Home » Leadership & Strategy » Entrepreneurial CEO » Zen and the Art of Startup Maintenance

Zen and the Art of Startup Maintenance

Watching Steve Larsen climb rickety steps into the unheated ad hoc conference room on the second floor of the funky, …

Watching Steve Larsen climb rickety steps into the unheated ad hoc conference room on the second floor of the funky, old, legacy Casa Mills building in Menlo Park, California, you’d have no idea that hours earlier he was finishing an assignment for RoadRunner magazine testing BMW’s new top-of-the-line, dual sport motorcycle on an eleven-day tour over the backroads of Turkey. The man likes speed. But we’re about to talk about another of his passions: driving startups hard.

Three weeks prior, in August 2005, after spending several weeks with Ken Krugler, a world-class software developer working out of the offices of Emergence Capital in San Mateo, Larsen had agreed to take on the role of cofounder and CEO of a new company with no product, no offices, no revenue, and no employees other than himself and cofounder Krugler.  By September, Larsen and Krugler had signed a $1.2M seed-round term sheet, which, when it closed, made Krugle a real company – while Larsen was cranking the BMW cycle all over Turkey.  Cofounder Krugler signed the papers, which Larsen then read, reviewed, and approved on his laptop in Turkey – during one of the few times the motorcycle wasn’t moving.

Steve demonstrating proper turning technique oh his Suzuki V-Strom to students of Lee Park’s “Total Control Advanced Riding Clinic”

ChasingSerial Guy’

With a track record of cofounding or running as many fast-growth startups as Larsen has, it stands to reason he’d be a juicy target for venture capitalists.  Successful serial entrepreneurs have a way of getting the attention of these folks.                                     

First, some history.  In the mid-’90s, Larsen had been part of the founding team at CitySearch (later acquired by Barry Diller’s IAC).  After helping build that business, Larsen was approached by Ann Winblad, a partner at Hummer-Winblad Ventures in San Francisco.  Winblad thought Larsen would be ideal for a gig back where he grew up, in Minnesota.  And she eventually recruited him to join that small team, for whom she had provided seed funding. It was a startup that wanted to turn “collaborative filtering” technology into a business. As senior VP of marketing and business development at this startup, Net Perceptions, Larsen went on to create an identity for the firm with a series of events he conceived, called the “Personalization Summit,” which helped enable the company to IPO in 2000.

With a 26X return to investors, Brian Jacobs, a partner at then St. Paul Venture Capital (SPVC), who had led the Series A funding of Net Perceptions, soon came knocking. He quickly recruited Larsen to SPVC with the hope he would add the same magic to a number of that firm’s other portfolio companies, and Larsen ultimately did that with LowerMyBills.com and others.

But it wasn’t long before Larsen was back to starting something from scratch, becoming the third hire in a Jacobs-incubated startup called Visage Mobile in the California offices of SPVC, backing the efforts of founding CEO Matt Johnson and CFO David Fraze. 

A year into Visage Mobile and soon after its Series A funding, Larsen was tapped to take on the interim CEO role at troubled BigFix Inc. While long on technology prowess, figuring out the right value proposition, marketing direction, and the “right ballpark in which to play,” as Larsen explains, took a bit of work. “The founding CEO had a vision and passion for the technology; it was just that it wasn’t aimed in precisely the right direction.” 

A year later, Larsen had replaced the sales, marketing, and product management teams and hired a president and COO who would take over the CEO slot after the board’s attempt to find a permanent CEO had failed.

“BigFix was and is an awesome company,” said Larsen, “with technology that seriously outperforms the current leaders in the security and systems management space. All they needed was a slight readjustment to turn things around and get them headed in the right direction. Dave Robbins, who’s now CEO, will ultimately and justifiably take the credit for getting BigFix into the headlines.”

What to Do Next?

In 2005, Larsen was being courted by a small group of VCs familiar with his track record – when they could find him, that is. “Larsen’s hard to keep track of,” says Brian Jacobs, now a General Partner at Emergence Capital Partners. “Once he’s out of a deal, he’s off riding motorcycles in Brazil, New Zealand, or who knows where,” Jacobs adds and then laughs. “But, though he’s hard to find, I’ve invested in him four times now, so I’ll probably keep doing it till we get it right.” 

