Update: I originally published this essay in late-2012. My attention recently drifted back to the topic of disruption in the telecom industry, prompting me to revisit what Mark Spencer achieved with Digium and the Asterisk product. I’m no less impressed today than I was four years ago, and so I put this at the top of the blog list again. What Mark accomplished is impressive and a worthy model for other disruptors to consider.
Mark Spencer presents me with a philosophical conundrum. Before an interview earlier this month I text him to say I’m so excited about our conversation that I hardly slept the night before. (He is understandably cautious of my enthusiasm.) After we spoke, I’m so confounded by the way he chose to tell his tale that I don’t sleep well for several more nights as my mind grapples with what it’s heard.
Mark is chief technology officer and founder of a company called Digium in Huntsville, Alabama. It supports and develops for an open source telecom platform called Asterisk which Mark (for lack of a better term) invented. Digium is to Asterisk what Red Hat is to Linux. And much as Linux evolved into the open source alternative to proprietary operating systems offered by companies such as Microsoft, so has Asterisk become an alternative to closed technology from the likes of telecom giants Cisco and Avaya.
Not everyone likes the open source model, but enough do that Red Hat has made a thriving business out of providing software, support and consulting services to those that do choose Linux. Likewise with Digium. It has found a loyal and growing base of followers who align with Asterisk’s open source philosophy, its price and its flexibility. Some subset of those Asterisk users, mostly small- and medium-sized companies, find value in paying Mark’s company to help them use the platform.
I have little doubt that Mark bristles at my word selection in the paragraphs above. That I call it “his” company and that he “invented” Asterisk. But those are accurate descriptions. Though he turned over day-to-day operations of the business to professional managers after raising a round of venture capital in 2007, he remains majority shareholder and has de facto voting control over board decisions. But Mark prefers inclusive language. Digium’s success is the result of the efforts of many, not just Mark, he points out several times in our interview. Asterisk’s adoption did not happen because of the code he originated, he adds, but because countless independent developers have committed their considerable energy and intellect to enhancing it and making it a better product for users.
Though its story is still being written, it’s not a stretch to call Digium a success at this point in its existence (which Mark characterizes as being in its late-adolescent or early-teenage years). The same applies to Mark. He has done more in his 35 years than most people could muster the ambition to even imagine accomplishing in their lifetimes. But he lets out a deep sigh when I ask him to tell the story in the context of this success.
Mark rejects my basic premise. “What is success?” he asks with implied disdain but unflinching politeness.
Call it success or call it something else, when he talks about where he is today and where Digium is, Mark has no interest in talking about the things he did. He takes me in a different direction altogether.
Let’s not tell a story about talent or skill leading to success, he intimates. Let’s talk about the importance of luck.
So it is this matter of luck, and its effects on outcomes – success or failure – that has kept me up more nights than I should admit since my last conversation with Mark.
The Problem of Survivorship Bias, a Disclaimer
Businesses fail much, much more often than they succeed…even if you define success by the impossibly low threshold of just surviving from one pay period to the next. Over 90 percent of all businesses fail. Success is a rare thing, and all those failures are quickly forgotten. You don’t find many studies – academic in nature or appearing on a blog like this one – seeking to find the unifying traits of failed companies. We’re much more interested in the successes and spend our time subjecting their leaders to batteries of tests to measure the precise reasons for their achievements.
In doing so we’re starting with a flawed premise: that in studying successes we learn what traits lead to success. Because we aren’t asking the same questions in the business graveyards, querying all the failures in search of what unifies them, we begin our exercise with a contaminated set of subjects – the successes; the survivors. We run the risk that those things they describe to us apply equally as well to the failures. But we’ll never know because we don’t ask the failures.
It’s called survivorship bias.
I’m holding up a mirror to myself here as this is obviously a flaw in my exceptional business series as it is with the more comprehensive works of, say, Jim Collins’ Good to Great or even Thomas Stanley’s The Millionaire Next Door. These exercises don’t actually prove anything, yet we’ll go on faith that there’s value in them regardless. (Noting, for the record, that it’s wise to be skeptical of what you learn. Take it all with a grain of salt.)
Before we get too deep into Mark and the Digium story, let’s setup a framework for how we think of the antecedents of success (and failure), those opportunities and decisions that (in retrospect at least) we point to as the reasons something either worked or didn’t.
