ArsDigita: From Start-Up to Bust-Up
If you’ve been hanging out around courthouses in Delaware lately, you may have heard about some legal acrimony involving ArsDigita’s venture capitalists versus the ArsDigita co-founders. This letter explains how it came about (from the perspective of one of the defendants).
Note that 99 percent of the information in this document is irrelevant to the lawsuit. The lawsuit has to do with the rights of the shareholders to control management based on some technical points of law and contract. In other words, the questions of who is best qualified to run the company and whether business decisions have been correct are largely irrelevant.
Let’s go back to 1993. That’s when we started developing the domain knowledge that led to the ArsDigita Community System product. Most people who’ve made money in the software business are those who wrapped their minds around a problem earlier than others. You can’t base a business on “we’ll be better programmers than the folks at Microsoft and Oracle”; each company has enough computer science PhDs and expert software engineers to bury 100 competitors. You can, however, base a business on “we’ll attack this problem a few years before Microsoft and Oracle notice it and recognize it as a problem.”
Adobe is a good example of a small software company that has thrived despite possible competition from much larger companies. Check out http://www.adobe.com/aboutadobe/pressroom/executivebios/johnwarnock.html and http://www.adobe.com/aboutadobe/pressroom/executivebios/charlesgeschke.html You can see that these two guys, who have managed Adobe since its inception, spent a lot of time at Xerox PARC and Evans and Sutherland grappling with substantially the same kinds of problems for which Adobe provides solutions.
Adobe’s engineers and founders aren’t smarter than Microsoft’s; they merely started thinking about graphics and publishing before the Microsoft folks did.
Fast forward to 1998. We have contracts with a few big companies: AOL, HP, Levi Strauss, Oracle. We are recognized as thought leaders (publication by Macmillan of Database Backed Web Sites) and market leaders (our open source software for online learning communities). As GE’s Jack Welch will tell you, it is a lot easier and more fun working for a company that is #1 or #2 in its market. For one thing, customers will knock on your door.
By 1999, customers were knocking like crazy. A good example was Siemens. They had a critical business problem that could be solved by the ArsDigita Community System. Recognizing the goodness of fit between our product and Siemens’s problem, Boston Consulting Group brought them to our old HQ (603 Franklin) and within two weeks we had a contract.
We expanded. We were still small, though, and we avoided direct confrontations with heavily financed competitors. They were closed-source; we were open source. We’d undermine them by creating a world-wide open-source standard rather than try to outshout them with full-page ads in Business 2.0. We laughed at most of the small closed-source companies, asking “What’s their marketing slogan? We’re just like Microsoft and Oracle but without the market leadership and profits? And how does that slogan work for recruiting?”
By March 2000 we had grown to 80 people. I was still CEO and beginning to feel nervous that, for every task in the company, I could not say exactly who was supposed to do what and by when. But we were profitable, with monthly service contract revenue coming in at a $20 million/year rate. We’d paid nearly $1 million in income tax on our profits for calendar year 1999. Not so bad considering that we built everything from a $10,000 investment.
We’d never sought venture capital but our revenue and profits were bringing some of the top East Coast firms to our door. Most of the time these guys were being forced by the frenzied times into investment in a company and figuring out how to get revenues later (and profits much much later). ArsDigita looked a lot better than than the typical “wing and a prayer” bunch of guys with a fancy spreadsheet. Despite 1000 percent annual growth, we had cash. Most of our revenue was recurring. Most of our customers were happy and loyal.
Companies don’t like to rely on enterprise software from small companies. There is too much risk that the vendor will go bankrupt. Open source ameliorates this risk to some extent but the tendency to stick to IBM, Microsoft, and Oracle is strong. We tried to present a face of financial invincibility to the world. We bought a Ferrari to give away to any employee who recruited 10 friends. In reality the car only cost $2,000 per month, the person who won it only got to drive it for as long as he or she was employed, and the cost of a Ferrari is much lower than 10 headhunter commissions. But sitting in the parking lot it gave us the appearance of extravagance while inside the building we were living the frugal life — in a world starved for software development talent, it would have been hard to lose money paying MIT-educated programmers $50–85,000 base salaries plus an end-of-year bonus based on accomplishment and the firm’s performance. We had a couple of other Ferrari-like schemes up our sleeves. One was a beach house on Cape Cod where teams of programmers would go to work and write. Another was ArsDigita University, a tuition-free post-baccalaureate one-year computer science program. These things sounded outrageous, gave people a way to remember who we were, gave journalists a reason to write about us (and they did), all while costing no more in total than our 1999 profit (i.e., practically nothing if our revenue had continued to grow).
