The Tim O’Reilly Hypothesis: Build a Market by Building an Ecosystem

Tim O’Reilly is the founder of O’Reilly Media. In this episode of the Product Science Podcast, we talk about his recent article in Quartz, “The fundamental problem with Silicon Valley’s favorite growth strategy,” how he grew O’Reilly media from a $500 publishing budget, and how you can build a market by building an ecosystem around it.

Subscribe for the full episode on Apple, Google Play, Spotify, Stitcher, and more. Love what you hear? Leave us a review, it means a lot.


Questions We Explore in This Episode

The Tim O'Reilly Hypothesis: Build a Market by Building an EcosystemHow did O’Reilly Media get started as a publishing company with $500? What did Tim do to grow his business when they didn’t have clients? How did they get started with their first conference and turn it into a multi-million dollar events business? What made him decide to build his own ebooks business?

What are Tim’s issues with the winner take all model proposed in Blitzscaling? What are the problems with trying to extract maximum value from a marketplace, rather than focusing on creating an ecosystem? How do you focus on delivering value to users and content creators instead? Why do many tech companies fall short and are their business models actually sustainable?

Why does the most innovation happen in areas where venture capital is scarce? How are examples like the early internet, search engines, and web advertising examples of this concept? How was a commercial website radical for its time, and how did they validate that idea?

How are the best innovations driven by a need to serve users? How does O’Reilly Media broaden that view to focus on serving their content suppliers, and even in some cases their competitors? Why is it so important to bring the entire ecosystem up with you? How did Tim come to these realizations?

How does Tim work to teach government how to do tech, and teach tech how important government is? What are the problems that platforms are discovering as they try to optimize their engagement metrics? How did Code for America work with San Francisco to improve the process of applying for food stamps? How did his team create an MVP to solve problems and test solutions? Who do we most need to convince in order to change models?

Quotes From This Episode

Our product business was done in the cracks of our consulting business. When we didn’t have a paying client I wanted to keep people busy so I said let’s write a manual we know we need that other people probably need. - Tim O’Reilly Click To Tweet A marketplace is a lot like an ecosystem—it has to be circular. It has to keep revenues flowing back the suppliers. The typical model is a lot more like the extractive model in other parts of our economy. - Tim O’Reilly Click To Tweet How do we continue to serve our suppliers: our authors, our conference presenters? How do we make this work for them so they keep doing what they do? And in some cases, how do we make this work for our competitors? - Tim O’Reilly Click To Tweet If I had been in a venture capital-type environment I would’ve never done any of the things that I’ve done, because the timeline to get it right is quite small, and that’s why so many companies fail. - Tim O’Reilly Click To Tweet


Holly Hester-Reilly: Hi, and welcome to the Product Science Podcast where we’re helping startup founders and product leaders build high growth products, teams and companies, through real conversations with people who’ve tried it and aren’t afraid to share lessons learned from their failures along the way. I’m your host Holly Hester Reilly, founder and CEO of H2R Product Science.
Holly: This week on the Product Science Podcast. I’m super excited to share Tim O’Reilly. I reached out to Tim after reading his article on the problem with Silicon Valley’s favorite growth strategy and asked if he would like to come and talk about it with me. And I’m super excited that he’s here. So welcome, Tim.
Tim O’Reilly: Thanks for having me.
Holly: So maybe you can share in case some of our listeners, I’m sure they know who you are, but tell us a little bit about how you built up to this perspective that you have now, and what you think about the… what led you to write that article?
Tim: Yeah. So to understand my background, you have to realize that I started my business now nearly 40 years ago. We were technical writing consulting company. We started saying that our clients all needed the same kind of manual. So we started retaining the rights and then we started publishing them as books. And so we became a publisher really for the Internet era. Very early on.
Tim: We published our first books in the mid 80s actually in this, to give you the idea of how far we were from the idea of a high growth tech company of today. The first print run of our first books was a hundred copies each and they sold for $5. So that was like a total potential revenue from that first print run of $1,000. And I started my company with $500, which was in the form of used furniture.
Holly: That’s awesome.
Tim: Basically, our product business really was done in the cracks of our consulting business when we didn’t have a paying client. I wanted to keep people busy. So I said, “Let’s write a manual about something that we know we need that other people probably need.” So we had this focus on the user and actually it was very central to who we were. It’s sort of astonishing to people today to understand how far we were from being user centered and the computer industry.
Tim: When I began as a writer in the industry, it was not considered appropriate to write in the second person. You simply describe the product, the device does this, the program does this. And it was all for the person. And one of the great revelations when we’re writing our own books, we could just talk to the user. It was also for bidden for example, writing a manual for one of our early clients to say, this feature is junky or doesn’t work. And again, once we were able to write our own books, we were able to say, “Hey, you’re not crazy. This doesn’t work very well, and explain work workarounds and so on.”