Why Success? Why Failure? A Matrix View of Preparedness and Luck
Let’s think in terms of the following matrix:
On one axis we have some combination of skill, talent, intelligence, and planning that we will call preparedness. This is the more popular way to interpret the causes of outcomes and therefore pervades the stories being told. Looking through this lens, entrepreneurs take credit for things turning out well. It was a product of their skills and talent; their insights and intellect; their ability to divine the future. The success was the result of how prepared they were. It creates a sense that events somehow came out as planned…that the person was somehow in control of the outcome.
This is the prevailing model because it makes for really good stories. It allows journalists to write about a leader’s charisma, his genius, his timing, his ability to prognosticate. It creates protagonists for whom readers can relate and cheer. And there are few things readers like more than a good hero who gains control en route to his success.
The other axis is luck, and it comes in two forms: good and bad. When you take the luck worldview, you’re confessing that there was little control in determining the outcomes. Success is the result of good luck; failure is the result of bad luck. Success came from being in the right place at the right time, being connected to the right people, or just some random combination of events that tilted the roulette wheel and dropped the ball in your slot.
Let’s consider Mark’s Digium story with this framework in mind.
Story 1: Digium Success as Preparedness
While preparing to interview Mark I read several stories that had been written about Digium in a variety of publications. They were told exclusively from the preparedness vantage point. It should be noted that they’re not factually inaccurate. At most they just leave out some details that might complicate the reader’s interpretation of events. As a result, they make for great stories. I loved reading them, and they played no small part in my strong interest in talking with Mark.
In composite form, they went something like this:
Wunderkind Mark Spencer, a computer science major at Auburn University, needs a PBX – a piece of telecom equipment – for his fledgling Linux consulting business. He can’t (or won’t) spare the $10,000 to buy one, so he applies his considerable genius to just inventing an alternative version instead. Philosophically he’s an open source kind of dude, so he opens his code for the world to make better. Asterisk is born.
The design is so brilliant and the demand is so great, Mark soon has more customers in queue than he can manage, and so he makes a bold entrepreneurial pivot by changing his company from its Linux bent into being a pure-Asterisk player. Dissatisfaction with proprietary technology from lumbering giants like Cisco and (what was then still) Nortel sends the users his way in droves. They want innovation and a lower price tag. Mark’s natural charm spreads the word further, and soon Asterisk is on a explosive growth trajectory.
The Digium coffers begin filling with dollar bills from grateful users. He secures funding from top-notch venture capital firms in 2007, hires a professional management team to run the day-to-day, yet stays onboard as chairman and chief technology officer.
Digium has become an undisputed success.
In our interview, Mark sensed that I was trolling for similar anecdotes to pluck from that complex history – those 15 years spent building Digium from scratch – and concoct them into the same simplistic story. He resisted.
“When you’re on the outside,” he says, “it’s easy to say it’s a great story. The reality is that’s not how it worked at all.”
His own philosophy pushes him to the other extreme of the Preparedness-Luck Matrix. He doesn’t want to talk about talent, skill, intuition or intelligence. He tells me the success (though that’s not his word) is mainly from luck.
Story 2: Digium Success as Luck
Mark urges us to look at the individual milestones with a different perspective. Below is how he characterized the story. Afterwards he challenges me to interpret the outcome from a series of inflection points as anything other than luck.
Mark becomes interested in Linux as a hobby, and in college starts a little business to provide support for other users of the platform. To earn his computer science degree on time, he elects to complete an engineering co-op with a technology firm and is paired with ADTRAN, a telecom equipment and services company in Huntsville, Alabama. (More on this later.) This exposes him to the telecom industry and specific types of telecom equipment. He knew little about these things before the co-op.
Mark clicked with ADTRAN, and relationships there proved enduring. He went back to Auburn to complete his degree and continue building his business. ADTRAN was aware of his Linux expertise, and when they needed help in the area they offered to make an investment in his company as part of a partnership. This gives Mark capital with which to grow. He moves the shop to Huntsville.
But the dot-com implosion wipes out nearly all of his customers, and Linux support will no longer fuel Digium as a going concern. Luckily, Mark had started work on Asterisk, something he was competent to do only because of the time he spent with ADTRAN. He decides to focus the company on this product because…well, the alternatives were not attractive.
Asterisk was intended for his own needs initially, but gradually it becomes clear that it can help other businesses with their telecom needs. Mark focuses on taking care of his early adopters, slowly creating Asterisk-based products and support as they need it. He has no visions of taking on Cisco or the other giants, he just wants to build something customers can use. It satisfies a need – providing options for price and flexibility that small- and medium-sized businesses didn’t have before – and it begins to pry open a niche in the market. Plus it’s fun, and he likes the people working with him.
