March 2009(This essay is derived from a talk at AngelConf.)When we sold our startup in 1998 I thought one day I'd do some angelinvesting. Seven years later I still hadn't started. I put it offbecause it seemed mysterious and complicated. It turns out to be easier than I expected, and also more interesting.The part I thought was hard, the mechanics of investing, reallyisn't. You give a startup money and they give you stock. You'llprobably get either preferred stock, which means stock with extrarights like getting your money back first in a sale, or convertibledebt, which means (on paper) you're lending the company money, andthe debt converts to stock at the next sufficiently big fundinground. [1]There are sometimes minor tactical advantages to using one or theother. The paperwork for convertible debt is simpler. But reallyit doesn't matter much which you use. Don't spend much time worryingabout the details of deal terms, especially when you first startangel investing. That's not how you win at this game. When youhear people talking about a successful angel investor, they're notsaying "He got a 4x liquidation preference." They're saying "Heinvested in Google."That's how you win: by investing in the right startups. That isso much more important than anything else that I worry I'm misleadingyou by even talking about other things.MechanicsAngel investors often syndicate deals, which means they join togetherto invest on the same terms. In a syndicate there is usually a"lead" investor who negotiates the terms with the startup. But notalways: sometimes the startup cobbles together a syndicate ofinvestors who approach them independently, and the startup's lawyersupplies the paperwork.The easiest way to get started in angel investing is to find afriend who already does it, and try to get included in his syndicates.Then all you have to do is write checks.Don't feel like you have to join a syndicate, though. It's not thathard to do it yourself. You can just use the standard series AAdocuments Wilson Sonsini and Y Combinator published online.You should of course have your lawyer review everything. Both youand the startup should have lawyers. But the lawyers don't haveto create the agreement from scratch. [2]When you negotiate terms with a startup, there are two numbers youcare about: how much money you're putting in, and the valuation ofthe company. The valuation determines how much stock you get. Ifyou put $50,000 into a company at a pre-money valuation of $1million, then the post-money valuation is $1.05 million, and youget .05/1.05, or 4.76% of the company's stock.If the company raises more money later, the new investor will takea chunk of the company away from all the existing shareholders justas you did. If in the next round they sell 10% of the company toa new investor, your 4.76% will be reduced to 4.28%.That's ok. Dilution is normal. What saves you from being mistreatedin future rounds, usually, is that you're in the same boat as thefounders. They can't dilute you without diluting themselves justas much. And they won't dilute themselves unless they end up net ahead. So in theory, each further round of investment leaves youwith a smaller share of an even more valuable company, till afterseveral more rounds you end up with .5% of the company at the pointwhere it IPOs, and you are very happy because your $50,000 hasbecome $5 million. [3]The agreement by which you invest should have provisions that let you contribute tofuture rounds to maintain your percentage. So it's your choicewhether you get diluted. [4]If the company does really well,you eventually will, because eventually the valuations will get sohigh it's not worth it for you.How much does an angel invest? That varies enormously, from $10,000to hundreds of thousands or in rare cases even millions. The upperbound is obviously the total amount the founders want to raise.The lower bound is 5-10% of the total or $10,000, whicheveris greater. A typical angel round these days might be $150,000raised from 5 people.Valuations don't vary as much. For angel rounds it's rare to seea valuation lower than half a million or higher than 4 or 5 million.4 million is starting to be VC territory.How do you decide what valuation to offer? If you're part of around led by someone else, that problem is solved for you. Butwhat if you're investing by yourself? There's no real answer.There is no rational way to value an early stage startup. Thevaluation reflects nothing more than the strength of the company'sbargaining position. If they really want you, either because theydesperately need money, or you're someone who can help them a lot,they'll let you invest at a low valuation. If they don't need you,it will be higher. So guess. The startup may not have any moreidea what the number should be than you do. [5]Ultimately it doesn't matter much. When angels make a lot of moneyfrom a deal, it's not because they invested at a valuation of $1.5million instead of $3 million. It's because the company was reallysuccessful.I can't emphasize that too much. Don't get hung up on mechanicsor deal terms. What you should spend your time thinking about iswhether the company is good.