Tim: So we were the voice of the user that who was a little ahead of you, who you wished you had looking over your shoulder. And so our business grew from there and we’re now roughly a $200 million a year revenue company. But each piece of it has been just paying attention to user needs and trying to fill them. Our conference business, which we launched in 1997 came about because our bestselling books were on the Perl programming language.
Tim: Sun Microsystems company, it was one of these high-flying companies that no longer exists. Had just launched a conference called Java one for the Java programming language. And I said, “Who’s doing that for Perl? Nobody is.” And so I really thought of the first conference really is just marketing this open source community. And it was just amazing coming together, people who had only known each other online and we said this is great, so let’s do it for other communities. So that led to, what we call it, the Open Source Software Convention, which was really our first conference.
Tim: And then from there we went on and built a multi-million, a multi-tens of millions of dollars events business. We launched the third laying of our stool in 2001 when there were a lot of high growth startups again in the eBooks space. We had been playing around with eBooks since the late 80s. We have some history there that is too deep to go into. But so these people came to us and we said, “That’s a shitty business model for us.” We’re not going to participate because they were telling me we have this subscription and it’ll be a thousand books and it’ll cost 99 bucks and we’ll share $20 on that with the publishers.
Tim: And I did the math with them. I said, “So you’re essentially offering me 6 cents a book per user. We can’t live on that.” And so I decided to build a service that was, and this is really an important thing, it was both user and supplier centric and that we were mindful that we had to make money for our authors, if we wanted them to keep producing in this future. And we also went to our biggest competitors and we really, so we built a marketplace business very early on 2001.
Tim: But if you look at the history of a lot of marketplace businesses, they end up screwing their supplier side a lot. In fact, I’m writing right now a sequel to the blitz scaling article which is really about the way more and more of the real estate on the Google search results page is devoted to Google and Google’s products themselves. And we’re also in the position that Google is then of being both a platform owner and a platform supplier i.e we’re competing on our own platform with our suppliers. But we’re really mindful that we have to actually be fair. And whereas I wrote about this in the blitz scaling article, you can see that Google, which started out with 50, 50 revenue from advertising on and only 18% on third party websites.
Tim: Now there’s maybe a lot of reasons for that and I go into that more in this new article I’m writing. But by contrast, we’ve really ended up maintaining a pretty consistent share of the revenues from the platform because we’re really mindful that if we take too much of the pie, our suppliers are going to go away. And so I’ve really ended up starting to develop kind of a theory of marketplaces that is built around how the way an efficient marketplace, it’s a lot like an ecosystem. It has to be circular. It has to keep revenues flowing back to the suppliers.
Tim: And the typical Silicon Valley model. It seems to me it’s a lot more like the extracted model in other parts of our economy. The goal is to keep growing at all costs and that you see these companies that start out as initially value creating companies for the help whole ecosystem. Back when I was first in starting my business, the big dog was Microsoft and they were creating value through building this universe of the personal computer software. And then bit by bit they took off all the best applications, and everybody then went to the Internet, which wasn’t yet commercial and where there was a green field opportunity.
Tim: And now I’ve watched over the last 15, 20 years as the greenfield opportunity of the Internet has been slowly blogged, if you like, by the winners. So we have this winners take all mentality, which is the other way of really looking at Reid Hoffman’s Blitzscaling book. He’s literally saying, all these internet businesses are winner takes all markets. And so you should basically be the winner. And here’s how to be the winner.
Tim: And my belief instead is that yes, there are some people who become the winners, but they have a responsibility to the losers. And it’s not that the others are losers. You can have a win win, you know? So if we had seen when we built our online platform, which was originally for eBooks, but eventually became for all kinds of online learning, we have video. We have interactive of content of various kinds, even the live training that’s delivered on over the net.
Tim: If we had seen this as a zero sum game with our competitors, we would have taken more and more of the value on the platform. And instead, we literally, we say, oh, we introduced this new live online training feature. Our supplier partners didn’t get onboard, even though we evangelize the feature of them. They didn’t get on board as quickly as we did. So our largest supplier, their revenue fell by half in the next month because it was such a popular feature. Then the revenue is allocated by usage of which content products people use.
Tim: And what did we do? We didn’t do a little victory dance. We actually dropped the price of our products radically so that we didn’t take too much of the pie. But we’re trying to calibrate it so that they would be incentivized to produce more of this content that we wanted on the platform. And so this balancing act of how do you actually serve users but also serve the people, make it possible for others to serve users through your platform, I think is central to my thinking.
Holly: Yeah. So can you actually tell us a little more about that? I think so I’m super fascinated about the ecosystem and deciding which levers to pull when you want to change something about it. And I’m curious how you made the decision with what you were just describing about what part of the pie they get and what in you dropping your price. Can you tell us a little more about what went into that and how you made the decision that that lever versus one of the other levers was what you should have moved?
