Premature optimization

Startups talk a lot about optimization—A/B testing, product enhancements, conversion funnels, email campaigns, etc.  These sorts of things can often produce 10% gains, and much more if you can compound several together. 

This is great, but only if the business is already working.  If no users love your product, you’re wasting your time working on anything but that.  It’s always painful to sit in a board meeting and listen to a company miss the forest for the trees—i.e., even if they make all these incremental improvements that the entire company is focused on, they’re going to end up at 2X, and they need to get to 100X. 

It’s dangerous to spend all your mental energy on incremental improvements when what you really need is a step change.

It’s a useful exercise to think about whether or not all the optimization work (or ad buying, for that matter) you’re doing is worth it, even if it all works.  It is usually not until you’re already quite successful.  Until then, you should focus relentlessly on making a product your users love (and making sure you're going after a large enough market).

Startups often ask me how to grow faster.  I usually say “build a great product, and you will be able to make it grow”.  The most sustainable (and cheapest) kind of growth is word-of-mouth growth.


Party rounds

There is a recent trend in Silicon Valley towards party rounds—in early financing rounds, instead of raising large amounts of money from a few large investors, companies are instead raising small amounts of money from many small investors.  The number of investors in such a round is commonly between 10 and 20, but I’ve seen rounds with over 50 investors.

I think the rising popularity of party rounds is bad for companies.

Ask some startups how useful their investors are, and you’ll get a variety of responses, but some commonalities emerge. It turns out that investors are generally about as involved as they are invested. Having at least one investor very focused on your company is valuable, even if the investor is not very good.  The cadence and rhythm of meeting with someone every month to review progress and goals turns out to be an important focusing function [1].

In a typical party round, no single investor cares enough to think about the company multiple times a day.  Each investor assumes that at least 1 of the N other investors will be closely involved, but in fact no one is, and the companies sometimes wander off into a very unfocused wilderness.

While it’s true that many investors are bad, I have found wonderful exceptions.  Good investors stay out of the way when you don’t need them and help you a lot when you do.  This is really valuable—it turns out that most of the great companies end up having great investors.

Additionally, a closely involved investor will help coordinate your next round, and can bridge the company if necessary (things often take longer than founders plan).  I’ve noticed that companies that raise party rounds seem to have a harder time raising their next round compared to companies that raised their first big round from a VC.

Perhaps another reason that party round companies have a harder time raising capital is that worse companies get funded this way—everyone seems less disciplined in deciding to invest.  Social proof plays a large role, and because the amounts are smaller people don’t think as hard about the quality of the business.

I’m all for seeing founders have more power and more ownership, and I’m all for VCs being forced to evolve, but I don’t think that having party rounds replace what used to be an A round is the answer.


[1] There is a version of this that makes Y Combinator work—the deadline of Demo Day is extremely good at focusing companies on the right things.


Man has a large capacity for effort. In fact it is so much greater than we think it is that few ever reach this capacity. We should value the faculty of knowing what we ought to do and having the will to do it. Knowing is easy; it is the doing that is difficult. The critical issue is not what we know but what we do with what we know. The great end of life is not knowledge, but action. I believe that it is the duty of each of us to act as if the fate of the world depended on him ... we must live for the future, not for our own comfort or success.

--Admiral Rickover

A founder-friendly term sheet

When I invest (outside of YC) I make offers with the following term sheet.  I’ve tried to make the terms reflect what I wanted when I was a founder.  A few people have asked me if I’d share it, so here it is.  I think it’s pretty founder-friendly.

If you believe the upside risk theory, then it makes sense to offer compelling terms and forgo some downside protection to get the best companies to want to work with you.

What’s most important is what’s not in it:

*No option pool.  Taking the option pool out of the pre-money valuation (ie, diluting only founders and not investors for future hires) is just a way to artificially manipulate valuation.  New hires benefit everyone and should dilute everyone.