(Similarly, founders also should not get hung up on dealterms, but should spend their time thinking about how to make thecompany good.)There's a second less obvious component of an angel investment: howmuch you're expected to help the startup. Like the amount youinvest, this can vary a lot. You don't have to do anything if youdon't want to; you could simply be a source of money. Or you canbecome a de facto employee of the company. Just make sure that youand the startup agree in advance about roughly how much you'll dofor them.Really hot companies sometimes have high standards for angels. Theones everyone wants to invest in practically audition investors,and only take money from people who are famous and/or will workhard for them. But don't feel like you have to put in a lot oftime or you won't get to invest in any good startups. There is asurprising lack of correlation between how hot a deal a startup isand how well it ends up doing. Lots of hot startups will end upfailing, and lots of startups no one likes will end up succeeding.And the latter are so desperate for money that they'll take it fromanyone at a low valuation. [6]Picking WinnersIt would be nice to be able to pick those out, wouldn't it? Thepart of angel investing that has most effect on your returns, pickingthe right companies, is also the hardest. So you should practicallyignore (or more precisely, archive, in the Gmail sense) everythingI've told you so far. You may need to refer to it at some point,but it is not the central issue.The central issue is picking the right startups. What "Make somethingpeople want" is for startups, "Pick the right startups" is forinvestors. Combined they yield "Pick the startups that will makesomething people want."How do you do that? It's not as simple as picking startups thatare already making something wildly popular. By then it'stoo late for angels. VCs will already be onto them. As an angel,you have to pick startups before they've got a hit—eitherbecause they've made something great but users don't realize ityet, like Google early on, or because they're still an iterationor two away from the big hit, like Paypal when they were makingsoftware for transferring money between PDAs.To be a good angel investor, you have to be a good judge of potential.That's what it comes down to. VCs can be fast followers. Most ofthem don't try to predict what will win. They just try to noticequickly when something already is winning. But angels have to beable to predict. [7]One interesting consequence of this fact is that there are a lotof people out there who have never even made an angel investmentand yet are already better angel investors than they realize.Someone who doesn't know the first thing about the mechanics ofventure funding but knows what a successful startup founder lookslike is actually far ahead of someone who knows termsheets insideout, but thinks "hacker" means someone who breaks into computers.If you can recognize good startup founders by empathizing withthem—if you both resonate at the same frequency—thenyou may already be a better startup picker than the median professionalVC. [8]Paul Buchheit, for example, started angel investing about a yearafter me, and he was pretty much immediately as good as me at pickingstartups. My extra year of experience was rounding error comparedto our ability to empathize with founders.What makes a good founder? If there were a word that meant theopposite of hapless, that would be the one. Bad founders seemhapless. They may be smart, or not, but somehow events overwhelmthem and they get discouraged and give up. Good founders makethings happen the way they want. Which is not to say they forcethings to happen in a predefined way. Good founders have a healthyrespect for reality. But they are relentlessly resourceful. That'sthe closest I can get to the opposite of hapless. You want to fundpeople who are relentlessly resourceful.Notice we started out talking about things, and now we're talkingabout people. There is an ongoing debate between investors whichis more important, the people, or the idea—or more precisely,the market. Some, like Ron Conway, say it's the people—thatthe idea will change, but the people are the foundation of thecompany. Whereas Marc Andreessen says he'd back ok founders in ahot market over great founders in a bad one. [9]These two positions are not so far apart as they seem, because goodpeople find good markets. Bill Gates would probably have ended uppretty rich even if IBM hadn't happened to drop the PC standard inhis lap.I've thought a lot about the disagreement between the investors whoprefer to bet on people and those who prefer to bet on markets.It's kind of surprising that it even exists. You'd expect opinionsto have converged more.But I think I've figured out what's going on. The three mostprominent people I know who favor markets are Marc, Jawed Karim,and Joe Kraus. And all three of them, in their own startups,basically flew into a thermal: they hit a market growing so fastthat it was all they could do to keep up with it. That kind ofexperience is hard to ignore. Plus I think they underestimatethemselves: they think back to how easy it felt to ride that hugethermal upward, and they think "anyone could have done it." Butthat isn't true; they are not ordinary people.So as an angel investor I think you want to go with Ron Conway andbet on people. Thermals happen, yes, but no one can predictthem—not even the founders, and certainly not you as aninvestor. And only good people can ride the thermals if they hitthem anyway.Deal FlowOf course the question of how to choose startups presumes youhave startups to choose between. How do you find them? This isyet another problem that gets solved for