Here is the uncomfortable finding underneath every honest conversation about angel investing: the asset class returns more than the average angel takes from it. The best UK data shows the opportunity is real. It also shows that most people investing in it end up on the wrong side of the median. The distance between those two facts is the subject of this piece, and it has a name. Call it the discipline gap.
The tempting explanation is luck. You backed the wrong ten companies, someone else backed the right ten, and that's early-stage investing. But the returns data doesn't read like a lottery. It reads like a game with a known shape that most players keep misplaying in the same few ways. This isn't the risk primer, which we've covered in the real risks, explained honestly. This is the layer above it: why, given those risks, most angels still do worse than they needed to.
The winners aren't hiding. Most books are just built so they can't catch them.
Why do most angel investors underperform?
Because the returns come from a tail most books are too narrow to catch, and most angels behave in ways that miss it. On the numbers, the majority of individual angel investments lose money while a tiny handful of winners carry the entire result. If your book doesn't hold that winner, the maths of the asset class simply doesn't reach you.
The best UK figures are dated but blunt. NESTA's Siding with the Angels (2009), across 158 UK angels, found a mean return of 2.2x over about 3.6 years. In the same data, 56% of exited investments returned less than the capital put in, and roughly 9% returned 10x or more. That last sliver produced the bulk of the total. The mean is healthy; the median investment is a loss. We run the full distribution in the realistic portfolio maths and the loss ratio, so we won't rebuild it here.
The point for this piece is the shape, not the digit. A positive mean sitting on top of a losing median means the winners aren't optional. They're the entire thesis. And most angels, without meaning to, run a book where the winner can't do its job.
What separates good angels from bad ones?
Mostly it's process, and it breaks in a handful of predictable places. A losing angel usually isn't unlucky; they've built a book that can't hold the winners, or they've handed the diligence over to enthusiasm. The failure modes repeat.
Too few names. If one in ten to one in twenty deals carries the return, a book of five companies is a coin toss dressed up as a strategy. Concentration feels like conviction; on this distribution it's mostly variance. No reserves. Winners keep raising, and an angel with nothing held back watches their best position get diluted round after round while newcomers buy the upside. Reserving for follow-on is planning to back what works; we cover the mechanics in keeping powder for follow-on rounds.
Thin diligence. The clearest single finding in the literature: hours in, returns out. Rob Wiltbank's work for the Kauffman Foundation (2007) found angels who did under 20 hours of due diligence saw a median of about 1.1x, while those who did more than 20 hours saw about 5.9x. Same deals, different work. Chasing heat. Backing a company because a name you respect is in, or the round is oversubscribed, outsources your judgement to a crowd that may be wrong together. Emotional follow-on. Putting more money into a struggling company to protect the original cheque, rather than because the evidence improved, is how the losers quietly eat the reserves the winners needed.
What makes a good angel investor?
The angels who beat the median tend to share a few observable habits, and none of them is a personality trait. They build a portfolio wide enough to give the power law room to work, they hold capital back for follow-on into the companies that earn it, and they run a repeatable process instead of a mood.
These are patterns in the data and in how experienced investors describe their own books, not instructions for yours. What shows up again and again: enough names that a winner can plausibly appear, rather than a heroic bet on one. Reserves set aside deliberately, so the follow-on decision is about evidence and not about whether there's cash left. Diligence done to a consistent standard before the cheque, which the Kauffman figures suggest is the closest thing to a lever this asset class offers. And a fast, clean no, because saying no quickly to the many is what protects the time and capital for the few.
Notice what's absent. There's no claim to pick better than everyone else on gut. The disciplined minority mostly win by building a structure that lets an ordinary hit rate produce an above-average result, then not sabotaging it. The judgement is in the system, not the swagger. The counterweight case, on holding versus selling those winners, is a separate decision with its own traps.
Is angel investing skill or luck?