At the end of March 2000 we closed a venture capital financing with Greylock and General Atlantic. By the time a couple of small checks arrived we had an extra $38 million to put in the bank. We figured that we could use the extra money to place some bets on product development and marketing. Under the product development rubric we thought we’d not make the client teams carry the full weight of ACS development on their shoulders. If they found a client whose needs were similar to what we wanted in the product, we’d do the job for a low-ish price to get experience with that problem (see the “domain knowledge” sentence above) and develop reusable code to enhance ACS. Under the marketing rubric we’d expand our “education marketing” program. Finally, we wanted working capital. A company with $20 million in revenue really needs to have about $10 million in the bank in case a customer doesn’t pay, the economy turns soft, an important project is late, etc. Because we’d been growing 1000 percent per year we never had more than a couple of million dollars in the bank.
The terms of the venture capital investment were that the VCs purchased stock that gave them about 30% of the issued shares. Under standard corporate governance, a minority ownership interest such as this would give the VCs little or no control over the direction of the company. So we also had a stockholder’s agreement that required the existing shareholders (myself and Jin Choi) to vote for a board of directors that consisted of
1 Greylock person
1 General Atlantic person
3 senior officers from ArsDigita, including the CEO
So the VCs would have 2 out of 7 board seats. The shareholders would elect the rest. Plus the VCs got veto power over certain kinds of big transactions, such as the buying of expensive capital equipment, the selling of the company, the acquiring of another company. Finally in the event that the company was sold, they were entitled to the first $38 million off the top of the deal (note that this makes all of the common shares theoretically worthless in the event of a sale for less than $38 million). The terms we’d been offered from the three other serious venture capitalist bidders were similar. They wanted “a seat at the table” but nobody was asking for absolute power over the company going forward; the firms proposed to help ArsDigita’s founders do what we’d been doing successful already.
In parallel to all of this VC stuff we’d been trying to recruit an “outside CEO”. Based on my conversations with successful business people around the world, I now believe this is a fundamentally bad idea. Even the most able person will need a few years to learn about a company’s market, challenge, mission, culture, and people. A fresh-from-the-outside CEO might be successful at a 50-year-old company with a huge bureaucracy that manages itself (cf. George W. Bush taking over the Federal Government). But young enterprises don’t have that kind of inherent stability.
Anyway, as it happens we recruited Allen Shaheen on the recommendation of Chip Hazard, a Greylock employee who would ultimately represent the firm on our Board. Allen came from Cambridge Technology Partners (CTP) where he managed a large group of consultants in the overseas division of this IT services firm. I knew that Allen had no background in the software products business and that he had not been responsible for establishing overall strategy and thought leadership at CTP. In short, he had always worked for someone else and in a less competitive business than software products. Still, the other candidates we’d interviewed had been either very poorly prepared (one from Lotus) or were very aggressive and in-your-face and my top managers at the time didn’t think that they could work with them. Allen seemed like the kind of guy who would work well with the difficult personalities populating the companies’ far-flung offices. He had experience managing a multi-national services business. So the plan was that I’d keep responsibility for engineering, education, and evangelism; Allen would build the rest of the business.
Within a few weeks of Allen’s arrival, I found people telling me that I had no power at all, pointing out that Allen and the two VCs could vote as a bloc on the Board. We had not yet filled the two outsider positions so this point was tough to argue. 3 out of 5 = absolute power. Period.
April 2000 through March 2001
For roughly one year Peter Bloom (General Atlantic), Chip Hazard (Greylock), and Allen Shaheen (CEO) exercised absolute power over ArsDigita Corporation. During this year they
- spent $20 million to get back to the same revenue that I had when I was CEO
- declined Microsoft’s offer (summer 2000) to be the first enterprise software company with a .NET product (a Microsoft employee came back from a follow-up meeting with Allen and said “He reminds me of a lot of CEOs of companies that we’ve worked with… that have gone bankrupt.”)
