Patrick McKenzie founded flashcard app Bingo Card Creator and programming game/hiring tool Starfighter, before joining the team at payment processing service Stripe to help increase the GDP of the internet through content and communications. He's well known for the startup advice he shares as @patio11 on Twitter and Hacker News—perhaps most popularly for telling people to "charge more."
Appointment Reminder was my second software business. I ran it from 2010 until selling it in 2016.
I would have done one thing differently: not started it at all. Peldi from Balsamiq, when I told him my idea, said "Patrick! Do you really want to spend 5 years of your life optimizing the schedules of dentist offices?" I remember, virtually word for word, "Oh God no, that sounds terrible, but it is a great business."
When you're at "Oh God no!" with respect to the business you are starting... do something else!
Two things which I've come to appreciate:
1) Good businesses and bad businesses take approximately the same amount of work.
2) I was so much more productive, and enjoyed life so much more, building things for software people at Starfighter (and then at Stripe) than I was on a per-unit-of-effort basis for AR or my prior businesses. I was also more able to actually summon the effort to work on them versus e.g. playing League of Legends.
I've had a number of weird hobbies for a very long time, but I'm a systems thinker and things which interact with money have always been fascinating for me. I was the sort of kid who papertraded in 8th grade not just to win extra credit on a "learn what the stock market is" assignment for social studies but because it was legitimately fascinating to me.
The Internet rewards having weird hobbies! You can find almost infinite depth on any of them. So when life would periodically bring me into contact with e.g. problems with my own credit report, I often dove and dove deep, and then had one more thing to incorporate into mental models going forward. (I ghostwrote Fair Credit Reporting Act complaints for a few years. Long story.)
I don't think I'm uniquely attracted to financial topics; I get the same sort of fascination from them that I do with e.g. computers, applications, hidden-in-plain-sight systems for organizing the world, history, games, poker, etc.
Fix engineering hiring.
We took a swing at this with Starfighter. It didn't work out. I think there are a lot of different potential takes on the problem space.
People will probably never buy houses and generally shouldn't buy cars with credit cards, given the positioning of credit cards among all payment types vis cost structure and guarantees.
All payment types conceptually bid for transactions against every other payment type. This is both the literal sense of bid in cost ("To whom?" an interesting and nuanced question there) and in the sense of "The payment method is a product and products have a number of features to them that might be more or less attractive." For example, there is a payment rail between an employer and an employee, and "How amenable is that payment rail to reversibility?" matters a non-zero amount to the employer but probably matters less to the employee than e.g. how much reversibility matters if I'm ordering from an e-commerce store for the first time.
By 2030 we'll see a lot more payments, because of growth of the economy and increasing inroads of markets into portions of one's life that have not been marketized as much as other places yet. (Biggest one that I think will surprise people: I believe strongly that cooking will go from something which "substantially everyone" does / expects to do to something closer to knitting your own clothes... sure, that's a hobby that exists, but you'd not predict that I would be a participant in that hobby in the same way you'd automatically predict that someone in my household cooks.)
I also think transactions will continue moving online, for some value of "online." Even retail face-to-face transactions at Starbucks may well move over app-enabled rails. "The future is here, just not evenly distributed." is a common quote and a good one.
The economics of doing this are quite attractive to firms. Large populations of users are becoming less reticent about doing this, including for transactions which are larger than the historical norms for online transactions.
Will we see people buy houses online? I mean, clearly yes, we will see that. Will it become dominant to do it without meeting one's counterparty and/or broker? That feels unlikely. But could that final signing move onto an entirely paperless system? Clearly it could; that was impossible 20 years ago and yet is epsilon away from the best versions of the experience now. The best experience available in the world will be the standard experience within 10 years; it almost can't fail to happen in that fashion, due to the efficiency and UX gains.
I don't know that I'd do all that much market research as market research is usually understood, largely because I'd be doing it for something where it is overwhelmingly obvious to me that there is a customer need. In lieu of market research, I'd probably be attempting to pre-sell the product.
This worked out well and not-so-well at Starfighter: we successfully closed a high number of conversations with hiring companies prior to having anyone to place as a candidate. However, that was less of a checkpoint than I had hoped, because the only thing the company has to promise is "Sure, we'll look at your candidates and pay you iff we hire one." If I were doing that again, I would try to get either a commitment to actually hiring candidates and/or a materially sized deposit up front, and use the fact of "Look, you're already $25k in here" to do much more in-depth research on their internal hiring guidelines than we did at Starfighter.
