How to build companies that matter

Eric Ries, a serial entrepreneur, most recently was co-founder and Chief Technology Officer of IMVU, his third
startup. He’s the author of the blog Lessons Learned, co-author of several books including The Black Art of Java Game Programming (Waite Group Press, 1996), and a Venture Advisor at Kleiner Perkins Caufield & Byers.
He. In 2007, BusinessWeek named Ries one of the Best Young Entrepreneurs of Tech. He’ll be presenting on “The Lean Startup: a Disciplined Approach to Imagining, Designing, and Building New Products” at Web 2.0 Expo.

We’re living in a time of renewed possibility for startups. Major trends – from the pain of the economic crisis to the disruption of web 2.0 – are breaking the old models and paving the way for a new breed of company. I call it the Lean Startup.

The Lean Startup is a disciplined approach to building companies that matter. It’s designed to dramatically reduce the risk associated with bringing a new product to market by building the company from the ground up for rapid iteration and learning. It requires dramatically less capital than older models, and can find profitability sooner. Most importantly, it breaks down the artificial dichotomy between pursuing the company’s vision and creating profitable value. Instead, it harnesses the power of the market in support of the company’s long-term mission.

Tim O’Reilly has recently been advocating that as an industry we focus on building stuff that matters. In response, I want to try and present a way of building startups that can realize that dream. In particular, he as articulated three principles:

(1) Work on something that matters to you more than money,
(2) Create more value than you capture, and
(3) Take the long view.

Given the hype and easy credit that has been the hallmark of technology startups the past few years, it’s been too easy for them to be unclear about whether they are really creating more value or just spending money to create the appearance of success. The lean startup approach tackles this problem from the very beginning of a startup’s life. My experience is that startups need to be built from the ground up for learning about customers and what they will pay for. That means an obsessive focus on finding out “is our company vision really the path to a brave new world, or just a delusion?”

Read the stories of successful startups and, if the founders are willing to be honest, you will see this pattern over and over again. They started out as digital cash for PDAs, but evolved into online payments for eBay. They started building BASIC interpreters, but evolved into the world’s largest operating systems monopoly. They were shocked to discover their online games company was actually a photo-sharing site.

Each of these companies were fortunate to have enough time, resources, and patience to endure the multiple iterations it took to find a successful product and market. The premise of the lean startup is simple: if we can reduce the time between these major iterations, we can increase the odds of success.

And here’s where working on something that matters to you more than money is critical. When you’re committed to something larger than yourself, every minute counts. Hype and transient success won’t keep you going. But the simple process of finding out whether or not your vision is right will. Because people who are dedicated to the truth are more likely to fail fast, learn, and try again.

This is one core tenet of the lean startup approach, called customer development. It has its roots in previous eras of startups (you can read more in the original customer development book, The Four Steps to the Epiphany), but changes in the industry are making it possible to iterate much faster than ever before.

There was a time when many technologies required proprietary licenses, big company permission, or custom deal-making ability to access. Those days are rapidly coming to a close. Most technology startups now have access to a staggering array of high-leverage technologies: free and open source software (and, increasingly, hardware), the data-driven services of web 2.0, user-generated content, and cloud computing, just to name a few. What all of these trends have in common is the increased leverage that development teams enjoy, meaning that for every ounce of effort they expend in building product, they take advantage of the efforts of thousands or even millions of others.

There’s no need to tell Radar readers that these technology trends make it cheaper to make new products. What’s striking to me is that they enable teams to make new products faster. It’s the speed with which companies can move through product iterations that will define this new era. Those that can experiment rapidly will be more likely to uncover what customers truly want; those that take advantage of these high-leverage technologies will be able to experiment the fastest.

Another way smart startups can work faster is to adopt agile product development practices. In traditional waterfall development, which assumes a top-down plan and stable development towards a well-defined outcome, a large number of software projects fail outright (and the lucky ones come in way over budget and way late). The rise of methodologies like Extreme Programming and Scrum has enabled teams to spend more time focused on focused on building products customers actually want and less time engaged in fruitless practices like writing documentation nobody reads or revising specs nobody adheres to.

Still, agile is not always well adapted to the startup experience. That’s not entirely surprising, because most agile methodologies have their roots in big companies. They are specifically designed to build products in situations where the problem is known but the solution is unknown. Thus, they engage in rapid communication between the engineers and an authoritative in-house customer or product owner, who can give them fast resolution on feature decisions as they come up. This is a huge improvement over ever-more-detailed specification documents. But startups routinely face the problem that they don’t even know what problem they are trying to solve.

The Lean Startup takes agile practices and evolves them for use in a startup. The net result is a focus on experimentation and extremely rapid deployment. At IMVU, my most recent startup, we built the systems that allowed us to deploy code to customers fifty times every day. When releases are measured in minutes, not months, you can build a company culture designed to avoid the biggest waste of all: building product nobody wants.

