It's hard to distill fourteen years into twenty-five minutes, so rather than try, I'm going to tell the story of Fieldglass in five acts and pull out a few lessons along the way:
- Founding, 1999 to 2000
- First of many, 2000 to 2001
- Becoming a contender, 2001 to 2005
- The MDP waypoint, 2005 to 2010
- The road to SAP, 2010 to 2014
But first, what does Fieldglass actually do? It's in an industry called VMS, for vendor management system: a software-as-a-service platform that large companies use to manage everyone who works for them but isn't an employee, meaning contractors, consultants, temporary staff and the services firms that supply them. Its customers range from mid-size companies to the largest banks. Today something like $30 billion runs through the network across roughly 300 large customers.
Act one: Founding (1999 to 2000)
Late 1999. The dot-com bubble is at its frothiest, and its hottest corner is business-to-business. Online exchanges from Ariba and Commerce One promise to rewire how companies buy everything, and "B2B" is the prefix every pitch wants.
Let me start at the beginning. In late 1999 the dot-com bubble was at its frothiest. Everyone had a dot-com business plan, and money was flying around. One of my colleagues at Giant Step introduced me to Jai Shekhawat.
Jai wanted to build software to manage the suppliers of contractors, starting with staffing companies. That ran completely counter to the thinking of the time, which was all about disintermediation: cutting the middlemen out. But once he explained it, I could see that he was deep in the domain and understood how the business really worked, from the inside. He also had one other person nearly on board, Chris, a seasoned staffing-industry sales executive; $500,000 in angel funding lined up; and, of course, a handful of big-name prospects who were "just about to close."
So I joined, to run technology, while Jai focused on strategy and fundraising and Chris on sales. In keeping with the times, the company was called b2bpeople. Hold that thought.
I showed up on my first day, January 12, 2000, to an office of maybe five rooms on Diehl Road. It was completely dark except for one office, where Jai was sitting. No furniture, no other people. That's when it hit me how utterly crazy what we were trying to do was. We were going to take on the massive staffing industry. We were going to convince big companies, full of tough procurement people, to give us real money for software we still had to build.
Luckily, I had the arrogance and ignorance of youth on my side. I shrugged and got to work.
It was obvious right away that the first few hires would be critical, and we learned how much those early hires shape everything that follows. We built a great core technology team. My first few key hires stayed the entire way, and my very first hire is now the CTO.
Act two: First of many (2000 to 2001)
2000 to 2001. The NASDAQ peaks in March 2000 and begins a fall that will last two and a half years. The B2B exchanges that defined the hype start to fold, and by 2001 the country is in recession, deepened by September 11.
The second act covers our first eighteen months or so, when we hit the critical early milestones: a product in production, a live customer, and institutional funding.
I'll start with the product. From the beginning we built Fieldglass as software-as-a-service. For all practical purposes, think of it as a cloud platform. But that wasn't some brilliant insight into the future of technology. It came from two things we saw around us.
Application service providers, or ASPs, were popular at the time, and they were crashing and burning hard. We thought that was because they ran other companies' software and could never reach economies of scale. But we liked their turnkey delivery and their control of the entire stack, which is essentially the cloud.
Portals were popular too, and they would later crash and burn as well. But we liked the idea of serving many buyers and many suppliers at once, which is the basis of multi-tenant software and of marketplaces.
We took a very pragmatic approach to development that put getting it done ahead of devotion to any methodology or architectural elegance. And we learned early that where you should sit on the curve between quality and speed depends on what a critical bug costs. Ours isn't a life-and-death application. When something goes wrong there's usually a manual workaround, so we could err on the side of speed.
Our first customer was Verizon Wireless. The amazing thing is that they signed up in 2001 knowing the product was still under development and not ready for prime time. They have the dubious distinction of being the only company ever to use our first live version, 2.5. But they understood the problem and believed in our solution. Most of all, they believed in the team and wanted the chance to influence us. That is exactly the kind of visionary client every business-to-business startup needs.
The obvious question is whether that was luck. Yes and no. We were certainly lucky to find a client like that. But we were also frugal with our cash and disciplined about where we spent it, and that culture, which became one of the hallmarks of Fieldglass, let us survive long enough, time after time, to get lucky.
Speaking of cash, our first two institutional rounds both closed in 2000. Jai, with the rest of us as his supporting cast, was fundraising from the very beginning. Since most of you weren't there to live through it, some context:
- When we started in January 2000, the NASDAQ was at 4,300.
- On March 10 it peaked at 5,048, and plenty of people believed 10,000 was coming, even inevitable.
- By March 20 it had dropped 10 percent.
- By April 14 it was at 3,321, down 30 percent from the peak.
- It kept falling until it bottomed out at 1,114 on October 9, 2002. People joked that their 401(k)s had become 101(k)s.