Both, and the reason discipline is hard is that luck dominates any single deal while skill only shows up across the book. On one investment, a disciplined angel and a reckless one can get the same result, which makes the discipline feel pointless in the moment. It only pays off over dozens of decisions, long after the feedback that would reward it has gone cold.
Several forces pull against it. The J-curve means a young book looks like losses for years before any winner matures, so the discipline is tested precisely when there's nothing to show for it. The hope-versus-outcome gap is measurable: the ERC's Nation of Angels (2015), the most recent UK survey, records what angels expect to make, and those expectations sit well above the 2009 realised losses. Optimism is the fuel of the asset class and also its main way of overpaying. Add social proof, where a respected co-investor substitutes for your own work, and FOMO, where a closing round rushes a decision that deserved the 20 hours. Each of these is rational in isolation and corrosive in aggregate. Discipline is mostly the refusal to be talked out of your process by any one exciting deal.
How do you close the discipline gap?
Treat discipline as a process you run, not a temperament you're born with, because the data rewards the process and is indifferent to the personality. Nothing above is a plan for your money. It's a description of where the returns actually come from and how the average angel gives them back.
The through-line is that the gap opens where structure is missing: too few names, no reserves, too little diligence, follow-on driven by feeling. Each is fixable in advance, by decision rather than by willpower in the heat of a pitch. Which is the quiet good news buried in a grim dataset. The tail exists, and the 5.9x-versus-1.1x split says the work is worth doing. The asset class doesn't ask you to be a better judge of founders than everyone else, only to stop making the same few structural mistakes that put most books on the losing side of the median.
None of this is a recommendation to invest, or a view on any particular company or amount. It's general information about how the asset class behaves and what experienced angels weigh, and your circumstances are your own. Angel investing is high-risk and illiquid; you can lose everything you put in. Before you commit capital, confirm the current rules and reliefs at GOV.UK and take advice from an FCA-regulated adviser. What you do with the pattern is yours to decide.
Frequently asked questions
Why do most angel investors underperform?
Because angel returns come from a small tail of big winners, and most books are built in ways that miss it. The typical failure modes are structural: too few companies to catch the tail, no capital reserved for follow-on, thin due diligence, and follow-on decisions driven by hope rather than evidence. NESTA's 2009 UK data found 56% of exited investments returned less than capital while the mean was 2.2x, so the winners aren't optional, they're the whole result. The cause of underperformance is usually behaviour, not a run of bad luck.
Do most angels lose money?
On most individual investments, yes. The best UK data, NESTA's Siding with the Angels (2009), found 56% of exited angel investments returned less than the money put in. But the portfolio can still come out ahead, because a small tail carries it: about 9% of exits returned 10x or more in that study, and the mean was 2.2x over 3.6 years. So the median deal is a loss while a wide enough book can still make money. Whether any given angel does depends on whether their book catches a winner.
What makes a good angel investor?
In the data, it looks like process rather than personality. The angels who beat the median tend to hold enough names for the power law to work, keep capital back for follow-on into the companies that earn it, do consistent due diligence before the cheque, and say no quickly to everything else. Kauffman Foundation work by Wiltbank (2007) found investors who did more than 20 hours of diligence saw roughly 5.9x versus roughly 1.1x for under 20 hours. These are observed patterns, not instructions, and no habit removes the risk of loss.
Is angel investing skill or luck?
Both, but they show up at different scales. Luck dominates any single deal, where a disciplined and a reckless angel can get the same outcome. Skill only shows up across a whole book, over dozens of decisions, which is why discipline feels pointless in the moment and matters in the aggregate. The Kauffman diligence figures and the NESTA distribution suggest the process is a real lever, but it works by improving the odds across many bets, not by guaranteeing any one.
Is this article investment advice?
No. This is general information about how angel returns behave and what experienced angels weigh, not financial advice and not a recommendation to invest in anything. It gives you the patterns and the trade-offs; the decision is yours. Angel investing is high-risk and illiquid, and you can lose all your capital. Confirm the current rules and reliefs at GOV.UK, and take advice from an FCA-regulated adviser before committing any money.