Tim: Well, we didn’t have very many levers. I mean, we have an algorithm for allocating revenue on the platform and it’s the price of the product times the usage of the product. And so since our usage was going way up the only thing we had to put down was the price so that we didn’t take too much of the pie. But why did we do it? I think that really goes back to our original orientation, which was that we realized that we were an intermediary between a network of people who knew something and a network of people who wanted to know something. And we didn’t know everything ourselves. I mean, many of our original books we wrote ourselves. We were really writing firm that started publishing our own books about things that we had learned.
Tim: But then we realized, oh, there’s all these things that we don’t know. And of course we started publishing books from others. And we have to incentivize those people. We have to give them value. And some of that value was in the form of revenue and royalties from the books, the successes of books. But some of it was also from the exposure, people who say that, hey, I got my new job because of the reputation that you helped me build. And then that led very naturally to the events because that was a way of not only of sharing knowledge, but of helping people build reputation.
Tim: So again, we’re thinking a lot about not just the benefit to our users, but the benefit to the people who will benefit with the users. And it’s a big contrast to how so many businesses operate and how you see the stories about Uber and effectively the feeling that the drivers who were originally given this great come on, it’s this great opportunity. You’ll make all this money. And then bit by bit they start squeezing the drivers while continuing to favor the users.
Tim: And it’s just not sustainable. Just any more than if you look at the changes in farming, where it was like, well, we’ll just have to keep putting more artificial fertilizers because we doing this extractive agriculture and something like half the top soil in America or worldwide has gone through that method and people start to go, oh, we actually have to discover how to build regenerative agriculture that builds the soil again. And it’s very similar. You have these companies that where venture capital is substituted for a real regenerative living system. Uber and Lyft are on the drip of venture capitalists, just like massive inputs of fertilizer from Monsanto keeping our farming ecosystem going and it doesn’t end well.
Holly: Yeah. So what do you think… you have such a breadth of years seeing all of these systems at play. Have there been periods where more companies were on the drip from venture capital and periods where it wasn’t as strong and how do you think that affected which companies were winning and not winning?
Tim: In my experience, most innovation happens in periods when venture capitalist is scarce And in areas where venture capital is not interested. If you think about the great companies of today, Apple was a service rogue company and this early PC era when all the giants of the industry were saying, “The PC is just a toy. Why bother?” Both Apple and Microsoft came out of that era, in the web Google was in a category that literally all the venture capitalists had abandoned. They said, “Well our search engines are… well, am I do anything?”
Tim: Amazon very early in E-commerce and in a category that nobody thought really was that important of what it’s a bookstore. And in fact, when we are actually… my company created the first commercial website and nobody thought that Internet had commercial potential. We were really on the very forefront of people. In fact, it was illegal for the Internet to be used for commercial purposes. And we had to get a special dispensation from the National Science Foundation to run ads on GNN, which was the Global Network Navigate.
Tim: It was like a predecessor of the Yahoo started about a year before Yahoo portal was the first site that had advertising. We actually had to do market research telemarketing where we had to call off 50,000 people and do research on demographics of internet usage because nobody believed that advertising could actually work on the Internet. And I should also add that our version of advertising was not the banner ad you could see today. And again, this also tells you something about product development.
Tim: You are shaped by who you are and where you are. At the time we were a publisher who sold largely through direct mail. We had actually begun to be discovered by bookstores. They found us, we didn’t see them as a target market. We sold directly to consumers. And so our experience of advertising was how do you get information about products to people? And at the time there were these crazy things which people of your generation won’t remember at all.
Tim: But you got a print magazine or newspaper and there was a tear off card in the middle that looked like a Bingo card, had little numbers, zero to 40, every advertisement had a number associated with it and you would circle the ones that you want to get a brochure mailed to you. And so what we realized was that the web would mean that all these brochures wouldn’t have to get mailed out anymore. People would just go to a website. And so was the idea of the commercial website that was literally a description of the products and services of a company that were the first Internet ads. Because up to that point, there were no commercial sites on the web.
Holly: And at that point in time you had to get legal permission to put that on the Internet?
Tim: Yeah the Internet was under the auspices of the National Science Foundation. And they had something called an acceptable use policy. And it was basically the Internet was for research and education. And it was funny because I had a conversation with this guy named Steve Wolf, who at the time was the administrator for the Internet at the NSF. And I explained what we were doing and he understood that we weren’t talking about sending out ads to people. We were just talking about building a site that had commercial content.
Tim: Because of who we were, it was mostly was educational informational content. And he said, “Well, if you guys aren’t about research and education, I don’t know who is, so go for it.” And it was a turning point on the Internet because there were a big mailing list. It was a mailing list called calm free of commercialization and privatization of the Internet. It was a government thing and the government set the rules and then they set it free.
