The Upstart

By Ted Greenwald | 05.18.12 5:18 PM

The Upstart

Scott Cook is conducting an experiment. “It’s the corporate-counseling version of speed dating,” says the spectacled cofounder of Intuit, the finance software giant. He’s gathered his troops in a brightly lit conference room, where members of four Intuit departments are seated in front of 300 colleagues—plus 1,500 more watching via webcast—to hash out some business predicaments. Each team will take five minutes to present its problem. Then special guest Eric Ries will come up with a solution.

Arun Muthukumaran, a group manager of Intuit Payment Solutions, kicks off the proceedings. He describes a feature that could dramatically increase the number of small businesses that sign up for the company’s payment services. But implementation would burn up 20 employees’ time for a month. What if customers don’t bite?

Ries, dressed casually in a blazer, pastel shirt, and black denim, suggests a test: Rather than building the service and trying it out on customers, create a sign-up page that merely promises to deliver this groundbreaking capability. Then present it to some prospective clients. Compare their enrollment rate with that of a control group shown the usual sign-up page. The results will give the team the confidence either to proceed or toss the idea into the circular file. No one would actually get the new feature yet, of course, because it hasn’t been built.

“I guess we could piss off a few customers instead of thousands,” Muthukumaran says. Laughter ripples through the crowd.

Ries glances at his watch. “It’s 4:18 pm on Monday,” he says with a puckish grin. “On Wednesday at 4:18 pm, I expect an email telling me how it went.” The team members exchange glances that are equal parts bemusement and worry: They make software, not concepts. They build code through painstaking cycles of design, programming, and testing. Customers depend on their products and trust their brand. And this guy expects them to offer a feature that doesn’t even exist? Nevertheless, the rest of Intuit’s employees are exhilarated. The room breaks into fervent applause.

It’s something Ries is getting used to. At age 33, he is Silicon Valley’s latest guru. In the four years since he first posted his theories about running startups on an anonymous blog, his campaign to replace the typical product development approach—build it and they will come—with a system based on experimentation has become a juggernaut. Ries’ book The Lean Startup, published last summer, has sold 90,000 copies in the US. His blog, Startup Lessons Learned, has 75,000 subscribers, and his annual conference attracts 400 entrepreneurs, each paying more than $500. Harvard Business School has incorporated his ideas into its entrepreneurship curriculum, and an army of followers are propagating his principles through their own books, events, and apps. Whiz kids looking for investors pepper their PowerPoint decks with Lean Startup lingo, which has become so pervasive that TechCrunch announced a ban on Ries’ term pivot. Tech darlings like Dropbox, Groupon, and Zappos serve as Lean Startup poster children, and now the philosophy is reaching established companies, including GE and, this afternoon, Intuit.

Back in the presentation hall, Ries walks his audience through the tenets of his philosophy. The core motivation is simple, and a single slide sums it up: “Stop wasting people’s time.” Entrepreneurs and their managers, minions, advisers, and investors routinely pour their lives into products nobody wants. The business landscape is littered with the wreckage of nascent companies built at monumental effort and expense that imploded on contact with the market. (Paging Webvan! 3DO! Iridium!) Unlike an established company, a startup (or a new division within an established company) doesn’t know who its customers are or what products they need. Its prime directive is to discover a sustainable business model before running out of funding.

The key to this discovery, Ries proposes, is the scientific method: the business equivalent of clinical trials. Assumptions must be tested rigorously, Ries says—and here he rolls out one of those increasingly ubiquitous Lean Startup phrases—on a minimum viable product, or MVP. This is a simplified offering that reveals how real customers, not cloistered focus groups, respond. It may be a functional product or, like the Intuit team’s sign-up page, a come-on designed to elicit a reaction. Once tallied, customer responses produce actionable metrics, as opposed to popular vanity metrics, which create the illusion of success but yield little useful information about what customers want. By repeatedly cycling or iterating through a build-measure-learn loop—a method Ries calls validated learning—the Lean Startup develops a verified perspective that enables it to identify and fine-tune the mechanism that will keep the company growing, aka its engine of growth. Or, failing that, it can pivot to a new strategy. This, Ries insists, is the quickest, most efficient route to product/market fit (a phrase adopted from Silicon Valley kingpin Marc Andreessen), defined as the moment when a product achieves resonance with customers.