you by syndicates. If youtag along on a friend's investments, you don't have to find startups.The problem is not finding startups, exactly, but finding a streamof reasonably high quality ones. The traditional way to do thisis through contacts. If you're friends with a lot of investors andfounders, they'll send deals your way. The Valley basically runson referrals. And once you start to become known as reliable,useful investor, people will refer lots of deals to you. I certainlywill.There's also a newer way to find startups, which is to come toevents like Y Combinator's Demo Day, where a batch of newly createdstartups presents to investors all at once. We have two Demo Daysa year, one in March and one in August. These are basically massreferrals.But events like Demo Day only account for a fraction of matchesbetween startups and investors. The personal referral is still themost common route. So if you want to hear about new startups, thebest way to do it is to get lots of referrals.The best way to get lots of referrals is to invest in startups. Nomatter how smart and nice you seem, insiders will be reluctant tosend you referrals until you've proven yourself by doing a coupleinvestments. Some smart, nice guys turn out to be flaky,high-maintenance investors. But once you prove yourself as a goodinvestor, the deal flow, as they call it, will increase rapidly inboth quality and quantity. At the extreme, for someone like RonConway, it is basically identical with the deal flow of the wholeValley.So if you want to invest seriously, the way to get started is tobootstrap yourself off your existing connections, be a good investorin the startups you meet that way, and eventually you'll start achain reaction. Good investors are rare, even in Silicon Valley.There probably aren't more than a couple hundred serious angels in the wholeValley, and yet they're probably the single most important ingredientin making the Valley what it is. Angels are the limiting reagentin startup formation.If there are only a couple hundred serious angels in the Valley, then by deciding to become one you could single-handedly make the pipelinefor startups in Silicon Valley significantly wider. That is kindof mind-blowing.Being GoodHow do you be a good angel investor? The first thing you need isto be decisive. When we talk to founders about good and badinvestors, one of the ways we describe the good ones is to say "hewrites checks." That doesn't mean the investor says yes to everyone.Far from it. It means he makes up his mind quickly,and follows through. You may be thinking, how hard could that be?You'll see when you try it. It follows from the nature of angelinvesting that the decisions are hard. You have to guess early,at the stage when the most promising ideas still seem counterintuitive,because if they were obviously good, VCs would already have fundedthem.Suppose it's 1998. You come across a startup founded by a couplegrad students. They say they're going to work on Internet search.There are already a bunch of big public companies doing search.How can these grad students possibly compete with them? And doessearch even matter anyway? All the search engines are trying toget people to start calling them "portals" instead. Why would youwant to invest in a startup run by a couple of nobodies who aretrying to compete with large, aggressive companies in an area theythemselves have declared passe? And yet the grad students seempretty smart. What do you do?There's a hack for being decisive when you're inexperienced: ratchetdown the size of your investment till it's an amount you wouldn'tcare too much about losing. For every rich person (you probablyshouldn't try angel investing unless you think of yourself as rich)there's some amount that would be painless, though annoying, tolose. Till you feel comfortable investing, don't invest more thanthat per startup.For example, if you have $5 million in investable assets, it wouldprobably be painless (though annoying) to lose $15,000. That'sless than .3% of your net worth. So start by making 3 or 4 $15,000investments. Nothing will teach you about angel investing likeexperience. Treat the first few as an educational expense. $60,000is less than a lot of graduate programs. Plus you get equity.What's really uncool is to be strategically indecisive: to stringfounders along while trying to gather more information about thestartup's trajectory. [10] There's always a temptation to do that,because you just have so little to go on, but you have to consciouslyresist it. In the long term it's to your advantage to be good.The other component of being a good angel investor is simply to bea good person. Angel investing is not a business where you makemoney by screwing people over. Startups create wealth, andcreating wealth is not a zero sum game. No one has to lose for youto win. In fact, if you mistreat the founders you invest in, they'lljust get demoralized and the company will do worse. Plus yourreferrals will dry up. So I recommend being good.The most successful angel investors I know are all basically goodpeople. Once they invest in a company, all they want to do is helpit. And they'll help people they haven't invested in too. Whenthey do favors they don't seem to keep track of them. It's toomuch overhead. They just try to help everyone, and assume goodthings will flow back to them somehow. Empirically that seems towork.Notes[1]Convertible debt can be either capped at a particular valuation,or can be done at a discount to whatever the valuation turns outto be when it converts. E.g. convertible debt at a discount of 30%means when it converts you get stock as if you'd invested at a 30%lower valuation. That can be useful in cases where you can't ordon't want to figure out what the valuation should be. You leaveit to the next investor. On the other hand, a lot of investorswant to know exactly what they're getting, so they will only doconvertible debt with a cap.