Holly: Wow. That’s a fascinating tidbit. I had no idea. Did the stuff that you put on the Internet back then really stick within this realm of, well it’s commercial, but it’s all education? And what was it like when you started to see things that weren’t about education?
Tim: Well, it became, it became ghastly. I mean the Internet, when the banner ad was invented, we’re seeing again, whether it be flashing, blinking, and that’s why like Google came out in 1998, and they basically said, “We’re not going to do any of that.” They just said, “Screw it. If that’s what it takes to run a website, we’re not interested, we’re just going to build a better service.” And so Google originally had this wonderful model where they produced the search results.
Tim: They had a better model for the product, but they also had a better model for the business. They said, look, a really good ad is going to be just like a search result. And they were just two parallel columns. It’s like, this will be something that people want and we’re going to try it rather than selling to the highest bidder, the original breakthrough Google’s advertising was, we’re going to sell… They had an auction model, but they also took into account the likelihood for the ad to be clicked on.
Tim: So if you think about how you make money on, it was also pay per click, which was not invented by Google, but it was invented by a company called Overture, which then got bought by Yahoo. But Google perfected the model and the model was instead of paying for, or you got a million impressions of your head or 50,000 impressions or 100,000, it’s this much per impression. It was like, we’re going to only charge you when somebody clicks on your ad.
Tim: And Google had figured out that they could do the math and say, well, an ad that somebody bids in the auction say, let’s just say a dollar for this position. And somebody else bids $2, but the $1 ad is going to be clicked three times and the $2 ad is going to be clicked once. We’re going to make more money from the $1 ad, right? So they’ll get $3 instead of two. So they actually had this really interesting innovation, which also was about serving their users.
Holly: Yeah. I think that’s great. It comes back full circle to their system had people on both sides and they designed the system in a way where the more value that was created, the more value everybody got including them.
Tim: Yeah. And I think that it’s just one of the missed pieces in so much of the Silicon Valley dogma, which tells you to focus on users. And I think the real secret is you have to focus on everybody who matters. Everybody who’s a stakeholder in your system. So when I think of our product innovations, a huge number then, come from thinking about how do we continue to serve our suppliers, our authors, our conference presenters. How do we make this work for them so that they will keep doing what they do?
Tim: And in some cases we even said, “How do we make this work for our competitors?” When we started, well at the time it was called Safari Books Online, and now it’s just integrated into O’Reilly. We literally invited our biggest competitor who had a group internally that they called the O’Reilly. We literally invited our biggest competitor who had internally that they called the O’Reilly killers to be our partner in the venture, because we thought that it would be better for the whole ecosystem if more people were at the time producing eBooks.
Tim: Remember, this is seven years before the Kendall is 2000, Kendall came out in 2007. eBooks were this new thing that publishers were very dubious about. And we said, “Well, if we just do it from O’Reilly, we’ll get a small, dedicated set of cutting edge users, but if we want to really take off, we have to get in the whole ecosystem onboard.” So we spent a lot of time evangelizing, not just our customers, not just our authors, but also our competitors. And even publishers in other parts of the industry because we realized you’ve got to bring the whole market up.
Holly: Where did you develop that learning? Because I do feel like it’s something where, when you explain it, it’s like, yeah, that makes sense, but I know I talked to business people who would fight tooth and nail about doing that. So how did you come to that realization?
Tim: It’s hard to say. I think some of it had to do with the fact that we were on a slow burn. We weren’t on this race where we had raised a bunch of capital and we had to get to some predetermined point that would make the venture capitalists happy. We were just trying to feel our way forward. We were exploring and thinking.
Tim: Now I do think our values were part of it. I grew up in a household where my father used to borrow money so he could meet his charitable obligations. I think, I had that background, but also I spent a lot of time, early in my career, before I was even in the technical writing area, I had written a book about Frank Herbert, the science fiction writer, author of Dune.
Tim: And he dealt a lot with ecological themes in this Dune series. And so I’d spent a bunch of time looking at ecology, not to mention being a fan of the whole earth catalog and that whole period. So I had ecological metaphors deeply rooted in my thinking. And so those were framing ideas for me. Natural Systems teach us a lot about the world. And when we think only of, again, back to this Blitzscaling idea, which is if you strip away the technological trappings, the Silicon Valley, Shiny Patina and you use that to describe, say an oil company, Blitzscaling or a gold mining company or an agribusiness swallowing up all the small farms everybody would go, is that really the way we want to build our economy? Is that really right?
Tim: And in some ways, maybe I’m a throw back to a different economy, but I think we can. I think it’s also the next economy because as we do want big globe spanning platforms. They create immense opportunity and information liquidity and in the same way you think about containers shipping, how it’s enabled free trade. But we’ve got to pair that with a kind of ecological circular model where we don’t turn those platforms into extracted machines, but into supporting platforms for a much more holistic economy in which everyone wins.