Jai flew around the country to meet more than ninety venture capitalists, and at some point he tore his Achilles tendon playing squash, so he did much of it on crutches. Fewer than a handful ever got serious. Still, in the first quarter of 2000 we raised our first institutional round, a small $1.5 million Series A. Ironically, after all that travel, half of it came from Prism Opportunity Fund, a small Chicago fund that my friend Steve Vivian had joined a few months earlier, and the rest from angels. We spent it building the product. In late 2000 we raised an $8 million Series B from BlueStream Ventures, a young fund out of Minneapolis.
Now, about that name. In early 2000, b2bpeople sounded exactly right. By that fall it sounded like the bubble that had just burst, and I had been grumbling about it from the start. In October, Jai put the question to the team: "Do we like our name?" Investors were telling him it was "a great name for a product, not for the company," that "b2b" was out of favor, and that "people" made us sound like a staffing firm. My entire reply was: "Need I express my views? Looks like time to get everyone together around a bar and come up with names." Two days later, Jai forwarded the news that Andersen Consulting was renaming itself Accenture. We were not the only ones.
What followed was six weeks of naming by committee. Justin, on our team, proposed Roundhouse and Stationhouse, after the turntables that switch railroad engines, and made an offer on stationhouse.com to the small leather shop in Connecticut that owned it. Jai floated Glass Engine. In December, Justin declared it decision time, and I agreed. Then he raided his CD collection: Mothloop, Songflute, Velvetwater, and Coffeesong, because "you can't go wrong with Sinatra." I voted for Lucid Engine. Jai answered with fifteen more of his own, from Dcifer to Glass Elephant. Number five was Fieldglass, a small telescope for studying terrain.
It stuck. Our new investors at BlueStream told Jai over Christmas that they liked it too, and that if we were going to change, we should do it sooner rather than later. We agreed on $6,500 for fieldglass.com, and by February 2001 we had the website and the business cards to match.
Act three: Becoming a contender (2001 to 2005)
2001 to 2005. Technology digs out of the bust. Venture money is scarce and cautious, Sarbanes-Oxley has every large company rethinking its controls and spending, and offshoring is reshaping how work gets done. Software as a service is a curiosity until Salesforce goes public in 2004.
Once we were no longer fighting for mere existence, we could turn to building the business. This act runs from the close of that first big round to the close of our last one.
On the leadership front, we added a CFO, whom I called our first adult, which probably explains why we started putting processes in place. Toward the end of this period we also changed sales leadership. That was very hard, because it meant a founding colleague and a good friend to all of us. By the end of 2004 we had the executive team that would take us through the Madison Dearborn Partners (MDP) deal, and nearly all of them through to SAP.
We also refined two things in this period: our channel strategy and our pricing.
Channel was an enlightening few years, to say the least. One thing I definitely learned is that if you plan to disintermediate a supply chain, be prepared for it to fight back. There was, and still is, an industry in our space called managed service providers, or MSPs, who manage a large company's relationships with its staffing suppliers. They were there before us; in fact our industry has its roots in MSPs automating their ordering, timekeeping and invoicing. So when we started, the MSPs eyed us with suspicion, and rightly so, because some of our competitors ran their own MSP businesses and competed with them. We never did. We kept to a pure software model, and over time we and the MSPs learned each other's strengths and built complementary roles. That became a set of trusted partnerships that gave us a big advantage, and our competitors have since had to copy the approach.
Pricing was hard at first. The prevailing model was the software license, but we were this odd ASP-like thing, and few companies would write a large check to an unproven company. Cost-plus pricing wouldn't work, because we didn't yet know what it would cost us once things settled, and if we'd guessed we would have been way off. Value-based pricing wouldn't work either, because the savings were too hard to compute and too easy to game. We ended up adopting the MSPs' pricing scheme, which happened to look a lot like what is now the classic SaaS pricing model. And we removed all the miscellaneous fees, which was an important lesson: they barely moved the needle, they weren't worth the bookkeeping, and they aggravated our customers.
We raised two more rounds in this period, for development, sales, marketing and expanding around the world. Our Series C, $17 million, was led by two East Coast venture firms, StarVest Partners of New York and HLM Venture Partners of Boston, with BlueStream Ventures back in. Our Series D, $11 million, was led by another East Coast firm, Grotech Capital Group, and it also gave us the chance to restructure the cap table. That is an important but long discussion for another day. The short version is that it was obvious we were going to be worth something one day, but the terms had reached the point where management had too little incentive. So that round, as late as it came, was a key one. All told, we had raised $38.5 million, about $38 million as the press later reported it.