- deprecated the old feature-complete product (ACS 3.4) before finishing the new product (ACS 4.x); note that this is a well-known way to kill a company among people with software products experience; Informix self-destructed because people couldn’t figure out whether to run the old proven version 7 or the new fancy version 9 so they converted to Oracle instead)
- created a vastly higher cost structure; I had 80 people mostly on base salaries under $100,000 and was bringing in revenue at the rate of $20 million annually. The ArsDigita of Greylock, General Atlantic, and Allen had nearly 200 with lots of new executive positions at $200,000 or over, programmers at base salaries of $125,000, etc. Contributing to the high cost structure was the new culture of working 9–5 Monday through Friday. Allen, Greylock, and General Atlantic wouldn’t be in the building on weekends and neither would the employees bother to come in.
- surrendered market leadership and thought leadership
How could these three guys have achieved such dreadful results? For that it is worth looking at what kind of leadership is required for a software products company. First, you probably want someone who has previously founded and run a company or been CEO at a company founded by others (i.e., not someone who has been an employee his or her whole life). Second, you probably want someone who has previous experience as an executive in the software products business. Third, you probably want someone with domain knowledge. Fourth, you probably want someone with technical knowledge.
Whatever strengths Peter, Chip, and Allen may have, all three were 0 for 4 on the qualifications listed above.
Software products is a rough business because it moves fast and attracts smart people. Furthermore you have companies like Microsoft where people work nights and weekends backed up by a cash hoard of $20 billion and a global brand. As an investor, you never want to send your company up against the Microsofts of the world unless your managers are smart, hard-working, and have the right experience. If they don’t, you need to look for a less competitive business. Maybe you can offer training or admin services for a Microsoft or Oracle product. Or maybe you should get out of the IT business altogether and apply your capital and employees to something like party equipment rental (you don’t see too many table and chair rental companies with $20 billion in the bank and MIT PhDs working nights and weekends trying to put their competitors out of business).
At this point you might ask “Hey, weren’t you still on the Board?” Sure. But for most of this year Chip, Peter, and Allen didn’t want to listen to me. They even developed a theory for why they didn’t have to listen to me: I’d hurt their feelings by criticizing their performance and capabilities; self-esteem was the most important thing in running a business; ergo, because I was injuring their self-esteem it was better if they just turned a deaf ear. I’m not sure how much time these three guys had ever spent with engineers. Chuck Vest, the president of MIT, in a private communication to some faculty, once described MIT as “a no-praise zone”. My first week as an electrical engineering and computer science graduate student I asked a professor for help with a problem. He talked to me for a bit and then said “You’re having trouble with this problem because you don’t know anything and you’re not working very hard.”
A seat on a Board of Directors that never meets
After December 2000 they stopped having board meetings altogether. Instead they had “investor meetings” that were attended by Allen, Ern Blackwelder (the COO, who’d already been told that he was going to be replaced), Greylock, and General Atlantic. In other words, all five board members except me would meet.
Board-level decisions were made not only without the chairman having an opportunity to vote but without the chairman (me) even being given notice. For example, after that final December 2000 board meeting, Allen, Ern, and the VCs (a) decided how much bonus to pay the CEO and COO for 2000, (b) named Jim Jordan to be Chief Financial Officer (see http://www.arsdigita.com/news/), (c) decided to eliminate me as Chairman and announced to the press that I was already gone (implying that I’d resigned though it was untrue), (d) hired and appointed Richard Buck as Senior Vice President of Engineering, (e) hired and appointed Dave Menninger as Senior Vice President of Marketing (again, see http://www.arsdigita.com/news/ ), etc.
What about the two outsider seats?
At this point you might ask “Hey, what about those two outsider seats?” At various times during the rule of Greylock, General Atlantic, and Allen I would push for the nomination of someone with software products experience. Nobody was ever approved. On November 22, 2000 I emailed Bill Helman and Bill Kaiser, two other Greylock employees who’d pitched ArsDigita, trying to set up a meeting to discuss getting a couple of good outsider board members:
… As an investor in the company, though, I’m concerned that ArsDigita is left without a single engineering expert on the board or on the management team. I’m not sure what your experience is but mine is that it is tough for tech companies to succeed without some engineering expertise at or near the top. …
They were reluctant to get involved, saying that normally everything should be piped through the Greylock employee actually sitting on a portfolio company’s board, in this case Chip Hazard, the very person whose lack of engineering experience was contributing to ArsDigita’s bleed. Kaiser agreed to meet me, however, after a couple of weeks. We walked around the MIT campus for 30 minutes. When I explained the problems with the product and the financials, Kaiser said “Isn’t it possible that this is just your opinion, that Allen and Chip would see it differently?”