Tim: And of course you see this in the political debate today. Where people are starting to say, again, windmill. We’ve built an economy where some people win really, really big and other people don’t. And again, it’s not that simple because even in the parallel between say an Uber in which passengers really win and Uber really wins but drivers get screwed. It’s kind of analogous to the free trade regime that we have in today where giant corporations really, really win, people in developing countries really, really win. And American you know, workers were left out in the cold.
Tim: And so the question is how do we build better markets? There’s a book I love. It was actually given to me by Uber’s chief economist is called Who Gets What-and Why and it’s by Alvin E. Roth, is a Stanford economics professor who basically founded a field called market design. And he got a Nobel prize for his work on redesigning a kidney transplant marketplaces so they could be more efficient. And this idea has obsessed me over the last four or five years has been the great opportunity of AI and big data systems is actually to design better markets.
Tim: And we think that we have this optimal market where it’s just the invisible hand and people competing and so on. But we’re really moving into a world of algorithmically controlled markets and have the best intentions, we’re going to bring the world together by sharing and discover that the market you’ve created is actually a very bad market. And then you have to redesign and intervene. And so this, what I’m really spending a lot of time doing is trying to wake up Silicon Valley to the idea that the market is not a given, that it’s something you have to design and you have to design it with everyone in mind.
Holly: Yeah. I love that. I’d never heard of this the book Who Gets What-and Why, but I will check it out because it sounds like it’s a lot of what I’ve been thinking about and why I was so fascinated by your article and perspective is not just markets but government and entire systems of society and organization, and how AI and big data and all of this infrastructure we have today. I mean, it just completely changes the raw materials that we have to design how the world works and-
Tim: Yeah, exactly.
Holly: … what if we think about that instead and what was this market supposed to get us and how do we redesign it with the materials we have today? It could do an even better job.
Tim: Yeah. In my book, which I published in 2017 called WT? What’s the Future and Why It’s Up to Us. I spent a lot of time really building some of the mental toolbox for thinking about market design, and looking at, for example, the way Google search works so well because they took hundreds of factors into account, they had an overriding objective function or fitness function, which was relevance for the user. But they took all of these different factors into account. And I go, could we start to do that in the real world?
Tim: But I also look at why do those systems break down? And I actually use the analogy that are the big data and algorithmic systems and AI that we build today. There’s a lot like the genies in Arabian mythology. If you’ve ever read Thousand and One Nights or even just seeing the Loudon, the story is always of you tell the genie what you want. You get three wishes or whatever and you always get them wrong. And that’s I think great mayhem results.
Tim: And so Facebook got their genie, and they told it and YouTube too, like do things that people find engaging. And unfortunately the things that people found engaging were not always healthy for them. And now Facebook is basically trying to give the genie other commands and this whole idea of debugging. And by the way, it’s one way back in my career, there was a guy named Andrew Singer. I was writing a manual for his company, and he says something marvelous to me, which I remembered ever since he said, “The skill of debugging is to figure out what you really told your program to do instead of what you thought you told your program to do.”
Tim: And it’s so relevant for algorithms, what is it really doing versus what you thought it was doing. And we have to actually do that at the level of our entire economy. There are people who said, who designed our economy just as surely as Google and Facebook design their algorithms. And they say, well, we’re going to have these tax rates, we’re going to have these incentives. We’re going to have these rules about trade, and it’s going to have these positive outcomes.
Tim: And suddenly they’re finding, well, the outcomes weren’t quite what we thought. And this is often on large timescales. I mean, this is true with a company like Google or back in the day, Microsoft, it took 20, 30 years to figure out that there was something slightly wrong here, and that needs adjustment.
Tim: And I think that we’re in a phase right now where we’re saying, wait, these economic algorithms that we put in place, are they really working? And that’s why I’ve been very focused really on thinking about the analogies between the great tech platforms and government as a platform. That’s just something I’ve been on for the last 10 years and I still a lot of work on that with my wife Jennifer to talk it through the organization.
Tim: She started Code for America, which really is trying to teach government how to do tech, and to teach tech people how important government is and how we actually… because this is a giant platform that we’re building, that we want to serve all of society, and we want it to be a generative platform as opposed to an extractive platform. And how do we do that and what lessons can we apply?
Holly: Yeah. In that work, what have been some of the most surprising lessons to the tech people that we’ve learned as you’ve tried to work with government?
Tim: Well, probably the most important lesson is one that it really isn’t about the tech. It’s about paying attention to the users which is or and in some cases, paying attention to the government as an intermediary. So if you look at the two biggest services that we have today rolling back 10 years, Code for America started as a fellowship program. Small teams of developers, designers are going out and working with cities. And we did that for three or four years.