Jacobs explained the latest deal he wanted Larsen for: “When Ken Krugler, a fraternity brother of mine at MIT, approached me with the idea for what would become Krugle, it was incredibly compelling.  Ken is a world-class developer and, if he was right, it would be major disruptive technology in the $10-billion application development space.  But I had to tell him that we don’t invest in technology, only companies.  Ken didn’t want to create and run a company, but he was willing to look at an outside CEO. I immediately thought of Steve and got the two of them together. I told them that if they were both committed, and my partners bought in, we’d fund it.”

“Steve’s never had a failure,” Jacobs continued. “I recruited him to St. Paul Venture Capital after he left Net Perceptions, which went public in 1999, and he spent three years in countless board meetings seeing what CEOs did that worked and what didn’t.  As a result, he totally nails what a board is looking for – clear, unambiguous visibility and intelligent insight into how the company is doing, with no attempt to sell.  And he knows how to build early-stage companies.” 

Larsen elaborated on Jacobs’ comments: “VCs really don’t want to run a company – they just want to be sure that someone else is and that someone knows what they’re doing.  And they want to ensure they’re heard, that the CEO listens to their point of view and considers it carefully before making the decisions and informing them.  Boards should really only make two decisions, hiring (or firing) of the CEO, and when and for how much to sell the company – other than that, they’re advisors.” 

Road Warrior

When not doing the startup CEO thing, Larsen has a second career as a moto-journalist, riding and writing about motorcycles and motorcycle experiences.  “I write for RIDER, Motorcycle Consumer News, and RoadRunner,” said Larsen. “My articles mostly focus on motorcycle training, some product reviews, and trips around the world.  I’m probably one of only a few riders who have done advanced training with top instructors who train professional riders for both SuperBike competition as well as off-road racing and trials competition.  I’m the only civilian to take both the Arizona Highway Patrol Motor Officer training program and spend four weeks in the Phoenix Police Department’s motor officer training program.”

As much fun as he’d been having riding, in mid-2005 he began making calls to select VCs to get a sense of what was out there.  But nothing seemed seriously tempting until Brian Jacobs called.  “Brian obviously thought a lot of Ken, and that meant a lot to me.  Then, when I ran Ken through the gauntlet of my technology advisors and he came through with flying colors, and Ken’s business advisors felt good about me, we felt we had a fit, that we could work together and we should go for it.”

Later, Larsen related to me his excitement about Krugle: “Its potential is incredible, because with a simple idea – applying a powerful search capability to a company’s software code base – we unlock literally billions of dollars of past investment.  Without Krugle, that past investment is simply ‘sunk costs’ – impossible to find, understand, or leverage.  We have the potential to double the useful output of a development organization by creating a groundswell of individual developer productivity, all by applying technology that deploys instantly and has no learning curve.”  I learned about one early user of Krugle who calculated that just one use-case alone could save their firm $10M annually.

February 7, 2006
DEMO, Phoenix, AZ 

Four months after his decision to co-found Krugle, Larsen is a little different from what I remember – more nervous, a bit sharper at the edges, and he’s talking faster. “We’re not quite ready, but we need to get this thing launched,” he explains. “We need developers to see and use this – real software geeks – and we need to line up someone to provide our next round of funding, and DEMO is the best place to do this.

“We didn’t have the time or money to really publicize this thing,” he said. “So we’re going to rely on the blog and talking to the media that DEMO attracts to get our message across.”  

The company had hired Silicon Valley veteran Don Thorson as interim VP of marketing to get them through the launch of the company. “I had no idea what I was getting into,” Thorson explained. “Larsen has this idea that constraints breed creativity and innovation – that is what start-ups are all about.  At one of his previous companies, he didn’t allow any money to be spent on advertising.  Apparently it worked for them, so now he wants me to launch Krugle, but without spending any money on advertising or public relations. I’m thinking, this guy is nuts.” 

“But we created a blog and, under the tutelage of some of Larsen’s friends, including Shel Israel, Doc Searls, and Chris Locke, pretty soon we had some good content and some regular users,” Thorson explained.