Never mind that this approach is a mashup of ideas culled from programming, marketing, manufacturing, and business strategy, leavened with hard-won insights that have circulated among Silicon Valley veterans for years. Ries makes no effort to hide his sources, and his presentation preempts his critics’ complaints. “Lean,” he explains, does not mean cheap; it means eliminating waste by testing ideas first. And it doesn’t mean small, but rather that companies shouldn’t ramp up personnel and facilities until they’ve validated their business model. His philosophy is not just for Internet and app companies—that’s just where it started. Reacting to customer behavior is not incompatible with creating breakthrough products like the iPhone, Ries says, which in the popular imagination sprang fully formed from the mind of Steve Jobs.

Right or wrong, the Lean Startup has a kind of inexorable logic, and Ries’ recommendations come as a bracing slap in the face to would-be tech moguls: Test your ideas before you bet the bank on them. Don’t listen to what focus groups say; watch what your customers do. Start with a modest offering and build on the aspects of it that prove valuable. Expect to get it wrong, and stay flexible (and solvent) enough to try again and again until you get it right.

It’s a message that rings true to grizzled startup vets who got burned in the Great Bubble and to young filmgoers who left The Social Network with visions of young Zuckerberg dancing in their heads. It resonates with Web entrepreneurs blessed with worldwide reach and open source code. It’s the perfect philosophy for an era of limited resources, when the noun optimism is necessarily preceded by the adjective cautious. In a highly connected, recession- ravaged economy, Eric Ries has created a story so persuasive that the Lean Startup may already be too big to fail.

Ries was fresh out of Yale with a BS in computer science in 2001 when Will Harvey, cofounder of a startup called, plucked his résumé from a tall stack. Ries was an impressive candidate. He’d published a book (Black Art of Java Game Programming) while still in high school and started a company that had actually received funding ( was a virtual world along the lines of Second Life, where members interacted through 3-D avatars. The young engineer quickly demonstrated an unusual quality: “He wanted to help everyone,” Harvey recalls. “He could understand someone else’s problem and solve it in a miraculously short period of time.” made a high-profile debut in 2003. But the public yawned, and poof—$50 million and years of dogged effort vanished. The following year, Harvey launched his next venture, a site called IMVU, which aimed to populate chat networks with 3-D avatars. He brought Ries aboard as CTO. He also roped in investor Steve Blank. A Silicon Valley legend, Blank had played a founding role in eight tech companies before withdrawing to the less stressful life of a university professor. He agreed to fund Harvey’s new company on one condition: that its top executives audit his entrepreneurship course at UC Berkeley.

For Ries, the class was a revelation. In the conventional view, startups are smaller versions of established companies, Blank explained. But in fact they’re not. Established companies know what to make and who will buy it. Startups don’t. When neither the product nor the customers are known, understanding the audience demands as much rigor as creating the product: It requires something Blank called customer development. Focus groups won’t do. The only solution is to put real features in front of customers, track their response, rinse, and repeat until the product gains traction.

Ries was already a convert to the school of agile software development. At the time he joined IMVU, most software projects used the waterfall method—programs were fully specified, built, tested, and deployed in one long, sequential march. The agile method, on the other hand, emphasizes rapid development and responsiveness to user needs. Designs evolve alongside user feedback, and products metamorphose through repeated cycles of design, development, and deployment. Ries realized that Blank’s process could augment agile software development by enabling IMVU to cycle rapidly through customer development in parallel with product development.

Ries spent the next four years fusing techniques from Blank’s self-published handbook The Four Steps to the Epiphany with ideas gleaned from Kent Beck’s Extreme Programming Explained, the lean manufacturing study The Machine That Changed the World, and Geoffrey A. Moore’s marketing classic, Crossing the Chasm. Fueled by these insights, IMVU tested unorthodox ways to learn who its customers were and what they wanted.

At first the company raced to develop a minimum viable product only to find that no one wanted to download it. Ries realized that he could have learned the same thing from a simple description of the product and a Download Now! button. Using Google’s AdWords program, which allows anyone to buy Internet traffic at a low price, he budgeted $5 a day to attract 100 visitors. Ries shunted them through alternate designs and compared the differences in their behavior (a practice called split or A/B testing). He tracked the shifting percentages as visitors registered, logged in, used the product, and upgraded to the paid version (known as cohort analysis). And he practiced something he called continuous deployment, updating code as frequently as 50 times a day. After many false starts, IMVU prospered.

A new CEO took the helm in 2008, and Ries decided to move on. Word of his innovative derring-do had made its way to the VC firm Kleiner Perkins, which gave him an office and a title: venture adviser. It was a place to hang his hat while he figured out what he wanted to do. But he never felt comfortable there. “Fund-raising strategy should be an outcome of business strategy, not the other way around,” he says. Six months later he went indie as a business consultant.