You're going to learn a heck of a lot when doing a business, regardless of what that business is and regardless of what your present skills are.
A friend of mine (Jay Winder) often talks to me about product/founder/market fit, and I think that is potentially a more powerful thing than merely having skills alignment. When you're "the right person to do this" (because of longstanding obsession with the domain, pre-existing connections, some "unfair advantage", etc) you're more credible to clients and investors and you tend to get even more credible as you execute than you would in the baseline case.
Skills you can learn (surprisingly quickly) and/or hire for. An underappreciated bit of starting a company: you're going to spend so much time on this problem, more even than many people who encounter it professionally, that you'll relatively quickly be one of the world's experts in it. This is a surprising insight but I think it rounds to true: you're not aiming to understand an entire field like e.g. dentistry. You're aiming at 5% of the operations of a particular employee at a dentist practice. How many years do you have to focus on that to have spent more hours on it, and more considered hours, than every person thinking about that problem? Probably something like 1, 1.5, 2 years, right?
I'm surprised that Heroku's model didn't win over AWS' model and that DevOps is accordingly a core competence at most SaaS companies. This seems obviously terrible to me every time I'm doing DevOps, which probably took ~20% of all engineering cycles at my last company for surfacing very little customer value.
So I have weird hobbies, and one of my weird hobbies is playing the points optimization game... casually.
This turns out to be a great hobby to prepare you for working at Stripe, but makes it socially difficult for me to answer this question in the maximally straightforward fashion.
What I can say:
I have approximately 20 pieces of plastic in my wallet and 10 in my desk drawer. That's higher than I "should" have given gamesmanship but my numbers get doubled by needing them in Japan and the US and doubled again by previously/currently running a business.
The exact names on the pieces of plastic? Pretty predictable given what I've said above.
I don't have extremely developed thoughts here. Broadly, I think economic measurement is hard.
An example: the impact of the Internet existing broadly doesn't show up in GDP numbers... but shouldn't it? We are far better off for having it than not having it. We're far better off that Google exists. We're far better off that the entire sum of human knowledge is accessible via your cell phone, to a degree far greater than the marginal increase in cell phone bills since 1990.
Depends on your risk tolerance and how sure you are that you want to run a SaaS company, IMHO. I will report that I started learning a lot faster and executing a lot better after having more cycles to spend on my companies, and basically every change in life which increased focus has been a net win for me. (Going full-time, winnowing down from ~4 simultaneous businesses to one, etc.)
One challenge: folks with expensive problems have high salaries and are behind gatekeepers.
People overestimate the degree to which that is true. Bill Gates has a gatekeeper, certainly. How many people do you think read the inbox of e.g. the head of application security at AppAmaGooFaceSoft? My prediction: one. Which is probably less useful from the building-things perspective (you're not going to sell to that person first) but is extremely useful to know about the world for e.g. career purposes.
Working in industry will expose you to problems very quickly. Failing that, talk to people. Complaining about work is a time-honored passtime. A surprising number of people will let you shouldersurf them when working; no engineer who has ever done that has said "Ahh you are working in the optimum fashion; such a pity software could not do this better."
Stepping back a bit, problems feel a bit less important to me as time goes on. The world is awash in problems. Problems you care about are just so much more rewarding to work on, and they pull better work out of you. Problems experienced by people you care about quintuply so, because if you start a SaaS business you'll spend much more time talking to customers and getting in their head than you will be on e.g. modeling their workflow in Ruby or optimizing your AWS spend.
So I'd suggest most people asking this question instead ask "How do I find a user population that buys software that I would enjoy having in my life five days a week for 5~10 years?" My answer for that, after a lot of soulsearching, is "I enjoy helping software people a lot more than I enjoy helping undifferentiated professionals. Maybe I should just do that. Maybe no maybe." Your answer may vary.
I'd separate what I want as a consumer from what I think is optimal as a person who has run software businesses and understands how they work.
There have been some apps where I thought "Oww, that's rich" on an emotional level, and then bought them, because I'm a responsible businessman and had a need for them which greatly outweighed the price. I don't think naming names is particularly helpful.
Note that this reaction is hardly limited to software. In-flight wifi: ouch! And yet I should pay for it literally every time without thinking about it. Accountants / lawyers / etc: ouch squared! And yet, again, clearly justifiable.