He went on:  “We have pretty simple goals for DEMO.  We want to get enough press coverage to get two to three thousand software developer types to signup for our beta, and, second, attract potential Series A investors. Larsen and John Mitchell, Krugle’s software architect, do a darn nice job (http://www.demo.com/demonstrators/demo2006/63003.php) and end up winning a DEMOgod award.  I’m pretty jaded with hundreds of product and company launches, but this would be the first time I’ve seen a product reach so far with so little expense.  The reach and velocity of the messages has to be attributed to the viral nature of blogs – ‘friends telling friends’ about a product or service they might find interesting.”  (And I confess I was one of those bloggers.)

April 27, 2006: Assessing the Launch

The DEMO experience has exceeded expectations.  The company shuts down beta signups at 35,000, well past its goal of 2,500 when it realizes it won’t be able to send emails to all those people.  More importantly, two venture capital firms submit term sheets in advance of the company’s April board meeting, and the company and existing investors decide to accept Rustic Canyon Venture’s term sheet, which values the company at $7M pre-money.

But now the company needs to execute.  February through April, Larsen’s time has been largely occupied with fund-raising and he’s frustrated. “This is one of the most important tasks that I have, but one of the least productive,” he explains. “While we get the money, we need to move forward, and none of this effort on my part actually moves us ahead.”

“Now, we have to get the product launched, generate usage, and begin plotting a very clear path to revenue.  We have to fill out the team, cement the value proposition, and figure out how we’re going to make a lot of money.” 

Krugle’s offices, although in the heart of Silicon Valley, are decidedly low rent.  While they’ve moved from the second floor offices in that funky Casa Mills building (with no heat, and ceilings that collapsed during a heavy winter rainstorm) to nicer remodeled offices on the first floor, they are still sparsely-furnished and cramped. Although the way Larsen shows them off, you would think you were getting a tour of the Googleplex or Taj Mahal, instead of a small startup.

“This is our global, worldwide, executive briefing center,” he proudly exclaims as we sit down in a conference room that doubles as the company kitchen.  The rest of the offices, while no means luxurious, are neat, clean, and organized.  Key slots on the team, both technical and business, have been filled with experienced professionals, but the company has run into problems.

“We’ve not had developers beating our door down,” Larsen explained. “But we thought that might be the case.”  The company had pitched early investors on two potential strategies.  One showed individual developers embracing and rapidly spreading the word about Krugle’s free public search engine and a pay-per-click ad model generating revenue.  The second showed an enterprise application that would reside behind a company’s firewall, providing visibility for the first time ever into a company’s code assets, which could be leveraged in a number of powerful ways.  

“We showed the VCs a revenue plan that was close to 50/50,” Larsen said, “with half of the revenue coming from the ad-driven public site, and the other half from our enterprise product. Every VC said it was not going to be 50/50 but ultimately 90/10.  We agreed, but then challenged them to pick which would be the 90 and which the 10.  We seriously had people argue quite convincingly on both sides.  Our stance was that we were unsure which one was going to be the 90% and which was going to be the 10%.  But, in either event, we needed to build the public site first.” 

Big Fork in the Road

By the fall of 2006, the company’s free web product, which searches for open-source code across hundreds of repositories, had drawn tens of thousands of fans and won tech industry accolades and awards, but lacked the millions of users necessary to prove out an ad model.  It was time to change strategies, and change quickly. 

Larsen tells me that the most successful people are those who are good at Plan B, which is why he always creates a Plan B, right from the start.  “It’s not that you don’t have faith in your initial vision,” he tells me.  “It’s just that when you are doing something brand new, it is so hard to get it exactly right.  You have to be ready to change and change quickly.”  Unfortunately for Krugle, shifting to developing an enterprise product requires different people and different thinking, and competition in the Valley for engineering talent early in 2007 has begun to heat up.

But Larsen is a fiercely competitive recruiter.  He finds and interviews Dean Pfutzenreuter, a first-class software engineer and manager.  But Pfutzenreuter is currently employed and has no desire to change careers or companies.  Unfortunately for Pfutzenreuter, Larsen has decided he’s the right guy.  A week after they meet and Pfutzenreuter had explained he was not looking to change jobs, Larsen sends him an offer letter.  Two weeks later he has Bob Cagle, Krugle’s VP of Engineering and an old friend of Pfutzenreuter, take him out and buy him beers and burgers and tell him how fantastic Krugle is and more about what he would be doing.  A week later Larsen drives by Pfutzenreuter’s office and leaves a business card on the windshield of his car and follows it a week later by emailing him another job offer, this time for a bit more money and a $10K signing bonus.  The message comes back from Pfutzenreuter through Bob Cagle:  “Thanks but no thanks.  I’m not leaving where I am.”

