This consistency helps us compare founders who already exited (post-exit founders) with those at early stages. While our data is self-reported (introducing potential bias), we designed the questions so they aren’t obviously “leading.” Founders thus have less incentive to “game” the system. There really isn't a right/wrong to the survey, just a distribution of probabilities.
From V1 to Our Two-Stage Model
The original work on the founder index was done in 2023 and used simple correlation to identify meaningful questions. Founder Index Version 1 (FIV1) scored founders on a single scale. It was decent at distinguishing total failures from founders who achieved some success. However, it struggled to differentiate between moderate exits and significantly larger ones. FoundersEdge then partnered with Gravitate AI , a leading AI data firm, for deeper analysis and an update to the model.
Initial accuracy was around 66%, better than random guessing but still leaving room for improvement. Figure 1 illustrates the distribution of the V1 scores across three groups: Group 1 (founders with no exit), Group 2 (founders with a moderate exit), and Group 3 (founders with a high exit). The blue line in the figure corresponds to no-exit founders, who cluster toward the lower score range, clearly separating them from the other groups. However, the green line (moderate exits) and the red line (large exits) show significant overlap, indicating that V1 could not effectively distinguish between those two levels of success.
Figure 1: FIV1 Score Distribution
The Two-Stage Approach
Through experimentation, Gravitate realized that a single-model approach was missing key subtleties. Some traits correlate strongly with simply “avoiding zero,” while others predict how far beyond zero a founder can go. We therefore split the analysis into two models:
Stage 1: Exit Likelihood
Stage 2: Exit Magnitude
By separating “likelihood” from “magnitude,” we significantly boosted our predictive power. Investors can also tune thresholds for each stage, matching different risk appetites or portfolio strategies.
In this updated model, we see traits that predict higher-value outcomes drive scores much higher. Group 1 is still non-exited founders, Group 2 are founders with exits, and Group 3 are founders with high exits. Now we can more reliably pinpoint those with higher potential.
Figure 2 shows the distribution of the FIV2 score for the three groups. Compared to those in Figure 1, we can now see clear separation between all three groups.
Figure 2: FIV2 Score Distribution
Interesting Correlations & Surprises
A few surprising findings stand out:
Has your co-founder or team ever let you down?
There are both positives and negatives to this question. While no one likes to be let down, co-founder “stress-tests” can build resilience, if they’re addressed proactively rather than ignored.
Do you set unreasonably high standards for yourself and others?
Founders who answered “yes” often have bigger exits, possibly due to strong execution cultures, higher hiring standards, or faster pivots. While this isn’t universally true, shooting for the stars often helps.
Have you ever managed a team before starting this company?
We found some mixed data with managers. We think that first-time founders can adapt more aggressively, while experienced managers take fewer risks. Note that generally more experience is helpful for success.
One question in particular was: “Has your co-founder or team ever let you down?” This doesn’t necessarily mean that being let down causes higher success, only that there’s a notable pattern worth exploring. While it’s difficult to know, our hypothesis on this question is that the startup journey is particularly difficult and founders who have experienced some of the challenges of co-founder relationships are better prepared to go the distance.
Data Thresholds: Narrowing the Field
No predictive model can guarantee success, but using a “score threshold” helps investors focus on the startups most likely to achieve large exits. For example, if an investor only backs companies scoring above a certain cutoff, they’ll likely include many top prospects—but might also miss out on some that could still succeed. The key is balancing how high to set that threshold so you don’t either ignore too many worthy founders or include too many less-promising ones. Figure 4 demonstrates this concept by applying a cutoff score of 6 out of 10.
Figure 4: Setting a score threshold of 6 will likely capture 23% of high-exit founders.
From the perspective of a seed fund or VC manager, these two pie charts illustrate how using a score threshold of 6 can shape your portfolio:
Scores < 6 (Left Pie)
In other words, if your portfolio largely comprises startups scoring below 6, you’ll be exposed to many companies that fail to exit, with only around 1% reaching a substantial exit.
Scores ≥ 6 (Right Pie)
By limiting your investments to startups scoring 6 or higher, you reduce your overall deal flow but significantly increase the proportion of companies that may generate large exits. Nearly a quarter of them could become top performers and another quarter being mid-range performance.
For a seed fund or VC, setting a higher threshold focuses resources on startups with a strong potential for big wins. The trade-off is a narrower investment pipeline—and the possibility of excluding some ventures that could still yield great returns. Depending on your risk appetite and fund portfolio strategy, you may choose a threshold that balances the pursuit of exceptional exits against the desire for broader coverage in your portfolio.
Limitations
Still a Work in Progress
This methodology is continually evolving. Some founders who score low end up achieving stellar exits, and vice versa. Overall, though, our data demonstrates a meaningful advantage in identifying high-exit founders at scale. In the earliest stages where revenue or customer signals are scarce, a bit of “early signal” can still be found in the founders themselves.
Our dataset is based on about 2000 founders and growing. We're constantly updating the model with new founders that further improve the model based on their performance and exit history.
Beyond the data we recognize that external factors, market shifts, funding climates, economic downturns and more play huge roles in startup success, no matter how promising a founder appears. Even exceptional teams and products can stumble. While more experienced founders might pivot and secure funding, there are no sure bets.
Self-Reported Data
Our data is still largely self-reported, which can introduce biases or attempts to “game the system.” Nonetheless, we designed the questions to be neither obvious nor easily manipulated. Founders reading this article wouldn't be able to manipulate their score as the Founder Index relies on 60+ data-points. We’re also working on ways to cross-reference founders’ answers with professional platforms like LinkedIn to enhance accuracy.
Key Takeaways for Founders
Balance Traits and Experience Resilience, control, and high standards help get your startup off the ground, while deeper expertise and relationship-building can magnify overall success.
Invest in Relationship Depth Founders who’ve known their co-founders for years often show higher success rates—especially for large exits—thanks to deeper trust and stronger teamwork.
Aim High Founders who push for audacious goals often see proportional rewards at exit; high expectations can drive a culture of excellence.
Want to learn more?
Founders: Think you have what it takes to build a high-value company? Share your profile with us at FoundersEdge. We invest in founders building exceptional products built on the experience of the founder and the latest in AI. Our data-driven Founder Index can highlight your unique strengths and opportunities.
Data-minded Leaders: Tap into the same AI-powered analytics behind the Founder Index. Our partners at Gravitate AI provide custom AI/ML solutions for smarter, faster, cheaper decisions in businesses of any size.
Comments