I should write a /books/ page on my site sometime. Thanks for the suggestion.
I would not write a /news/ page on my website. I am broadly very bearish on traditional journalism; the percent of value in the field of journalism and commentary created by Matt Levine specifically is far, far higher than it should be. I'd rather follow experts who know about X/Y/Z (and Matt Levine is plausibly than more than a commentator/journalist). Happily for me, the Internet makes that relatively straightforward. Unhappily for journalism, the Internet gives users much better options for in-depth explanation of e.g. Japanese trade policy than "Here's what a 24 year old who doesn't speak Japanese and has studied this topic for 2 weeks thinks as a result of consulting 4 people in their rolodex."
I am going to take the liberty of substituting "a high-status well-run software company" as the target here rather than Stripe.
You should certainly want to work at one of them for your first job in industry. It will help you learn more quickly than working at a company which happens to employ engineers but where software isn't key to the success of the enterprise like e.g. a typical insurance company. I would try to find companies in the goldilocks zone of "Big enough to have mentors available who will have bandwith to help you grow; small enough such that your personal contributions actually matter and that you will own things that you can point to in the world and say 'That thing, that thing was mine' versus e.g. optimizing the SQL query which renders a preference page for the company's 37th most important product.
Given that you want to work at one of these companies, see career advice I have given before. The unique factor bootcamp grads have is that you as soon as practical want a better branding for yourself than "bootcamp grad" because the industry has not done you favors in teaching decisionmakers at software companies that that branding leads to successful engineers.
Specific things you should cultivate:
The great thing about our industry is that, given 2 years of execution at an employer, you're just "an engineer" not "an engineer who graduated from a bootcamp." A less great, but true, thing about our industry is that it matters a non-trivial amount what the status of that engineering employer is. If possible, consider getting one with good "resume value"; this matters disproportionately early in your career and when you otherwise do not have many proof-points that you'll probabilistically be a great engineer.
Why not do both? There are infinite needs out there; you've got 5~10 shots at running a company in your life.
Why not take the best shots you can?
I enjoyed the Phoenix Project a lot but don't have super-developed thoughts on the Theory of Constraints. (More business advice should be placed in a well-executed fictional narrative.)
That's like asking me for my favorite book... there are just so many that I love for so many different reasons.
If you put a gun to my head, I've given the answers Twilio and Stripe for the last ~10 years and, in the course of the next 10, will probably say Superhuman in the same breath with them a lot.
I think the relatively easy part of valuing a stock option is understanding how the math works out given the company achieves $X valuation at exit (covered in many places on the Internet; the Holloway guide is pretty good) and the relatively difficult part is understanding whether a given company is likely to win or not. It hugely, hugely, hugely matters that you pick right on a company which wins big; contingent on this, most other infelicities in math or negotiation magically go away.
Plausibly if you can see that a company is going to succeed in 6+ years you should be an investor rather than an engineer, but leaving that aside: isn't it sometimes kind of obvious to geeks with taste which companies are going to win among the geeks with taste crowd? Stripe is a healthy business in 2020. Would you have picked Stripe for Stripe-shaped things in 2014? Yes? Then that is a data point. It isn't a sufficient data point, probably, but it is a material data point.
As to the tradeoff between companies justifying a pay cut: you're making a bet on some sort of growth as a result of choosing small-and-nimble over big-and-stable, right? Growth in equity, growth in your skillset, growth in opportunity selection, etc. You should be very explicit with yourself about that bet, and then go about derisking it. Don't just passively consume the company's hiring process; try to understand a) what are their levers for growing and b) to the limit you can perceive externally, are they already growing quickly? Ask for their model and ask questions; develop your own model. Ask probing questions of existing engineers with regards to growth in opportunities, etc. Do some soulsearching on whether your work style and temperament comports with high-growth opportunities in (without loss of generality) AppAmaGooBookSoft, which do still have many high-growth opportunities available, or whether you're best positioned working near a frontier.
I think that generic project tracking tools have to compete with both Excel and every other generic project tracking tool, and would personally struggle to find meaningful competitive differentiation, but I think there is huge, huge, huuuuuuuuuuuuge opportunity for "It is the X business in a box" which includes salient projects for the X business and which uses deep understanding of their core projects to ROFLstomp over the experience of trying to run the same things out of Excel/Basecamp/etc.