It just grows organically, hits revenues of $10 million in 2007 while growing quarter-over-quarter for more than five years and maintaining profitability. He looks at getting venture funding in 2006, but there are no attractive options. He tries again in 2007 and finds that the environment has changed and he can dictate the terms of the investment. He gets capital he needs to grow the business, hires a professional management team, steps-back his day-to-day involvement, and he remains majority shareholder and chairman of the board of directors.
There was no grand plan, Mark tells me. He just kind of stumbled from one thing to the next and events that needed to break in his direction…well, they did. And that’s the luck component. If the results of all these tiny little events and decisions had been different, Digium might not exist today or might be a different kind of business altogether.
That’s Mark’s interpretation. We’ll consider it some more later.
Preparation Meets Luck: There’s a Reason for the Cliché
The great Roman stoic Seneca is credited with the statement, “Luck is what happens when preparation meets opportunity.” While that spins the logic of our matrix a bit, I don’t think it’s inconsistent with the point I’m trying to make. Luck does not generally occur in a vacuum. It tends to smile on the prepared. It’s a cliché, but its wisdom abides nonetheless.
To push the preparedness-luck matrix to its logical conclusion, we need to consider what happens when the different extremes are combined.
Those who rely too heavily on the preparedness vantage point of the Digium story are not wrong, but they don’t incorporate the nuance that luck brings to outcomes. As we see from the matrix, high preparedness can lead to failure when combined with bad luck. But we don’t conduct our surveys in the graveyards of failed businesses (the “unlucky”), so our survivorship bias blinds us to how well-prepared many of the failures might have been.
Luck clearly plays a role in success. Any entrepreneur willing to gaze thoughtfully at his navel for more than a few minutes – willing to think critically about his experiences – would readily confess this. It can be so powerful that it overcomes any amount of preparation. We heard this point from Tom Pirelli in a previous interview in the Exceptional Business series. He tells us how his chance connection with Bob Noyce of Intel led to an investment in Enterprise Systems, Inc. that ultimately saved the company. He was a very prepared entrepreneur (smart, good market focus, customer-centered, loved by his employees), but without this bit of luck he would have been forced to shutter his business.
Likewise, if Tom had not been prepared, Bob Noyce would have recognized that ESI was a bad investment. Luck would have passed him by.
We should abide by this same logic when considering Mark’s interpretation. His argument for luck fails to consider how prepared he was when the inflection points came. I will contend that his unique preparation tilted the roulette wheel in his favor, helping land that spinning ball in his slot. He ends up in the top right corner of “Deserved Success” because luck happened to meet a tremendous amount of preparation.
To belabor it a bit more, let’s consider two of the inflection points upon which Digium’s success clearly depend.
Digium Inflection Points: Where Luck and Preparation Collide
Luck Event 1: All These Little Ideas
Mark told me that he constantly has all these little ideas. “Even the ones I think are pretty good,” he said, “turn out not to be. I’d say 99.9 percent of them fall by the wayside.”
His implication was clear: it was luck in the first place that the Asterisk inspiration popped into his head, perhaps only chance that it survived the battle with all the other thoughts competing for his conscious attention, and perhaps nothing more than serendipity that he toyed with it long enough to be developed.
Here I largely agree with Mark. In the primordial soup that is his mind, how is it that any idea fragments can combine with others, synthesizing their genetic pieces to produce a coherent thought?
This really drives to the heart of the inspiration question. Why do we ever have a new thought? That goes beyond the scope of this essay (yes, wise guy, there is a scope!), but suffice it to say that even in the mystical arts of creativity, the good ideas tend to present themselves to the most prepared minds. (For more on this topic, read Steven Johnson’s excellent book Where Good Ideas Come From.)
I had the good fortune of knowing Mark pretty well when we were just boys. We went to elementary, middle, and high schools together. While I was fretting about my teenage social status, Mark focused his mental energies on designing a microchip that helped neural networks learn. Intel’s wafer just couldn’t match the heavy duty processing needed for the artificial intelligence algorithms he was conjuring in his mind. He went so far as to create a working prototype.
My point is that Mark had been preparing his mind since a young age to seek out ideas, evaluate their potential, and then choose the few worth turning into functioning products.
I see at least two gates people would have to pass through to do what Mark did in originating Asterisk. First, they would have to possess broad familiarity with the problem. They would have to understand how a PBX functions. By itself, I would venture a guess that this would eliminate some 99.99 percent of the world’s population.