Relativism. It was impressive in a way to see Protagoras’s sophism alive and well after 2500 years. But the “all points of view are equally valid and supported only by someone’s opinion” ignores the fact that it is easy to measure the correctness of business beliefs: some people are losing money and some are making money; some companies are gaining market share while others are losing market share.
We gave up on the idea of finding any help from the Greylock corner.
With no voice in company operations and with a board seat in a company that did not have board meetings, it seemed that there was no longer anything that I could do to express the co-founders’ wills as shareholders. Keep in mind that ArsDigita had been running in exactly the opposite direction from the way that we wanted it run. We started aD slowly and carefully. We ran it profitably. We placed small bets. We handled money conservatively (though we tried to give the appearance of wildness and fantastic prosperity to the outside world there is actually nothing extravagant about having a fancy beach retreat for a team of programmers that is excited and working 6 days/week, 12 hours/day). We made sure that we were working as hard as teams at Microsoft and startup companies.
By contrast, Allen, Greylock, and General Atlantic presented us (Common shareholders) with a strategy of “here’s this spreadsheet that shows us going bankrupt in one year unless a big stream of license revenue starts coming in.” And, oh yes, the revenue would be coming from a product that had never been built, purchased by customers to whom we’d never sold anything.
Do these kinds of risks bother venture capitalists? Having a first-time CEO with zero experience in the industry? Staking everything on a to-be-finished software product? Perhaps not. General Atlantic has $10 billion under management, according to their Web site (gapartners.com). If they point ArsDigita in a direction that leads to tankage, they can fall back on their $9.98 billion in other investments. What about the Common shareholders, though? We never signed up for this kind of risk and we don’t have substantial other investments. I put 8 years of my life into ArsDigita Community System. Jin put in 4 years. We would be unhappy to see the company spend through its accumulated profits plus $38 million in capital merely so that three guys in suits could learn a little something about what it is like to run a software products company.
March 2001: The Final Shove
March 2001 was a dark time from our shareholder perspective. Some of our greatest assets were pushed out the door. David Rodriguez, for example, a man who had worked like a monster and delivered huge projects to happy clients. Did he refuse to implement abstract URL on the World Bank knowledge management system until I nagged him via email from Australia? Yes. Did he say that ACS 4.0 was unusable? Yes. Did he tell Allen “You talk like a press release”? Yes. These things disqualify dvr from diplomatic service. But as shareholders we didn’t like to see someone who had personally delivered more than $1 million in revenue while costing us perhaps $200,000 being pushed out the door. To a non-owner manager, it might make sense to get rid of someone who’d offended you with a harsh word. You’re still going to get more than a third of a million dollars in base plus bonus, even if the shareholders take a beating. But as an owner-CEO I would let an employee vent his spleen at me, secure in the knowledge that at the end of the day this guy was building the value of my shares.
Our co-founder Aurelius Prochazka was also axed in the March 2001 massacre. Have Jin and I had to clean up some of his code in the past? Yes. Was Aure rather discouraged and unproductive in the past few months as ArsDigita’s financial and market position slid and he reflected on the fact that unqualified people were managing the firm? Yes. Was Aure too quick to criticize highly paid executives whose intellectual abilities fell short of the standards he absorbed at Caltech? Yes. But what kind of a company can you have when you fire someone who is (a) a founder, one of the people who built a $20 million profitable enterprise on capital of $10,000, (b) someone who’d previously built a successful business and sold it, and (c) responsible for the innovative ideas and interface behind some of the ACS’s most interesting modules (e.g., file storage)? Aure’s PhD is in engineering and not in charm. But if shareholder value were related to average employee charm, Microsoft shareholders would be rather poor indeed.
So what were the shareholders doing in March 2001? Planning some research projects at MIT and Orange/France Telecom. Giving some one-day courses in Thailand and India. Revamping our Software Engineering for Internet Applications course at MIT (recently accepted by the faculty into the core curriculum and renumbered 6.171). In short, getting on with our lives and personal technical goals, working full-time for other companies. We cried if we thought about ArsDigita’s financial performance but mostly we tried not to think about it.