*The company doesn’t have to pay any of my legal fees.  Requiring the company to pay investors’ legal fees always struck me as particularly egregious—the company can probably make better use of the money than investors can, so I’ll pay my own legal fees for the round (in a simple deal with no back and forth they always end up super low anyway).

*No expiration.  I got burned once by an exploding offer and haven’t forgotten it; the founders can take as much time as they want to think about it.  In practice, people usually decide pretty quickly.

*No confidentiality.  Founder/investor relationships are long and important.  The founders should talk to whomever they want, and if they want to tell people what I offered them, I don’t really care.  Investors certainly tell each other what they offer companies. (Once we shake hands on a deal, of course, I expect the founders to honor it.)

*No participating preferred, non-standard liquidation preference, etc.  There is a 1x liquidation preference, but I’m willing to forgo even that and buy common shares (and sometimes do, although it has implications on the strike price for employee options so most founders don’t want it).  In early-stage investing, you should not focus on downside protection.

I have an allergic reaction to complex deal structures, as they invariably end up with all sorts of unintended consequences.  Also, getting this right in early rounds is important—future rounds tend to do whatever the previous rounds did.

(What I do care about is ownership percentage and pro rata rights to maintain that ownership percentage in future rounds.  Most of the rest I don’t care about, but it’s never contentious anyway.)

Startup Advice

In honor of the new YC batch starting tomorrow, here is some of the best startup advice I’ve heard or given (mostly heard):

1.     Make something people want.

2.     A great team and a great market are both critically important—you have to have both.  The debate about which is more important is silly.

3.     Write code, talk to users, and build the company (hire the best people you can find, get the culture right, fundraise, close sales, etc.)  Most other things that founders do are a waste of time.

4.     Set a clear, easy-to-understand vision for your company, and make it be a mission people believe in.

5.     Stay focused and don’t try to do too many things at once.  Care about execution quality.

6.     You have to have an almost crazy level of dedication to your company to succeed.

7.     In general, don’t start a startup you’re not willing to work on for ten years.

8.     Be relentlessly resourceful.

9.     In the current pivot-happy world, good ideas are underweight.  It’s worth the time to think through a good one.

10.  Growth solves (nearly) all problems.

11.  While growth is critical and you should focus on it, occasionally consider where you’re going—you need both growth and to be growing towards something valuable.

12.  Obsess about the quality of the product.

13.  Overcommunicate with your team.  For some reason most founders are really bad at this one.  Transparency is your friend.

14.  Move fast.  Speed is one of your main advantages over large companies.

15.  Hire slow; fire fast.  Hiring is the most important thing you do; spend at least a third of your time on it.

16.  Occasionally think about why the 20th person will join your company.

17.  Hire smart and effective people that are committed to what you’re doing.  The last five words there are important.

18.  Hire friends and friends of friends.  Go after these people like crazy to get them to join.  Some other candidate sources are ok, but I always got bad results from technical recruiters.

19.  Generally, value aptitude over experience.

20.  Hire people that you could describe as animals.

21.  Eliminate distractions.

22.  Don’t die.

23.  Be frugal.

24.  You’ll often hear conflicting advice about everything but “build a great product”.  This means you can go either way on much of the rest of it and it doesn’t really matter.  Just make a decision and get back to work.  Product/market fit is what matters.  You can—and will—make a lot of mistakes.

25.  You make what you measure.

26.  Startups are very hard no matter what you do; you may as well go after a big opportunity.

27.  Momentum is critical.  Don’t lose it.

28.  Keep salaries low and equity high.

29.  Keep the organization as flat as you can.

30.  When working on a deal—raising money, trying to get a partnership, etc.—it’s important to create a competitive situation.

31.  Schleps are good.

32.  Don’t forget to make money.

33.  Journalists like hearing directly from founders.  If you hire PR people, resist their desire to control all the contact.