One of his first clients was IMVU investor and superangel Mike Maples. “At lunch in summer 2008,” Maples recalls, “Eric said, ‘I’m working on this new set of ideas called the Lean Startup. What do you think?’ He showed us a bunch of raw slides. We said, ‘This is cool! This framework could have legs.’ He said, ‘If this works out, I won’t just consult. Maybe there’s a book to be written.’”

Other clients weren’t so receptive. “I would tell stories of what we did at IMVU, and they would look at me like I was a lunatic,” Ries says. Tired of justifying his advice, he decided to start a blog. “Then I could say, ‘Read this essay. If you think I’m crazy, don’t hire me.’”

Initial posts on the blog, Startup Lessons Learned, were anonymous. Then, to the author’s surprise, a trickle of readers started showing up, mostly hungry young entrepreneurs who came across the site while searching Google for guidance. True to Blank’s customer development credo, Ries asked them what they wanted to get out of his writing: “Lo, my 5 subscribers,” he addressed them in September 2008. The response was unequivocal: They wanted to know who he was and why they should take him seriously. It was time to pivot from private adviser to public pundit. So in October, after two months of anonymity, he peeled off the mask and revealed his identity.

Then the economy tanked. Sequoia Capital issued a presentation to portfolio companies titled “R.I.P. Good Times,” advising CEOs to “spend every dollar as if it were your last.” There couldn’t have been a better advertisement for a philosophy called the Lean Startup, and a rising tide of distressed entrepreneurs found their way to Ries’ blog.

Suddenly Ries found himself in high demand. Executives requested meetings. Publisher Tim O’Reilly invited him to speak at the Web 2.0 Expo. Harvard Business School offered Ries a spot as entrepreneur-in-residence. Lean Startup groups sprang up in cities across the globe. In his November reader survey, Ries called out: “Lo, my 1,032 subscribers.” By January 2010, with his posts ricocheting among startup blogs and business sites, he had 18,891 followers. And, by September, 57,692.

The Lean Startup had escaped into the wild, where it was mutating into events, courses, software tools, and blogs. But Ries was more interested in propagating his ideas than in controlling them. He trademarked the name as a stick with which to beat scam artists, but he resolved not to charge anyone who used it honestly. He publicized meetup groups on his blog and went out of his way to attend their initial gatherings.

One hectic year after his first blog post, he was ready to push the Lean Startup out of its beginnings in Silicon Valley and into the mainstream. Published in September 2011, The Lean Startup debuted at number two on the New York Times advice best-seller list.

On an unseasonably warm January day, Dogpatch Labs swarms like an anthill. Fifty would-be entrepreneurs have come to Palo Alto from as far away as Finland for an event called Lean Startup Machine. This is not an Eric Ries production but a hackathon of sorts organized by 24-year-old impresario Trevor Owens. Participants form teams, rally around a product idea, and then test it. Next they “pivot or persevere” (a Ries adage) based on what they learn, aiming to complete as many revolutions around the build-measure-learn sequence as possible before Sunday night. The team that demonstrates the most validated learning wins workspace, legal services, and consulting from LSM mentors.

Ries was aghast when Owens emailed in mid-2010 asking him to appear at the first LSM. “I said it’s fundamentally wrong to judge a startup in 48 hours,” he recalls. But Owens persuaded him to participate, and Ries was amazed by what he saw: teams pivoting several times in a weekend, learning in three days what might otherwise take weeks or months.

For the January event, Ries Skypes in from the UK—where he’s speaking at the London School of Economics and launching the Dublin meetup group—for a 30-minute pep talk. While Ries spent his childhood programming multiuser dungeons, today’s entrepreneurs, coming of age in the era of Angry Birds, have at their fingertips everything they need to create the next Twitter or Zynga. They can cadge open source software from SourceForge, rent processing power from Amazon Web Services, buy traffic from Google, and distribute through the App Store. The Lean Startup gives them a guide to spinning these resources into gold.

Indeed, LSM is a hive of activity. One group is posting ads on Craigslist to test demand for a scam-free classified ad market. The team that calls itself CrowdJewel spams the Etsy ecommerce network to promote a jewelry-design contest. Participants are shooting demo videos, mocking up screens for mobile apps, bolting together open source code, and researching cloud-computing resources.