Like, you could run a kitchen countertop installation business out of Google Docs and an inbox... but Moraware has persuasive reasons why that would be a poor idea versus just using Moraware. ("What reasons?" I do not have a high-fidelity model for them but bet I would if I did countertop installation to feed my family.)
There are some industrial byproducts of the cryptocurrency space which are plausibly businesses.
e.g. "Cryptocurrency exchanges have a problem with KYC. Gee it sure would be great to figure out a way to demonstrate identity over the Internet." is a plausible thesis. If you successfully build that, don't sell it to cryptocurrency companies. They are incorrigible muppets who largely don't want to be effective at KYC. The broader financial industry (and e-commerce, and the million other use cases for identity) care much more, have deeper pockets, and are less conflicted about consuming this SaaS.
My goal over the 45+ years of my career is to maximize for value created for software people. (I usually articulate it as "the product of number of people helped times the median concrete improvement" or similar.)
I work at Stripe because I think working at Stripe is the best option available for doing that at present. (Considering my values, skills, likelihood of having an impact on the margin, magnitude of margin if successful, personal circumstances, etc.)
Are you a fan of Star Trek by any chance?
O'Brien learns to cook, in a world where infinite free tasty food is available at the touch of a button. I think it's instructive to think why one would still do that. One potential answer is: the ritual of cooking for someone, and for incorporating someone into that ritual, plausibly matters a lot more than simply feeding them does.
If this doesn't sound like it matters to you, no worries at all. The world is moving in a direction where cooking will become a niche preference and where lack of cooking skill will be the least stigmatizing it has ever been. This prediction does not necessarily decrease the returns to cooking skill!
It is not obvious to me that I am correctly calibrated on this score.
(Do I have imposter syndrome about having imposter syndrome?!)
I am broadly skeptical about the discourse of "top talent", which suggests that it is static, and think there is some combination of person, opportunity, and supporting infrastructure which results in outsized productivity (or other outcomes).
That said, productivity differences between people, opportunities, and supporting infrastructure (keeping other factors constant) is one of those brute facts of the world that will continue negatively surprising someone until they come to terms with it.
What makes people productive? High energy levels, good discernment of what to work on, ability to leverage the work of others (particularly as they move forward in their career), good alignment with the organization they find themselves in, being good at the thing they do (a surprisingly underrated source of competitive advantage!), caring an unreasonable amount about the thing they do (a surprisingly underrated source of competitive advantage!), an unwillingness to settle for their current level of productivity, etc.
Do you want to have a job? Then get a job at an employer who values software engineers. They'll pay a lot more than $5k a month.
Do you want to have a business? Then you should focus on selling something similar to what you've got for other companies; plausibly 2~3 sales gets you to sustainability.
Do you want to have an interesting flexible lifestyle? Then plausibly keep doing what you're doing and attempting to walk up the value of the one account that you have. (And/or execute on the business and then just operate it in a flexible fashion.)
Go work for, without loss of generality, Ramit Sethi for a year. Learn how the business of teaching functions at scale. Then make the business that enables the business of teaching.
(You can pick other examples; I picked Ramit because I think he's savvy, because he runs a savvy shop with lots of needs, and because my understanding of many of those needs maps fairly directly with things one could build.)
A variation of this advice: work for Teachable/etc.
A variation of this advice: sell courses for a living on whatever it is that you're an expert in.
I broadly don't think about "movements" that frequently. Digital nomads have many things in common with my lifestyle and many things very not-in-common. I describe myself as a "digital homebody" in conversations with them but don't e.g. think that there is some sort of great cosmic tradeoff happening.
Do I feel like living my entire adult life in Japan has colored my way of thinking about the world? Certainly so. Do I feel like having counterfactually chosen to live in e.g. San Francisco or Tulsa for my entire life would have colored my way of thinking about the world? Certainly so. I think people far underestimate the impact of their community, their style of engagement vis the community, their environment, their peer set, etc, even among people who think they are correctly calibrated on this issue, likely including me.
(It is extremely material to my thinking on this that I live in Japan but a lot of the communities which matter to me live on the Internet, FWIW. Like, I aspire to be a pillar of the community and responsible citizen locally, too, but it does not escape my notice that this sentence is in English.)
I think that cryptocurrency enthusiasts are constantly searching for a problem to which cryptocurrencies are the solution, and that the search is as yet basically unsuccessful. Feel free to tell me or the world generally why we have a non-fungible token shaped hole in our life.