34.  It’s standard for founders to keep board control in the first round.

35.  Listen to everyone.  Then make your own decision.

36.  Remember that you are more likely to die because you execute badly than get crushed by a competitor.

37.  Get lucky.

38.  Have a direct relationship with your customers.

39.  Be formidable—do not be easy to push around.

40.  Don’t let your company be run by a sales guy.  But do learn how to sell your product.

41.  Have a culture that rewards output.

42.  Don’t hire professional managers too early.

43.  Simple is good.  Be suspicious of complexity.

44.  Get on planes in marginal situations.  In-person is still better than tele-anything.

45.  Most things are not as risky as they seem.

46.  Be suspect of anyone who says the word process too often.

47.  Raise a bit more money than you think you need.

48.  Ignore the fact that “the press loves [you]”.

49.  Have great customer service.

50.  You can create value with breakthrough innovation, incremental refinement, or complex coordination.  Great companies often do two of these.  The very best companies do all three.

51.  The role of the board is advice and consent.  If the CEO does not lay out a clear strategy and tries to get the board to set one, it will usually end in disaster.

52.  Board observers are usually a headache.

53.  If you pivot, do it fully and with conviction.  The worst thing is to try to do a bit of the old and the new—it’s hard to kill your babies.

54.  It’s better to make a decision and be wrong than to equivocate.

55.  Set goals for the company and motivate people to get there.

56.  Always praise good work.

57.  Celebrate your wins as a company.  Get t-shirts for big milestones.

58.  Have a good operational cadence where projects are short and you’re releasing something new on a regular basis.

59.  You can win with the best product, the best price, or the best experience.

60.  Meetups and conferences are generally a waste of time.

61.  If the founders of your company seem to care more about being founders than they care about your specific company, go join another company.

62.  It’s easier to sell painkillers than vitamins.

63.  Be suspicious of any work that is not building product or getting customers.  It’s easy to get sucked into an infrastructure rewrite death spiral.

64.  It’s better to have a few users love your product than for a lot of users to sort of like it.

65.  Learn how to stay extermally optimistic when your world is melting down.

66.  Startups should require as few miracles as possible, but at least one.

67.  You have to have great execution—far more people have good ideas than are willing to roll up their sleeves and get shit done.

68.  Don’t have a diverse culture in the early days.

69.  Keep a to-do list every day.  At the top of it, put the one or two big things you want to work on.

70.  Being the CEO is miserable more often than it’s good.  But when it’s good, it’s really good.

71.  On the really bad days, remember that tomorrow will be better—it’s hard to see it being much worse!

72.  Sleep and exercise.

73.  Success in a startup is usually pass/fail.  Worry more about making sure you pass than an extra point of dilution.

74.  Good investors are worth a reasonable premium.

75.  Give your investors something to do.

76.  Go for a few highly involved investors over a lot of lightly engaged ones.

77.  Raise money on promise.  Raise money on clean terms.

78.  Do reference checks on your potential investors.  Ask other founders how they are when everything goes wrong.

79.  Investors love companies other investors love.

80.  A lot of the best ideas seem silly or bad initially—you want an idea at the intersection of “seems like bad idea” and “is good idea”. (It’s important to note you need to be contrarian and right, not simply contrarian.)

81.  Surf someone else’s wave.

82.  Sometimes you can succeed through sheer force of will.

83.  All startups are fucked in at least one major way.  Keep going.

84.  Keep an eye on cash in the bank and don’t run out of it.

85.  Pay a lot of attention to the relationship between cofounders, especially if both/all of you want to be CEO.

86.  Stay small and nimble.

87.  Have a staff meeting at least once a week.

88.  Find a mentor that will teach you how to manage.

89.  Keep burn low until you’re sure everything is working.

90.  Be suspect about buying users.

91.  Lead by example.

92.  Have the right kind of office.  The proper office for a very small company is an apartment or house.

93.  Share results (financial and key metrics) with the company every month.

94.  Have a table in your offer letters that shows how much the stock you’re granting a new hire could be worth in various scenarios.