They’re also cranking out online ad campaigns for products that don’t exist and web pages featuring bogus Buy buttons. The absence of real products is not a problem, Ries tells them. Don’t worry about negative fallout from an undercooked MVP—a follow-up apology will mollify mainstream customers and actually make early adopters more enthusiastic by letting them in on the product’s genesis.

This degradation of the minimum viable product into blatant deception is one of several axioms that rankle Ries’ admirers and critics alike. For instance, the Intuit Payment Solutions employees working on their promising new feature ultimately decided they felt uncomfortable misleading customers. Instead, they invited a select group via email to participate in a private beta.

Other aspects of the Lean Startup philosophy have met with more-vocal opposition. While debating with Ries at a startup conference, David Heinemeier Hansson, creator of the Ruby on Rails programming tool, professed the credo “We only build software for us.” Ben Horowitz, cofounder of the VC firm Andreessen Horowitz, picked up the baton with a blog post titled “The Case for the Fat Startup,” arguing that smart entrepreneurs don’t conserve capital; they raise enough cash to crush competitors. Defenders say that Horowitz was misinterpreting Ries’ use of the word lean, but the feeling lingers that the Lean Startup philosophy trades vision for plodding validation. “The whole philosophy is barren and derivative,” says Keith Rabois, the superangel and PayPal exec who now serves as COO of Square, a service that allows anyone to accept credit cards. It’s incapable, he says, of inspiring companies like Apple or Tesla, which have made the words startup and innovation synonymous.

Ries doesn’t see any contradiction between visionary leadership and sussing out the market through experimentation. “The Apple I was a classic MVP,” he says with a shrug. Steve Jobs cycled through designs until he hit on one that clicked. A method wasn’t available to turn his experimentation into a structured process; he had to rely on intuition and browbeat his organization into following it. “The next generation will have the tools to prove they’re on the right track,” Ries says. “They won’t have to deal with that kind of bullshit.”

Not that the LSM participants have come up with an Apple this weekend. As Sunday draws to a close, Owens proudly announces the winner: CrowdJewel. Headed by Courtney McColgan, an entrepreneur-in-residence at Morgenthaler Ventures, the team staged a 15-hour jewelry competition that attracted 75 designs from 14 countries and garnered 600 votes. After announcing a victor in its contest, it confirmed one sale by tracking traffic from its landing page to the designer’s site. The only negative indicator: Googling the term “CrowdJewel” leads to a discussion about whether the company is a scam.

For all the Lean Startup’s emphasis on the scientific method, no one knows whether it really works. An argument can be made that the companies cited as Lean Startup success stories aren’t strictly evidence of Ries’ wisdom. Dropbox famously debuted with a video depicting a beta of its yet-to-be-released product—spurring 70,000 users to sign up for the service within 24 hours. But the company didn’t base its entire strategy on Lean Startup principles.

Ries himself is confronting the question head-on. Determined either to validate the efficacy of his theory or iterate his way to something more effective, he has teamed up with Nathan Furr, an assistant professor of entrepreneurship at Brigham Young University, to subject the Lean Startup to rigorous analysis. Furr is hammering out before-and-after case studies and examining correlations between learning, shifts in strategy, and performance.

But Ries’ apostles aren’t waiting for data. Top business schools are cranking out candidates for the next generation of Lean Startup case studies. Tom Eisenmann is leading 100 Harvard MBA students through Lean Startup principles in his course Launching Technology Ventures. Some 80,000 people signed up to take Steve Blank’s Lean LaunchPad, a 10-week online boot camp meant to foster functioning businesses. The National Science Foundation is funding him to do the same with a handpicked corps of scientists and engineers.

And the Lean Startup is advancing into the public sector. Tim O’Reilly, as board member emeritus of the nonprofit geek squad, Code for America, is helping push Ries’ principles into municipal governments. “We’re teaching cities to use the Lean Startup approach instead of a yearlong procurement process,” he says.

If Ries can keep the movement yoked to the scientific method, it might just isolate principles that steer companies toward real, lasting, massive success. Until that happens, he will remain a reluctant guru. “I want the message to be heard,” he says. “The movement needs a face, and if it has to be me, I can accept that. But I’d be happier if it could be the faces of the obscure entrepreneurs doing the real work of figuring out whether these ideas are true.”

Which is not to say he isn’t enjoying the ministry. “I will do this as long as it’s having an impact and it remains intellectually interesting,” Ries says. “So far, it’s off the charts on both fronts. If it turns out to be a fad, I’ll go back to what I was doing before.” After all, he’s comfortable with failure. It’s the only way to learn.