“I figured that was the end of the story,” Bob Cagle tells me. “But not for Steve.”  Two weeks later, Larsen crafts another offer letter for Pfutzenreuter, this time for even more money and a $25K signing bonus, and sends it off on a Friday.  “I didn’t expect anything and didn’t bother to check with Dean over the weekend,” he said.  “I get a call at 6:30 on Monday morning from Dean telling me to check the fax machine.  Sure enough, there’s his signed offer letter.  He’d given his two-week notice.” 

On his first day on the job, Dean stood in the door of Larsen’s office, just grinning.  They looked at each for a few seconds, then Larsen said,  “So, you figured out we were never going to quit, did you?”  Dean thought for a second, then said, “Yeah, I guess more than anything else, that was probably it.”   Larsen has a way of getting what he wants.


April 2007, Garden Court Hotel, Palo Alto, CA:

Meeting of the Minds

Five companies gather in an upstairs conference room for a day-long briefing.  They’ve agreed to test Krugle’s software for discovering software code inside a company’s firewall.  The stakes are high, as these companies are substantial and represent a cross-section of industries that could be Krugle’s future customers:  communications leader Motorola, aerospace giant Raytheon, financial services trading company Chicago Mercantile Exchange (CME), huge defense contractor SAIC, and a large Indian outsourcer, Persistent Systems.

After hundreds of whiteboard drawings and months of coding, the first Krugle enterprise appliances are scheduled to ship to these users in the next few weeks, and this session is to get their buy-in and input on the initial set of features and their feedback on what needs to be in the product longer-term.

“Krugle Enterprise” is like Krugle’s public site but pointed at internal code bases with the same ability to search on projects and source files.  Developers inside a given company are able to browse the code of any project, add notes, and share just about everything with coworkers.  This view into the world of private and public source has never been possible before and has the potential to save development organizations months of work and millions of dollars.  Krugle’s mantra is “If you can’t find it, you can’t fix it, refactor it, learn from it, or avoid breaking it.”

The mood is upbeat, as a month earlier the Yahoo! Developer Network (YDN) had adopted Krugle’s search product for crawling and indexing YDN code and technical pages.  Having a major search company acknowledge the value of code specific search, and that it’s something impossible to do using normal text search and page-rank algorithm, is a major breakthrough and validates many of Krugle’s efforts, including the time and effort the company has made to patent the key aspects of what they are doing.  While not announced with the YDN deal, the company has also been working with IBM, who will soon make similar announcements, as will SourceForge.net and Amazon’s developer network. 

October 2007: The Inevitable Speed Bumps

Larsen calls to share some news: Dale Fuller, the former CEO of Borland and McAfee, has joined Krugle’s board.  They’ve hired a VP of sales that understands this space and he’s building out the sales team.  And they’ve made a solid hire on the product marketing side as well.  The product is doing well.  

“We’ve had good news and bad news,” Larsen explained.  “The good news is that there’s been very strong interest for the product within larger companies with larger software development organizations.  The bad news is that they have more specialized requirements that go beyond our initial launch version of the product, which was targeted for smaller companies or departments of larger companies.  So, the version we have of the product needs greater performance and scalability.  And we need to make it far easier for prospects to evaluate the product, without us having to send them a box.  So, we’re creating a VMware version of the product that should address this shortcoming and speed the sales cycle.”

Larsen and CFO Bill Daniher go on to explain that while the deal size will be much larger, the sales cycle will be longer, and they’re about five months from running out of money.  “The key to getting funded and, more importantly, to getting the kind of valuation we want in the next round, will be the number and quality of our customers.”  Seeing the runway shrinking and cash getting short, Larsen and Daniher are working on a far less dramatic, less aggressive plan.  This plan has been heavily influenced by the newly-hired sales VP, Mike Rich.  “It’s amazing how much insight you get when you have someone who’s spent their career selling products into this space,” Daniher says.

