Essay · The Losses
What losing my first angel investment taught me about the difference between believing in a person and underwriting a deal · Juan Vegarra · July 2026
The first angel check I ever wrote, I lost in full. Two hundred thousand dollars, gone, on the first independent investment decision of my life. I have written a good deal on this site about how to read a market before the consensus does, how to spot the signals that a category is about to reorder, how to build the engine that turns a thesis into a company. All of it is true, and all of it was earned. But a body of work that only ever describes the method working is telling you half of a life, and it is the less useful half. The method came from the times it did not work. Here is the first of those times, told at the same length and with the same care, because I would rather you learn it the way I did not, which is before writing the check instead of after.
It was right after I left Microsoft, around 1999. I had spent the better part of a decade there, I had done well, and for the first time I had capital that was mine to deploy rather than a salary to live on. I set aside a portion of it for high-risk, high-reward bets, the aggressive slice of what I thought of as my retirement, the part meant to compound into something over the decades ahead. That framing matters, because it is easy to read a two hundred thousand dollar loss by an ex-Microsoft manager as tuition a wealthy man could shrug off. It was not that to me. It was a meaningful piece of the money I was counting on for the long run, and losing it was going to hurt, and it did.
The company was an education technology startup, in a year when nobody yet had a framework for education technology at all. The founder was someone I knew, liked, and believed in. That sentence is the entire essay, and at the time I could not hear the problem living inside it.
I did not underwrite a company. I underwrote a person, and told myself I had done the harder thing.
Everything I could point to as diligence was really a description of the founder. He was smart, and I had seen him be smart. He was driven, and I had watched him work. He explained the vision with a fluency that left me nodding. I walked away from those conversations feeling that I had done my homework, because I had spent hours on it and come away more convinced. What I had actually done was spend hours letting someone I already trusted deepen a conviction I had arrived with. That is not diligence. That is the feeling of diligence, which is a far more dangerous thing, because it is indistinguishable from the real one right up until the money is gone.
I never built the model that would have told me how many students at what price it took to reach a sustainable business, or whether that number was reachable with the capital on hand. I never mapped who actually pays for education technology, which turns out to be one of the hardest questions in the sector and the one that quietly kills most of the companies in it. I never asked what had to be true for this to work, and then gone looking, honestly, for the evidence that it was not true. I did none of the things I now consider the floor, not the ceiling, of an investment decision. I skipped all of it, and I skipped it precisely because I liked and trusted the person asking. The trust did not supplement the analysis. It replaced it.
Belief in a person is the most expensive substitute there is for underwriting a deal, because it feels like the most rigorous thing you could possibly do.
The company did not make it. The specifics of how it came apart matter less than the fact that when it did, I had no independent read on what was happening, because I had never built one. I had outsourced my judgment to my affection for the founder, so when the affection was tested I had nothing else to fall back on. I lost the entire position. And the honest, uncomfortable part is that I could not even be usefully angry, because the error was not his. It was mine. He ran at a hard problem and lost, which is what founders do most of the time. I had handed him money without doing the one job an investor has, which is to decide clearly whether the bet is worth making before making it.
It would be more flattering to file this under a warm heading. I trusted too much. I believed in people. I led with my heart. Every one of those framings is a quiet way of turning a discipline failure into a character virtue, and I want to resist all of them, because the flattering version is useless to you and to me. The truth is plainer and less comfortable. I confused knowing someone with underwriting them, and those are different activities that happen to feel similar in the moment.
They feel similar because both produce conviction, and conviction feels like knowledge. When you have spent real time with a capable founder, you end up certain, and certainty is a sensation the mind does not easily distinguish by source. The certainty that comes from having built the model and stress-tested the assumptions feels exactly like the certainty that comes from having simply spent an afternoon with someone impressive. The first is earned and the second is borrowed, and you cannot tell them apart from the inside. That is the trap, and it is not a beginner's trap that you age out of. It is the permanent condition of anyone who invests in people they have come to like, which, if you do this long enough, is most of the people you invest in.
There is a version of this story where I let myself off the hook, and I want to close it off before someone offers it to me. It goes: you were early, edtech was unproven, nobody had the tools in 1999, you could not have known. Part of that is even true. It was a genuinely hard sector to read, and I was making my first decision without a framework anyone had handed me. But the excuse does not survive contact with the actual failure. I did not lose because the sector was hard to read. I lost because I never seriously tried to read it. The tools I lacked were not exotic. A basic unit-economics model, a clear map of who pays, an honest inventory of what had to be true, these existed and were learnable, and I simply had not done the work of learning them yet. Being early is a defensible reason to lose. Not doing the work is not, and I will not dress one up as the other.
A couple of years later I went back to school for an executive MBA, and for the first time I was handed the actual apparatus of underwriting a business rather than admiring one. I learned to build the model that turns a story into a set of numbers that either close or do not. I learned to separate the person from the plan, to hold that a founder can be genuinely excellent and the deal still be a poor one, and that the two judgments have to be made independently or they contaminate each other. None of it was complicated. Most of it, I could have taught myself years earlier if I had known to look. The lesson of the loss was not that investing is hard, though it is. The lesson was that I had been skipping the part of it that is actually the job.
What I did with that, over the years since, was turn the discipline into something external, so it would not depend on how I happened to feel in the room. Feelings in the room are exactly the thing that failed me the first time, so the fix could not live in my feelings. I built a scorecard, a fixed set of questions every potential investment has to answer before I am allowed to write a check, regardless of how much I like the founder. I have written separately about that scorecard and how it works. What matters here is where it came from. It came from this loss. The scorecard is the machine I built so that the two hundred thousand dollars bought something, and so that liking someone can never again quietly stand in for having done the work.
The guardrail is in the past tense. The bias it guards against is not. I still have it, and I always will.
That distinction is the one I would most want a younger version of myself to hear. I did not fix the flaw. You do not fix this flaw. I am, to this day, more inclined to believe in people I like than the evidence alone would support, and if I ever tell you I have grown out of that tendency you should discount everything else I say, because I will be lying or fooling myself. What changed is not the instinct. What changed is that I no longer let the instinct near the decision unsupervised. The scorecard sits between my affection and my capital, and it asks the questions my affection would rather skip. That is not a cure. It is a structure I built because I know the cure does not exist.
The reason to publish a loss and not just the wins is not confession for its own sake. It is that the wins, told alone, teach almost nothing transferable. A string of good outcomes reads as luck or genius, neither of which you can copy. The loss is where the actual instruction is, because you can see exactly which step I skipped and exactly what it cost, and you can decide not to skip it. If you take one thing from this, let it be the discipline to notice when your conviction is borrowed rather than earned. Ask yourself, before the next check, before the next hire, before the next partnership, whether you have underwritten the decision or merely spent time with someone who made you want to. They feel the same from the inside. Only one of them protects you. I paid two hundred thousand dollars to learn the difference, and I am handing it to you here for free, which is the closest thing to a return that loss will ever produce.
Juan Vegarra is the author of An Outsider's Playbook (forthcoming). The views here are his own. More from the Notebook · Continue the conversation
READ NEXT