Tim: And then we hit on a couple of projects so we said these need to go to scale. And the first of those was a project that had been started with the city and county of San Francisco to improve basically the process of people applying for food stamps because California was spending millions of dollars actually literally on the system, hundreds of millions of dollars, almost a billion dollars to try, and get people their benefits.
Tim: And it started in San Francisco. They’re like, “Our participation rate is abysmal. People who are eligible aren’t signing up, why is that?” And they originally thing, well you guys will make an app for that. We didn’t make an app. That was not how we started. How we started. It was just debugging that process. And we said, “Well, you guys have a digital front end. You’re spending millions of dollars advertising this digital front end.” It was called MyBenefits CalWIN and you start with the name pretty terrible, but and a CalFresh is the name of the program in California. And he said, “Let us try applying.” And they discovered that A, it didn’t work on a mobile phone and it didn’t work in libraries.
Tim: And those are the two ways that their target audience would try to use an online service, right? Like this. They don’t handle network home computers often. And so why didn’t it work on library computers? Well, the library computers have a half hour timeout and their application took an hour to fill out. And they gave people no way to save their work. They clearly never tested it in the environment of the people who they… and why did it take an hour? Well, because it was this consortium of counties, this is where you have to look to the supplier side and say what’s wrong here?
Tim: And they’d all stuffed in their own questions. And it turned out you only needed a half dozen questions to get an application started. So we built a little mobile app that took the minimal set of questions, and it was really a very much an MVP and the lean startup sets it literally people filled out a form filled out this mobile app. And then, we generated a fax, which we sent in to the social services office.
Holly: That’s amazing.
Tim: Yeah. Over time, we actually built in the application, we figured it out by actually texting with the users, and we started finding, again, you had to work both sides of the market. You had to listen to the users and go, oh, and by the way, the thing was open source, and it spread to six or eight other counties and that eventually the state said, “Hey, we want you to do this for the whole state.” And we’ve built a model of where of this really took over and now we’re taking it to other states.
Tim: But the critical thing was just debugging the process by texting, which is a super power that we have the government doesn’t really understand texting with the users as they went through the process. Just heavy, intense follow your users, and you go, oh. I had to start over because I missed my appointment because they send out a paper letter telling you the date of your… it’s a telephone appointment, but they’re going to call you and if you miss it, you have to start over. And it turns out that the 25% of the letters in one county were going out after the date of the appointment.
Holly: Wow.
Tim: And so they were just throwing away half of the 25% of they users just because of bad process. Other cases, we just did pipeline analysis. So when you send out this form or you asked this question, this is where you lose 10% here use 5% here. Things that Silicon Valley knows, but there really was no process on the government side to actually even asked those questions. So we became an intermediary help being the voice of the user into the government and changing their processes. And Jen some times it calls it that apps to apps, you’re building an app to instrument the government process so that you can make the process smarter.
Holly: Were there any hurdles that you had to overcome with rules and regulations around talking to citizens about using these things or was that pretty straight forward?
Tim: Well, part of our success I think was that we were a quasi rogue operation and we had a partner inside government, but we weren’t actually working for the government, particularly after we did. We did this first fellowship year. But it really was this real entrepreneurial group of fellows who after the project year ended, the model normally was we’d do a project and then we would turn it over to the government and maybe try to hire the fellows or they would figure out other ways to carry it on, some times the fellows say, oh, this could be a startup, but we actually birthed about 10 startups in the early years.
Tim: But this one, we had three fellows who said, “We just want to keep working on this.” And they literally, we scraped together a little funding, although I think originally they were just doing it because they didn’t want to quit. And so they were rogue operation and that gave them the freedom not to ask permission. And so we’ve really built a model now where we build things with philanthropy, and then we scale them with government dollars.
Tim: So get CalFresh is now paid for almost entirely by the state of California. Through basically they had to tweak some of the rules. They realized they had a source of funding where they would pay food banks to help sign people up and they said, “oh, well we can just use that same money for you guys to do online outreach.” But we built it with philanthropy because that gave us the freedom not to have to ask permission.
Tim: And I think that is often difficult in a government context, but usually it’s a partnership on other signature project today, which is sort of clearing criminal records that are eligible to be cleared from, if you just give you a background on that, people who have served their time or maybe were had a criminal conviction for something that’s no longer a crime, such as marijuana usage, and actually it turns out that was a proposition. Another proposition 47, which turned a whole set of things from crimes into misdemeanors.
Tim: There’s on consequences if you have a crime on your record. You can’t get public housing. It turns out you can’t even drive your kid on a school field trip. You can’t get off and can’t get a job. So clearing your records a big deal. And people had spent tens of millions of dollars getting these propositions passed, but they never thought about the implementation. And that’s again, part of the key insight that we’ve had at Code for America, which is the implementation matters. When people pass a law, if they don’t think about the implementation, it’s just may turn out to be nothing like what they imagined.