Daniher has created a plan with a sales forecast based on a new, lower sales floor, a higher goal, and a middle forecast positioned as the most likely case. The plan they’re presenting at the November 28 board meeting calls for spending in 2008 to be tied to the lowest bookings number.  The board likes the direction but suggests the company go back to the drawing board and find even more to cut.

The board has asked for a phone update late in December to get the status on sales, enhancements to the product, and the new financial plan with the more aggressive expense cuts.  The board call went well and the new, stricter expense plan, which includes eliminating two VP positions and sharply reduced technical consulting expenditures, is accepted.  It should extend the company’s runway to provide more time for some of the larger sales to come in. 

Although no revenue was promised to the board in 2007, the company gets its first check from a real customer who’s adopted the company’s application on the last day of the year.  Will it be enough and will others follow suit and arrive in time? 

The long, hard ride continues.

 Postscript: Doing What He Loves

Larsen now confides to me that he had hired a CEO coach early in 2007. He tells me about a dinner meeting he had with a business associate and her husband, a former musician and retired professional poker player, who now makes his living as a poker coach.  “When I suggested that perhaps her husband could give me some hints, she politely told me that his clients need to show poker income of at least $500K for three years before he takes them on.  I was shocked. Hell, if you make half a million dollars playing poker, why would you need a coach?”  She went on to explain that Tiger Woods has a golf coach, and Michael Jordon – at the top of his career – had a basketball coach, as do most professionals attempting to reach the top of their game.  “As I drove home, I had this epiphany,” said Larsen. “I’d spent thousands of dollars and weeks and months of time seeking out and learning from the people who could best improve my motorcycle riding. And how much money do I make riding or writing about motorcycles?  Answer: not much.  And how much money do I make as an early-stage startup CEO?  Quite a bit!  And how much time and money do I spend getting coaching to make sure I’m always becoming a better CEO? Nothing.  Nada!  I saw this as a great opportunity.”

Enter the Coach: Larsen had soon hired Linda Tirado, Ph.D., of The Strayer Consulting Group, (www.strayerconsulting.com) of Los Gatos, California.  Tirado has coached more than 300 startups and CEOs since 1986, mainly Silicon Valley technology firms. Early in her career, she primarily coached CEOs who were troubled in their new role, often being hired by the firm’s venture capitalists.  Now, said Tirado, CEO coaching is more accepted, and she works with solid leaders looking to become even better.  “It’s a special breed of person that does well with early-stage startups,” she said.  “They have to be very smart, fast-moving, able to deal with a lack of structure, ambiguity, and be risk-takers. These are big-idea people who really enjoy the process of bringing something nebulous into a better shape and form.  They have to be able to capture lightning in a bottle, to capture that big idea and present it to others in a way that attracts. They have to get people excited, and create a business model that gets funding.”

Startup CEOs have to have a vision of the technology, but also need to be aware of the market, which is hard to do, said Tirado.  “They’re more often involved in the visionary side, and not the business side.  That’s why Steve is particularly valuable in the marketplace – and he’s been paired up with Ken Krugler, who realized he needed a business partner.”

Typically, a CEO that thrives in the early stages of the company doesn’t necessarily do well as things grow up, according to Tirado.  “They can’t wait around, they just have to move things along – they’re not interested in a lot of structure and process.  As the company grows and gets more complex, the leader has to learn to leverage the product by scaling it. As things get bigger and more elaborate, you need more structure. There are some people that are particularly well-suited to the rough ride of those early stages, who are really motivated and can attract talent. There’s a sort of DNA that thrives in that environment, and Steve is a good example of that. The excitement for him was bringing those people together.”

Lessons Learned: I asked Tirado what other CEOs can learn from Larsen. “Some people think they should know all the answers, or that there’s some problem in acknowledging that they can’t do everything. Steve’s a very humble guy, and he always knows there’s a way to do it better – always open to improving the company and himself. He’s very confident, and very good at attracting talented people. One thing he wasn’t as aware of was how to make all those talented people work together. He worked individually, hub and spokes, with each person. But, at the growth stage of the company, it’s important to get other people to leverage each other’s talents, to ensure that the whole becomes greater than the sums of the parts.

About graeme thickins