At IMVU, we shipped a product in just a few months. It was terrible. But we decided to charge for it anyway, iterating our way to a freemium business model that brings in more than a million dollars a month. In the early days, though, nobody was buying. It took months of constantly shipping features, measuring the results, and trying again before we realized what was wrong. Although we were able to get a few users to try the product for free, that wasn’t good enough. We wanted to validate the riskiest part of our business plan: that we could get people to pay real money for virtual clothes.

We were always convinced that the next feature we were about to ship would be “the big one” that would fix the product and help us make our paltry monthly revenue targets – only $300 a month in those early days. But these dreams of the instant fix never materialized. No one major release solved the problem, because the problem wasn’t a lack of features.

We eventually realized that our initial product concept, which had seemed so brilliant at the whiteboard, was fundamentally flawed. But because we took a disciplined approach to learning we were able to find out before it was too late. Because we had tried every variation of features, had measured the behavior of the people we were bringing in, and were committed to a revenue plan, we were forced to change direction. It was painful, but if we hadn’t done it, we would never have been able to chart a course that led to our eventual success.

In fact, it wasn’t the risky part of our original vision that had to change. It turns out that people really will pay good money for virtual goods. By discovering the other problems with our concept early, we were able to preserve our larger long-term vision. And, most importantly, we were then able to prove that it could work.

The Lean Startup is a vision for how startups could be built differently. Instead of focusing on hype and mega-growth, we can focus on building companies that serve customers in a fundamental way. That’s what matters.

  • It is really nice to see that these hard times are forcing everyone to get back to basics in a lot of areas.

    I feel that the concepts presented above are not new. However we all forget in the shuffle where we come from and how to be true to what we do.

    Thanks for putting this out there to help all of us hopeless entrepreneurs find our way.

  • Develop a really great service, build a very tight, strong fence around the idea, grow it out slowly and confidently.

    (I hope)

  • Evelyn Jabri

    Reading this post reminded me of the book “Small Giants: Companies That Choose to Be Great Instead of Big” (by Bo Burlingham). The founders of the small companies discussed in this book share similar views – focus on what matters to you.

  • Thanks for highlighting this upcoming talk! — this post jolted me into writing a few reactions about how these ideas can guide the startup I’m currently working on. My own post, “The Lean Startup”: Agile business development? is at but to “cut to the chase,” my favorite quote from Eric that’s worth celebrating is:

    “A line of working code is not a valid unit of progress”

  • J S

    The key to the story and success is the line figure out what the customer wants and will pay for. You can have all the passion you want but if it’s not directed at paying customers then the business is in trouble – either a brilliant initial idea or quickly prototyping your way to greatness can work. Keeping costs low so there is a slow burn rate allows you to survive longer if those iterative circles keep going.

  • bowerbird

    eric said:
    > We wanted to validate the riskiest part of our business plan:
    > that we could get people to pay real money for virtual clothes.

    i must admit that i would not have thought that this
    is what tim meant by “working on stuff that matters”.

    but, as evidenced by his invitation to have you speak
    at his conference, it’s obvious i must have been wrong.


  • What I find most important about Customer Development is the perspective that you know nothing. Which is the same line made famous by William Goldman, referring to how Hollywood has no idea how well a film will do – “Nobody knows anything”

    Looking forward to your presentation at Web 2.0. Your blog already got me reading “The Four Steps to the Epiphany”.

  • I love this post. It’s good to know that we aren’t the only ones out there committed to having a business model. And to staying lean enough to be able to make a freemium approach work.

    The term “Lean Startup” is a great one and it’s good to have a name for what we’ve been doing.

    Do you have any specifics on how to continue to attract new customers during that initial start up period. You know, just enough of them to pay for those server costs. :-)


  • I’ve been reading your stuff too. Very enlightening. Thanks for sharing. I’m with you on the need to continue to measure and keep things disciplined.

    We are trying to do ‘stuff that matters’ at If you are interested in disrupting education, then sign up for our beta.

  • BOB

    These comments helped us decide to follow our hearts and start


    Let me respectfully remind you
    Life and death are of supreme importance.
    Time swiftly passes by and opportunity is lost.
    Each of us should strive to awaken……awaken.
    Take heed
    Do not squander your life.


    as the old Californios said,
    “…la Tomba se avre para el Jinete descuidado.”
    (the grave yawns wide for the inattentive horseman)

  • @Evelyn already mentioned the book Small Giants. We’re putting on a small conference inspired in large part by that very book. Love for you to check it out. I think the genetics are in sync with this post.