95.  The best startups are defined by exceptions; all of these rules are probably breakable, but probably not all at the same time.

By endurance we conquer

I'm reading The Endurance and "by endurance we conquer" (Ernest Shackleton's family motto) struck me as a great piece of startup wisdom.

Everyone knows that you need a great team, great execution, and a great idea.  Less obvious is that you have to have great endurance.  It's very tough to keep going when everyone tells you your idea sucks and it will never work (especially when things are plainly not working).  It's tough to keep going when everything goes wrong, which it almost certainly will. And it's tough to keep working when you're really tired, but very often that extra 5% at a critical point is how you beat out a competitor for a critical deal and then they disappear in the rearview mirror.

Most startups don't die at the hands of a competitor.  It's more often something like an internal implosion, the founders giving up, or not building something people want (and failing to remedy that situation). You can win by endurance.

Upside risk

Everyone claims that they understand the power law in angel investing, but very few people practice it.  I think this is because it’s hard to conceptualize the difference between a 3x and a 300x (or 3000x) return. 

It’s common to make more money from your single best angel investment than all the rest put together.  The consequence of this is that the real risk is missing out on that outstanding investment, and not failing to get your money back (or, as some people ask for, a guaranteed 2x) on all of your other companies.

And yet angel investors continue to ask for onerous terms to mitigate their “downside risk”.  All this does is piss founders off, misalign incentives, and harm the investors’ chance of getting to invest in the best deals, because those are usually hotly pursued and good founders check references.  An angel investor is much better off focusing on investing at a reasonable price [1] and not trying to “win” on any other terms.

Instead of downside risk [2], more investors should think about upside risk—not getting to invest in the company that will provide the return everyone is looking for. 

[1] Speaking of price, the mistake that founders make (corresponding to investors focusing too much on downside terms) is focusing too much on getting a high price.  I have seen many founders price out good investors and put the company in a bad situation facing a down round a year later, all because they were obsessed with getting a high sticker price for their company.  I think it’s because it gives founders something quantitative to compete on. 

[2] As a side note, I think most people have terrible intuition about investment risk/reward tradeoffs in general—this is not limited to private companies.  It feels like every time I turn on CNBC (which is thankfully very infrequently) they’re talking about an impending total collapse.  The end of the world only happens once; it’s very unlikely to be Monday morning.  But we seem hardwired to focus on downside risk.  The CNBC watchers would be better off keeping a cushion in cash and not selling their stocks after every panic. 

Software to avoid the software people

A few years ago, many of the Y Combinator B2B startups wrote tools for the developers in other companies--metrics software, deployment software, monitoring software, build software, development frameworks, etc.  The startups would want to meet with the technology people at companies to sell their service.

There's been a significant shift--lots of the YC B2B startups are now building software to help non-technical people in companies (usually large ones with a primary business that is not writing software) avoid their internal IT department when they need software to help them get something done.  It's faster and easier.  So now the startups are trying to avoid the developers at the other company (so they don't get blocked) and sell to the person who is waiting in the internal development queue. 

It will be interesting to see how far this trend goes, and something to keep in mind if you’re starting a new company.

More interesting dinner conversations

When seated at a table with people you don't know, ask "what are you interested in?" or "what have you been thinking about lately?" instead of "what do you do?".

(Surprisingly often you get a look of utter confusion, followed by fifteen seconds of hemming and hawing, and then a version of "man, i really need to take some time off".)


I don't believe that any of the blurry UFO photos are real, for a very simple reason--they all resemble a slightly more advanced version of the the then-current technology and style.  It seem very unlikely a spacecraft would fit the gestalt of the decade during which it crashed into earth.

There is a version of this to keep in mind when listening to startups make certain types of extraordinary claims.

(As a random aside, I went back and looked at some old photos of purported UFOs and real aircraft.  A hat-tip to the designers of the SR-71 for creating the only aircraft that still manages to look super-futuristic 50 years later.)