[2]The expensive part of creating an agreement from scratch isnot writing the agreement, but bickering at several hundreddollars an hour over the details. That's why the series AA paperworkaims at a middle ground. You can just start from the compromiseyou'd have reached after lots of back and forth.When you fund a startup, both your lawyers should be specialistsin startups. Do not use ordinary corporate lawyers for this. Theirinexperience makes them overbuild: they'll create huge, overcomplicatedagreements, and spend hours arguing over irrelevant things.In the Valley, the top startup law firms are Wilson Sonsini, Orrick,Fenwick & West, Gunderson Dettmer, and Cooley Godward. In Bostonthe best are Goodwin Procter, Wilmer Hale, and Foley Hoag.[3]Your mileage may vary.[4]These anti-dilution provisions also protect you againsttricks like a later investor trying to steal the company by doinganother round that values the company at $1. If you have a competentstartup lawyer handle the deal for you, you should be protectedagainst such tricks initially. But it could become a problem later.If a big VC firm wants to invest in the startup after you, they maytry to make you take out your anti-dilution protections. And ifthey do the startup will be pressuring you to agree. They'll tellyou that if you don't, you're going to kill their deal with the VC.I recommend you solve this problem by having a gentlemen's agreementwith the founders: agree with them in advance that you're not goingto give up your anti-dilution protections. Then it's up to themto tell VCs early on.The reason you don't want to give them up is the following scenario.The VCs recapitalize the company, meaning they give it additionalfunding at a pre-money valuation of zero. This wipes out theexisting shareholders, including both you and the founders. Theythen grant the founders lots of options, because they need them tostay around, but you get nothing.Obviously this is not a nice thing to do. It doesn't happen often.Brand-name VCs wouldn't recapitalize a company just to steal a fewpercent from an angel. But there's a continuum here. A lessupstanding, lower-tier VC might be tempted to do it to steal a bigchunk of stock.I'm not saying you should always absolutely refuse to give up youranti-dilution protections. Everything is a negotiation. If you'repart of a powerful syndicate, you might be able to give up legalprotections and rely on social ones. If you invest in a deal ledby a big angel like Ron Conway, for example, you're pretty wellprotected against being mistreated, because any VC would think twicebefore crossing him. This kind of protection is one of the reasonsangels like to invest in syndicates.[5]Don't invest so much, or at such a low valuation, that youend up with an excessively large share of a startup, unless you'resure your money will be the last they ever need. Later stageinvestors won't invest in a company if the founders don't haveenough equity left to motivate them. I talked to a VC recently whosaid he'd met with a company he really liked, but he turnedthem down because investors already owned more than half of it.Those investors probably thought they'd been pretty clever by gettingsuch a large chunk of this desirable company, but in fact they wereshooting themselves in the foot.[6]At any given time I know of at least 3 or 4 YC alumni who Ibelieve will be big successes but who are running on vapor,financially, because investors don't yet get what they're doing.(And no, unfortunately, I can't tell you who they are. I can'trefer a startup to an investor I don't know.)[7]There are some VCs who can predict instead of reacting. Notsurprisingly, these are the most successful ones.[8]It's somewhat sneaky of me to put it this way, because themedian VC loses money. That's one of the most surprising thingsI've learned about VC while working on Y Combinator. Only a fractionof VCs even have positive returns. The rest exist to satisfy demandamong fund managers for venture capital as an asset class. Learningthis explained a lot about some of the VCs I encountered when wewere working on Viaweb.[9]VCs also generally say they prefer great markets to greatpeople. But what they're really saying is they want both. They'reso selective that they only even consider great people. So whenthey say they care above all about big markets, they mean that'show they choose between great people.[10]Founders rightly dislike the sort of investor who says he'sinterested in investing but doesn't want to lead. There arecircumstances where this is an acceptable excuse, but more oftenthan not what it means is "No, but if you turn out to be a hot deal,I want to be able to claim retroactively I said yes."If you like a startup enough to invest in it, then invest in it.Just use the standard series AA terms and write them a check.Thanks to Sam Altman, Paul Buchheit, Jessica Livingston,Robert Morris, and Fred Wilson for reading drafts of this.Comment on this essay.