Act four: The MDP waypoint (2005 to 2010)
2005 to 2010. Technology booms again. Amazon launches its cloud services in 2006, the iPhone arrives in 2007, and cloud software becomes respectable. Then Lehman Brothers fails in September 2008, credit freezes, and private equity deals collapse across the market.
The fourth act runs from our last venture round to MDP taking a majority stake in 2010.
By now our industry was broadly accepted. Forrester and Gartner covered it, a cottage industry of consultants had sprung up around it, and every staffing company worked with one or more VMS providers. Fieldglass was consistently number one or a close number two in a shrinking field of maybe four or five real competitors.
In 2007, IQNavigator, something of an arch-nemesis, made a big splash with a sale, large for the time, rumored at around $100 million, to GTCR, a private equity firm in Chicago. It was demoralizing, since IQNavigator parlayed it into the story that they were clearly number one. But it also validated our industry and showed us a big outcome was possible.
So in 2008 we began a process of our own, which ended with a private equity buyer, not yet MDP, at a number close to IQNavigator's. But 2008 turned out to be a rather bad year for financial firms. The price came down twice, and we started to balk. Our venture investors, though, were ready to close out their funds; some had been in for six to eight years. Then one day we heard that the private equity firm was being shut down and the deal was off.
That was a double serving of humble pie: our competitor had sold at a premium, and our own deal had fallen apart at the altar. Fortunately, we had been cash-flow positive since July 2006, with margins rising ever since, so we could ride it out. We went back to what we did best, which was running our business, and improved it in several ways.
On the product side, we moved to a date-driven release cycle. We went from an erratic schedule driven by what was in each release to three releases a year, at the same times every year. In my opinion it was one of the best things we ever did. It gave partners and customers plenty of notice to prepare, it helped our implementation teams schedule launches around the enhancements clients needed, and it forced us to pay close attention to scope. We also became obsessed with user experience and simplicity, which became the new mantra for how we designed the application. And we became much less concerned with what our competitors were doing, and much more focused on the problems we could solve for our growing number of customers.
On the services side, we stopped treating professional services as a profit center and started treating it as the thing that makes recurring software revenue possible. We changed our integration strategy to accommodate customers' systems and lighten the load on their IT teams. Most importantly, we built new capabilities on top of our data and our community of practitioners, and both became huge differentiators.
In 2010 we began another process, but this time our record of growth, our financial strength and our growing dominance of the market put us in the driver's seat. In a very competitive process, MDP took a majority stake in a deal valued at more than $220 million, with management, Jai included, keeping a significant share. That was more than twice IQNavigator's price three years earlier, and more than three times what we had nearly sold for two years earlier. MDP is a great private equity firm and a pleasure to work with. Their approach is to pick winning teams and back them.
Act five: The road to SAP (2010 to 2014)
2010 to 2014. The economy recovers, the cloud has won the argument, and the big enterprise software companies are buying their way into it. SAP alone buys SuccessFactors in 2011 and Ariba in 2012.
The fifth act, and the final one for me, was a period of huge acceleration. We started 2010 with 120 people and ended 2013 with 440, and we were trying as hard as we could to hire many more. We scaled the business significantly, in technology, people and process. We became the undisputed market leader, winning three out of every four deals. And we doubled down on the things that had made us successful.
We learned a lot about scaling. I'll share three lessons.
First, you need instrumentation for your business. Beyond a certain size, metrics are the only way to see the big picture. I try always to have a trio of metrics that capture growth, throughput or efficiency, and quality.
Second, the people side of the business becomes critical. This was particularly hard for us because, like many entrepreneurs, we had a "suck it up and own it" attitude, and that doesn't scale. Below a certain size, you build good leaders deliberately, by hiring good people and investing in them. Above it, people begin to emerge as leaders on their own, and management's job shifts to finding those emerging leaders and accelerating them. Part of that meant creating an environment where people could take risks and fail safely, which was incredibly hard to do. Letting go like that was essential to scaling. People also needed career paths: a way to see their jobs as a chance to grow into more senior roles.
Third, process. Yes, the bane of every entrepreneur. But processes are what guide your people when you can no longer do it all yourself. They are what make delivery consistent. They are, in fact, the key to scaling.
Around mid to late 2013 we informally began to explore our options. We found ourselves in an amazing position: we had the growth of a high-growth SaaS company and the margins of an old-school software company. From a strategic buyer's point of view, contingent labor had become a large enough share of most companies' headcount and spend that they needed systems to manage it. Fieldglass was the clear winner in the space, and the next two players were undesirable for various reasons. A Fieldglass exit was inevitable, and each potential buyer had to consider what it would mean if we went to a competitor.
Unlike the MDP deal, which was a formal process, the sale to SAP was informal, and it was fast. We began talking in late 2013, announced the deal in March 2014, at about $1 billion, and closed in May. It was a great outcome for MDP and for all of our investors.