B2B SaaS is a very large field of human endeavor. If you're looking for an entry point, I wrote one, but the field is multifaceted and I have no recommendation or set of recommendation that covers every facet.
I will say that producing beats consuming in business and that many people have a tendency to hoard great information about B2B SaaS businesses without talking to potential customers, selling SaaS, or coding SaaS, but that collecting guides/interviews/etc has built precisely zero SaaS companies and talking/selling/coding has created many.
I do not have a very developed point of view on the strategy here qua immigration.
I think over a 5+ year timespan you're going to increasingly see the world move to approximately two salary schedules for engineering jobs, one pegged to SF/NYC and similar "most expensive places in the world" and one for everywhere else. To the extent you believe this, you should probably believe that German engineering wages will narrow the gap with non-SF/etc US-based engineering wages.
I think very little about having "an audience" on a day-to-day basis. I think more frequently about the community I'm serving (software people, broadly writ) and the specific subcommunity of software people who know about me, have gotten value from my work, would happily break bread with me, etc.
I orient work-related parts of my life around those communities quite a bit and the existence of them / my recognition of their importance to me have been enormously positive for my life, across many dimensions.
I will acknowledge that it has occasionally been weird getting stopped on the street with "Excuse me, are you patio11?!" (Once in Mountain View, which makes sense to me, and once when walking around in Tokyo, which was extremely surprising.)
Relevantly to the first part of the question: I do sweat the question of "Are you a professional... social media influencer?" and anything which makes that answer round to Yes would be an alarming change in state for me.
Do you think you're going to get better at selling this after you have a product available? What is the narrative by which that happens? If you do not get better at selling this, running a SaaS company on one customer is rather tough.
I'd probably put most of my cycles into getting better at pitching or getting something better to pitch rather than building something given this thumbnail sketch of your present situation.
I like to think of myself as a patron of the arts, inclusive of the art of small businesses providing personal entertainment services via presence at poker tables. If you're asking this question, you should want to be at tables with me.
The direct answer to the question: I perceive I am ahead by a little of the average 1-2 table in Las Vegas, ahead by a lot in either in-person tournaments with ~$50 buy-ins, and ahead of an Pokerstars MTT with a ~$20 or less buy-in. (All assumes I'm rested and playing well.)
Fans of poker math will note that there's no way that playing at these stakes makes economic sense for me even if I won them 100% of the time, and I'm OK with that. It's a hobby for me.
I understand little about the entertainment industry, but to a gross approximation, I think it is largely an iterated game where ~2,000 people come together in clumps of ~10 to build a defined project on a tight timeline, and then they disperse again, and many of the clumps for a given project have worked together before, and if you stay in the game long enough you become known for being good at whatever hyperspecialized thing you do.
This does not describe the traditional practice of building startups, but I think as the field gets more understood and as velocity increases that there is a chance that this model gets adopted more frequently.
I could certainly envision a future where a lot of SaaS is produced by semi-persistent flocks of teams in a studio-esque system, with the goal of "Derisk the business, build the product, install the marketing/sales engine, turn crank three times, sell to PE, cycle rinse repeat." In this model, the people stay the same (out of some broader cast of characters) while the products rotate relatively quickly. We see the opposite in startups historically; a startup is, if successful, a 10+ year commitment, with different strata of employees staffing it at different times for 3, 4, 5 year stints.
If you predict we move to the entertainment model, you should also predict that there will be some people with "star power" who have played the game multiple times, won it multiple times, and are able to make marginal projects successful largely because other people say "Oh, X is signed onto this? Eff yes I want in on that adventure then." Serial founders are a bit of a thing in startups. Engineering managers (and sales managers, etc) quite commonly bring along people who worked with them previously. You would expect to see the impacts of these behaviors get much more acute.
Ask Erin and Thomas; they have expressed a desire to host it for marketing purposes sometime. That is, unfortunately, not a trivial endeavor.
(Blame the guy who wrote large portions of it.)
I'm sad it isn't available on the Internet today; it is by far the coolest thing I've ever built.
Speaking only for myself, I'm extremely skeptical that the cryptocurrency community has produced anything of value. 10 years. Billions of dollars. What has changed about the world? Some people got rich; virtually all of it through the losses of other speculators. Where is the value creation? Who benefits from cryptocurrency daily? What is the narrative, not the sales pitch for number go up but the actual honest-to-goodness prediction about the future, that the next 10 years and next billion dollars spent on this area of activity will create something that the last iteration manifestly has not?