Tim: And in this case with these propositions that decriminalized various things, their implementation was, well, people who have these crimes on the record can apply. They have to go to the DA’s office, get their record, they have to fill out a bunch of paper and they probably have to hire a lawyer. Again, not thinking about the actual users. They can’t afford to do this. They don’t have the time to do this.
Tim: And so, despite the tens of millions of dollars to pass these laws, only a few thousand people have taken advantage of it. And so we started out actually building an app. It was a lot like get CalFresh. It was designed to help people through the process. And then we realized this process doesn’t need to exist at all. And actually are about the same time the District Attorney in San Francisco realized the same thing. And he was like, “Why are we making people apply? We should just automatically clear the records.” And unfortunately, they weren’t tech savvy. And their idea of automatically clearing the records was hiring a bunch of paralegals who would fill out the forms without people having to ask.
Tim: But it was still as painful process because this is just changing our record in the state’s own database. Again, so we built a hack from the outside, which is, okay, we’ll download all the records for San Francisco. We had already built a little rap sheets, they’re coded like medical records. You’ll see all these strange little codes and you have to know where they are. And we’d written the program that could read a rap sheet and decide, oh, well this is a violent offense, so therefore, it’s not eligible. This is a nonviolent offense for something that’s been cleared. You’re eligible. Good. You’re good to go.
Tim: And then automatically fill out the application to the court to get these clear. We cleared 9,000 in a couple of hours, and it was in San Francisco and they were like, “Oh my God. It was that easy?” And so we’ve taken that actually statewide and nationwide also. But it’s this like, sometimes you have to get close enough to the process to understand what’s wrong with it and then throw it away. And that’s another key piece of a product design, which is getting close enough to make it better.
Holly: Yeah. I think there’s a lot of lessons in there for product. Understanding the real problem you’re solving and not just the problem someone tells you you’re supposed to solve for them.
Tim: Yeah. And that’s often an iterative process. Throughout my career, I’ve watched the game of leapfrog and I actually, in my book, I talk quite a bit about the history of online ride hailing in that way, all of the steps, and this came from one come in, this came from another and how they leapfrog each other to finally develop the model. And I think the same is true in many other areas online eCommerce, how various obstacles had to be overcome and somebody would take a step forward and then suddenly go, oh, if we can do that, then we can do this as well.
Holly: I’ve been reading your WTF book and so have some more ways to go. And I’m wondering what are your prescriptions on the most effective way for us to overcome these challenges? Like sometimes, these things they just seem like there’s herculean task of trying to write the system and I’m curious, I can see that you’re spreading the message and you’re writing and talking about it. But who are the people we most need to convince? Do we need to convince the founders or the venture capitalists or the citizens or everybody, what do you think?
Tim: Well, I think in some ways we need to make space for people to build different kinds of businesses. I think that if I had been in a venture capital type environment, I would never have done any of the things that I’ve done. Because the timeline to get it right is quite small. And that’s why so many companies fail, right? You don’t really have much time to get it right.
Tim: I’ve been doing my business for 40 years, and in the course of that, I built four or five, maybe six businesses within my own business, some of which we’ve continued, some of them had a great random success. So, it’s a now forgotten piece of history. But O’Reilly created not only the first commercial website but we also created the first PC-based web server which we sold and validated the market because we said everybody who has a web browser should have a web server.
Tim: I grew the business about $3 million before Microsoft woke up and said, “Oh, we really need to do this. Sorry, we have to kill you.” But we were okay with that and not we would’ve liked to have that go on and succeed, but part of our goal has always been to catalyze the market. And having the ability, because everything we’ve done is pay as you go. We try things and we just have had a lot of time to tweak and explore and figure out what things really resonate with us, and we’re not desperately going, we have to hit this some hyper growth need.
Tim: We basically, had slow and steady growth over a very long period whaling a typical venture capital firm only the last 10 years. They have no provision for company to last 40 years unless it exits. And we need a lot more slow businesses as I’m wondering, say we need fewer startups and more finish ups, and that’s how business really used to be. Let’s not say there isn’t, there’s a real place for venture capital. It really can create enormous fuel for a rocket ship, but if you’re not a rocket ship, and you try to take on rocket fuel, usually burn up.
Tim: And so what we’ve been trying to do in our venture firm one of the businesses we started as a small, early stage venture firm called O’Reilly Alpha Tech Ventures, but my partner, Bryce Roberts in our fourth fund is really focused on something he’s calling Indie VC,, which is really the message to founders is we want to be the last money you need to take. The goal is catalyzed companies that are focused on cashflow and profits, and they actually pay the investors back through dividends rather than through an exit.