Sing, O goddess, the anger of Achilles son of Peleus, that brought countless ills upon the Achaeans. Many a brave soul did it send hurrying down to Hades, and many a hero did it yield a prey to dogs and vultures, for so were the counsels of Jove fulfilled from the day on which the son of Atreus, king of men, and great Achilles, first fell out with one another.
— Iliad, Book I
Peter Bloom, the General Atlantic employee representing their interest on our board, was not crying in March 2001. He was angry, as he had been for many months. Though Agamemnon had not taken his prize girl Briseis to replace the daughter of the priest Chryses, Bloom’s anger was not less than that of the great son of Peleus. My habit of pointing out that he’d accomplish more if he picked more important opponents (e.g., Microsoft and Oracle rather than a 37-year-old living in a 2-bedroom apartment in Cambridge) did not cool him down. What really sent him over the edge, as far as I can tell, was when I related my response to a member of the Harvard faculty who asked me what it was like to watch venture capitalists and professional managers run ArsDigita (I replied “like watching a group of nursery school children who’ve stolen a Boeing 747 and are now flipping all the switches trying to get it to take off”).
Peter Bloom sent me an email message on March 28, 2001: “Since you are so troubled by the direction that the company has taken, you can choose to resign from the board before our next meeting. This is your decision to make, but it is a course of action open to you to avoid the public humiliation and significant professional impairment of being removed as Chairman from a board of directors. … The actions you have taken and the written communication you have directed at individuals has now gotten you in very serious trouble and you need to turn to someone you trust for counsel. I sincerely hope that your trusted confidants will tell you the truth about the impending consequences of your recent communications and accusations before you irreparably impair your reputation and financial future.”
Shortly after I received the email message, I stopped by ArsDigita HQ to pick up Alex from Eve. My card key no longer opened the door.
When a company with $10 billion in assets threatens “irreparable impairment of one’s financial future” it is time to see a lawyer. So, thanks to Peter’s initiative, I trundled down to see Sam Mawn-Mahlau and Paul Mahoney at Edwards and Angell. They prepared a “shareholder’s consent” that would change the company by-laws so that, until ArsDigita went public, the CEO and president would be directly elected by the shareholders. The next item on the list was the election of Philip Greenspun as CEO. Another item in this shareholder’s consent was to elect two existing vice-presidents of the firm, Tracy Adams and Eve Andersson, to the board. The stockholder’s agreement said that the three insiders on the Board had to be “senior executives” so we promoted them to “Executive Vice President” just to be safe.
The effect of this shareholder consent was to trim the venture capitalists back to what they’d bargained for, i.e., two board seats plus veto power over major transactions.
Our shareholder vote happened to occur on the same day that CNET carried a story about how ArsDigita would henceforth abandon its open-source strategy in favor of traditional licensed software and how Philip Greenspun, the “former chairman”, had left the company. The next morning, April 6, a courier arrived at 80 Prospect Street (ArsDigita HQ) with a letter for Allen notifying him that he’d been demoted from “President and CEO” to “President”. I telephoned Allen to assure him that I didn’t want to make any major personnel changes immediately, that I’d be happy to consider the entire last year as water under the bridge and work with him under our original agreement (I’d keep responsibility for engineering, education, and evangelism; Allen would build the rest of the business). I said that I wanted to spend the next few weeks just coming up to speed on the status of the product, the customers, and the company. Allen told me just what I wanted to hear and I was encouraged by the idea of working through him.
April 2001: Allen and the Venture Capitalists File Suit
On April 11, 2001, the following lawsuit was filed in Delaware chancery court:
ALLEN SHAHEEN, ERNEST )
BLACKWELDER, GENERAL ATLANTIC )
PARTNERS 64, L.P., a Delaware )
limited partnership, GREYLOCK )
X LIMITED PARTNERSHIP, a )
Delaware limited partnership )
and ARSDIGITA CORPORATION, )
a Delaware corporation, )
v. ) Civil Action No. 18821
PHILIP GREENSPUN, EVE A. )
ANDERSSON and TRACY E. ADAMS )
As you see from the caption of the case, the lawsuit was filed by Allen, Ern, the VCs, and the corporation itself. We were quite confused by the form of the case, given that this is fundamentally a dispute between two groups of shareholders (the VCs versus the founders). So we’d not expected the corporation itself to have any interest in the case one way or the other.