  • D’oh… the URL for the conference is

  • @bowerbird –

    No, selling virtual clothes is not what I had in mind as “stuff that matters,” but learning (and teaching) how to build startups that are rapid learning engines most certainly fits that bill. If IMVU is where Eric learned what he knows, it *became* something that matters.

    I think it was A. R. Orage (maybe quoting Gurdjieff) who said, “If you can serve a cup of tea perfectly, you can do anything.”

    I suspect that Eric can serve a cup of tea perfectly.

  • Great post.

    I think it’s important to have a strong belief in a general direction, but not get too bogged down in the specifics of how you achieve it.

    When we launched we thought about launching a sister site called oriented around airfare information. But being agile and lean, we decided instead to just quickly integrate airfare into, and now its the most clicked on item on the site. Turns out people really wanted to compare the cost of flying vs. driving, and nobody offered that in the marketplace, so in this case operating on a lean model forced us to be creative and led to innovation.

  • In my experience as mentor and coach to entrepreneurs, I find that money is really their main motivator. It is the environment (analysts, VCs, etc) that motivate them to talk in terms of money. Entrepreneurs tend to be dancing to a different tune and tend to live for the idea.

  • bowerbird

    tim said:
    > I suspect that Eric can serve a cup of tea perfectly.

    then, if what you say is true, he can do _anything_.

    and in that case, i’d suggest he do something that _matters_,
    instead of getting people to pay real money for virtual clothes.

    i hope we weren’t expected to simply ignore the contradiction…

    and it’s not that i haven’t squandered my talents before. i have.
    and that experience is what spurred me to offer the advice i did.


  • “No, selling virtual clothes is not what I had in mind as “stuff that matters,””

    I have to differ on that. It seems silly but what the virtual economy is going to do is satiate the need for growth without a direct tie to physical goods. Think about it. Capitalism needs to grow at rapid rates. The only way we can keep up without exhausting all resources is creating an extension virtual economy to allow compounding exponential growth that is needed.

    Virtual goods will be one area, along with tv online and other services like on demand gaming, that allow many smart people to build markets and jobs for people to work in that entertains and makes experiences fun and worth money. Is it as cool getting a virtual Ferrari as getting a real one? Probably not but it is still pretty cool depending on how a game or system used that.

    Building virtual economies is “stuff that matters” if you abstract from it a bit.

  • The practice of rapid iteration in the marketplace (rather than in the lab) with a focus on features (rather than design) has given us this unholy plate of spaghetti that is the Web.

    Granted, the web was never designed to do what it currently does, but there is no doubt in my mind that with a little bit of foresight Berners-Lee’s document sharing system could have evolved into something much more powerful and stable.

    I would argue that this race to get products and features out of the door as soon and as often as possible has come at the expense of even the most basic software engineering principles.

    If the current web were an airplane, I wouldn’t fly on it. If it were a nuclear plant, I’d relocate to the moon.

  • @David S.: If the web evolved the way you are proposing, we’d still be using lynx as a browser and amazing our grandmothers that we could send letters without a stamp. No thanks.

    @Eric: Great article – one that I hope inspires a few fence-sitters like myself to jump onto what’s important.

  • bob

    @David Semeria: It’s very easy to talk about foresight after the event.

    TBL had an itch which he scratched – how was he supposed to foresee the modern web?

  • @Todd & @Bob – none of what I said should be interpreted as criticism of TBL. He set the ball in motion, and for that he deserves our respect and admiration.

    I’m just pointing out that this breathless, “let’s get our half-assed product out of the door and hope it gains traction” approach does not necessarily lead to the best technology being adopted.

    Everybody knows Betamax was miles better than VHS, but the latter won because it had more traction.

    That’s all I was saying.

  • bowerbird

    > Capitalism needs to grow at rapid rates.
    > The only way we can keep up without exhausting all
    > resources is creating an extension virtual economy to
    > allow compounding exponential growth that is needed.

    great idea, let’s have people pay real money for virtual _houses_!

    yes, that’ll give us the “compounded exponential growth” that
    we “need” in order to get over this sticky financial crisis mess
    bankers created where we paid virtual money for real houses.

    > Building virtual economies is “stuff that matters”
    > if you abstract from it a bit.

    i’m not sure i want to abstract in the direction you’re proposing.

    after all, “virtual clothes” is what the emperor was wearing, not?


  • I would like to add that it is important to put together supporting cast that “simply gets it”, especially on the software development team. It is important that your startup is supported by a team of developers who not only understand, but appreciate that “…startups routinely face the problem that they don’t even know what problem they are trying to solve”. It is not enough to implement a particular or hybrid development methodology. Your chosen methodology must be implemented by developers who are entrepreneurial, enterprising and who are at their creative best in the face of challenges and disappointments of the startup process. In a startup, every minute is too important to spend debating why the requirements changed on a dime or why we have to rewrite some code, you need a supporting cast that just “simply gets it”.