Tim: Now, not to say that, if you hit a rocket ship, then you can take, additional venture capital go for the exit and we’ll convert and go along for the ride, but you don’t have to. And because there are a lot of businesses that are not venture bubble that get burned up by the need of venture capitalists to have them grow faster. And you can have a very, very nice business that serves the world without becoming a rocket ship.
Tim: And I like to think, we’ve had a lot of impact on this industry. We’ve created employment for thousands of people over the decades, and creating value should be part of that circular economy. And I think we’ll get there with tech. I do think that this is period in an industry where pouring so much fuel on the fire does lead to a lot of churning through opportunities. And the problem with that for entrepreneurs is you’re part of the fuel. You’re not actually the guaranteed winner. The investors are the guaranteed winners.
Tim: It’s interesting because VCs will say in their defense, well, most VCs don’t actually make their returns. Yeah, but they get enormous fees in the meantime. They get paid very well, even if they’re failing. And entrepreneurs are often just the hired help in the VC model. And yes, some of them get really rich, but most don’t. And so building a business that you can control serving people at a much slower pace is often… well, let me put it this way, it’s another model that entrepreneurs should consider. So if you go to the NDVC site, you’ll see sort of some of the background on how Bryce is approaching that.
Holly: Well, to me that is so important until, and I think it would be an understatement to say that O’Reilly media has made a huge impact. And the idea, a lot of entrepreneurs want impact or lasting, they want to do something big. And sure a lot of them want to get rich, but I think there’s a lot of people who are looking for other things as well, especially in this day and age with the way the world is today that I feel like the more stories we can share about this kind of success and impact and all the things that you’ve spun off and the people who’ve grown from the content they get is I think you’re right that it’s something that you if you had tried to raise venture funds to do this, you wouldn’t have had the time to learn and adjust and make sure it was making the impact you wanted.
Tim: Yeah, I think coming back to the regenerative agriculture model, one of our slogans at O’Reilly for many decades has been create more value than you capture. And if you think about farming in the regenerative model, the soil is better after you finished growing your crops. In the extractive model, the soil is worse. And that is the secret to a circular system. You make everything around you better.
Tim: And there’s a wonderful passage in Victor Hugo’s Les Miserables, the novel that was the basis for the musical in which the main character at one point is a businessman and there’s this description about how he made everyone around him more prosperous. And I think that should be our ideal, that we live our life, and we run our business in such a way that we make everyone around us better. And that doesn’t mean that we don’t compete. There’s a kind of healthy competition in which we make our competitors better as well, and they make us better. And there’s a kind of unhealthy competition where it’s seen as winner takes off.
Holly: Yeah. I don’t know if you’ve seen, there’s actually some data on, this is just a specific example, but one of my favorite ways to underscore point for people who are skeptical is there is some data on pricing models for subscription SAS companies. And how the different pricing models correlate with growth. And one of the things that I’ve seen, and I can put it in the show notes, is some data that when you tie the pricing to the outcomes you’re driving for your customers.
Holly: When you tie it to the value that they’re getting, and when your pricing scales, but the more value they get from you, the more they pay you, but they don’t just pay it all up front. That is the best pricing model for growth. And I think it comes back to this idea. If you create more value than you capture, you will capture more value than if you just capture all the value, because then there is none left to be creating.
Tim: Yeah. I think that’s absolutely right. Well, on that we’re probably at a place to wrap, that you think?
Holly: Yup. I think so. I really want to thank you, not just for talking with me today, but I also wanted to mention that when I got my start in tech, I had not studied software and my husband had a Safari bookshelf and that’s how I started learning about the software world. So I feel like enormously grateful that you created that product. And this was back when it was like PDF eBooks and other forms of eBooks before the Kendall.
Holly: And I read the Cathedral and the Bazaar, which I know you wrote about in your book. And that’s how I learned what product management was and how I managed to walk into companies being the person who knew more than the others about agile software development. So thank you so much for creating this ecosystem because it’s impacted me personally as well as tons of people around the world.
Tim: Well, just to be clear, we didn’t create it. We tried to help grow it. And I think that’s really central to our story is that, our mission statement is changing the world by spreading the knowledge of innovators. We don’t see ourselves as green. We see ourselves as helping people who are inventing the future and we’re trying to support them and to share what they know and help them share what they know with others so that they can follow in their footsteps.
Holly: Well, thank you so much for doing that. It’s so valuable.
Tim: AlL right. Thank you.
Holly: The Product Science Podcast is brought to you by H2R Product Science. We teach startup founders and products leaders, how do you use the Product Science methods to discover the strongest product opportunities and lay the foundations for high-growth products, teams and businesses. Learn more at Enjoying this episode, don’t forget to subscribe so you don’t miss next week’s episode. I also encourage you to visit us at to sign up for more information and resources from me and our guests. If you love the show, a rating review would be greatly appreciated. Thank you.