A conversation with ArsDigita Corporation’s corporate counsel, Jay Hachigian from Gunderson, shed some light on the matter. It seems that Allen dipped his hands into the company checking account and scooped out a quarter million dollars to pay the venture capitalists’ attorneys in this matter. Jay cautioned the group that this was perhaps not the best idea but they apparently went ahead anyway. Thus we now have the spectacle of a group of shareholders trying to increase the level of accountability of a management team who has, in their view, been doing a bad job. That group of shareholders is being sued by the managers who want to avoid accountability. The lawsuit is being funded with the defendants’ own money!
The crux of the plaintiffs’ case is that Jin and I signed various agreements promising to do various things, e.g., always vote for a Greylock and General Atlantic representative on the board. The closing documents for our financing formed a stack about the size of a Manhattan Yellow Pages. Supposedly somewhere within this stack it is said that the Board of Directors of ArsDigita won’t amend the corporate by-laws without the consent of the venture capitalist members. Nowhere does it prohibit the shareholders from doing this, however. Greylock and General Atlantic would like to read this interpretation into the documents. If memory serves, those documents were drafted by Paul, Weiss (paulweiss.com), General Atlantic’s lawyers. So under standard legal doctrines, ambiguity ought to be construed against them. Of course, I’m not a lawyer and nobody can say what another human being, in this case a Delaware judge, is going to do. I personally think it would be a bit shocking for the judge to rule in favor of Greylock and General Atlantic. The effect of such a ruling would be to make the shareholders’ voting rights worthless, i.e., the judge would be saying that the VC firms could exercise absolute power forever as if they’d bought the voting rights on our Common shares at the time of the investment.
So, that’s the story. Keep in mind that most of what is in this document may well be irrelevant to the outcome of the lawsuit. A court generally does not want to decide which group of people is likely to make better business decisions. A court looks at issues such as “What rights do the owners of a majority of the shares of a company have to control its direction?” or “Can the venture capitalists add extra restrictions, a year later, to an agreement that they made and that is reflected in documents drafted by their own lawyers?”
It will take a couple of months to take everyone’s deposition and get through discovery. Then we’ll have a trial, maybe in June 2001, in front of a judge in Delaware. The judge might decide the case based on just a few documents, in which case all of the discovery will have been a waste of time.
Why I wrote this
Generally people try to say as little as possible during litigation. However, I’ve been getting progressively more and more annoyed listening to other folks’ characterizations of the lawsuit. ArsDigita management was running around the building telling folks that “Philip sued the company”, something that was plainly false. Allen was telling people “the venture capitalists have a very strong case” (how come he needed to pay a retainer of $250,000 of the shareholders’ money to defeat a couple of individuals if the venture capitalists have such a strong case?). Reporters were making it sound like this was a dispute about ego and control. In a way that was true; Allen, Chip, and Peter did not like being characterized as fools whose ignorance was costing the company $millions. But as far as Jin and I were concerned, the reporters’ spin was not true. We’d have been delighted to be passive shareholders in a successful profitable company. What we didn’t like was being passive shareholders in a company bleeding cash.
There were some simple practical motivations for writing this article. One of the beauties of the Web is that it can save one from having to repeat oneself. On any normal day, I get 50 email messages from readers of photo.net and philip.greenspun.com asking various questions. A few times every week, a reporter will email or telephone to ask a question. After Greylock and General Atlantic filed suit, the stream of questions about photography and computer science was supplemented by a flood of questions about the lawsuit. If I hadn’t written this article I might have gotten RSI in my wrists simply from typing “no comment” 200 times every day.
Finally, there are customers who’ve adopted ArsDigita Community System to consider and friends that we recruited to work at ArsDigita, the company that made a profit every month. These folks have a right to a better explanation than they’re going to get from a 500-word newspaper story or a corporate press release. The company’s birth and growth were public, chronicled in Chapter 2 of Philip and Alex’s Guide to Web Publishing. The folks who were kind enough to pay attention and support us are entitled to know how the rest of the story unfolds.