Second, they would have to possess enough knowledge of engineering principles to even begin to create a solution. Of the infinitesimally small remainder of humans that made it through the first gate, I would guess less than one percent of them could marry the ability to recognize the problem Mark saw with the ability to solve it from an engineer’s perspective.
Was it luck that he had this idea? Yes. But that luck collided with a tremendous amount of preparedness.
I don’t, however, have a strong counter to Mark’s argument that it was plain luck that his brain somehow attached to the Asterisk idea, filtered out all the competing ideas, and stuck with it to create a product. Perhaps he’s right. But I suspect there’s some meaningful intuition guiding his thoughts, even if it’s in that hard-to-define realm of the subconscious.
This same level of judgment and intuition is not available to the unprepared mind. Give me a million years, for example, to let my mind wander in its own inspiration process, and I’ll guarantee there are no little particles in it that could randomly combine to produce Asterisk.
Luck Event 2: The ADTRAN Scheduling Incident
Mark possesses some slight elements of the classic absent minded professor, and there are times when the pesky details of managing daily routines get short thrift. So it was in college. Mark was late submitting his schedule to the Auburn University registrar, this caused him to get locked out of the classes he needed to take, and this forced him to take an unexpected co-op semester in which he would intern for a business that would mentor a young college engineer in search of an apprenticeship.
Mark’s flub led him to ADTRAN, a Huntsville-based manufacturer of networking and telecommunications equipment. There he made enduring friendships and gained deep exposure to the telecom industry. Without this experience, Mark might have been able to recognize his $10,000 PBX problem and perhaps even create Asterisk, but he probably would not have recognized that it was a compelling business opportunity worth pursuing.
But ADTRAN has been more than that to Mark. In a nested series of chance occurrences, ADTRAN invested a substantial sum in Digium in 2000. Not so much because they were financially interested in the future of its Linux support services, but because they needed that support for themselves. It nests further as this deeper relationship with ADTRAN introduced Mark to the executive team that he would eventually hire at Digium in 2007 to provide a level of professional shepherding as the company moved into its next stage of evolution.
Consider all that ADTRAN has meant to Digium’s existence, Mark argues, and then go back to the fact that he only met ADTRAN because he forgot to register for his computer science classes in time.
Is there a better definition of plain luck?
Perhaps. But Mark’s premise here reveals his humility. He suggests that he was just another engineering student seeking to check-off his co-op prerequisite on a predictable path to graduation. He was anything but typical. Among other things, he was already running Digium’s predecessor company, a viable business predicated on Mark’s deep, self-taught expertise in Linux…an open source code base to which Mark had already made substantial contributions as a young coder.
No, he was not just another college student. He had already developed rare and valuable skills that made him exceptional. The ADTRAN accidental connection was fortuitous, but for whom? Which party got more value out of the transaction?
I won’t argue that it set Mark and Digium down a certain path, and that in tracking the plotted points along that line would create the illusion that each successive event owed its existence to the happy chance that Mark forgot to file his schedule on time. But I’m skeptical of that simplistic telling of the story just as Mark is skeptical of the preparedness telling. I suspect that while history happened to follow this path, there are many other paths it could have followed to end up in substantially the same place.
In the final analysis of this series of events, I’ll concede that the manner in which Mark connected with ADTRAN was luck. But the events that it set in motion were possible only because that lucky encounter matched the skills of a highly-prepared engineer-cum-entrepreneur with a company that would soon have desperate need for those particular skills.
A Place to Roost
This is where the cliché shows its wisdom. Luck does have a tendency to find the prepared. At some point Mark’s suggestion that luck plays the singular role in this outcome just doesn’t pass muster. It’s a refreshing display of humility from an entrepreneurial star, but I don’t think it’s fitting as a true explanation for why things turned out the way they have.
His argument for it during our interview was convincing. Indeed, the subsequent sleepless nights were the byproduct of me trying to reconcile his points with what I believe to be a more rounded thesis. It did not come easily to this underpowered brain.
His explanation is more a reflection on the character of Mark Spencer than anything. Mark’s innate humility places him alongside Gandhi and Mother Teresa in the pantheon of history’s tiniest egos. His tendency is to deflect credit, and a person with this trait who exhibits such success has few choices but to point his finger at lady luck.
But we know that if Mark had not spent his entire life preparing his mind to be a world-class engineer, he never would have recognized the need for an alternative to existing PBX technology. Nor would he have possessed the skills to develop Asterisk into a platform to satisfy those needs.
If Mark had not honed his abilities to both generate new ideas and infect everyone around him with his enthusiasm for his creations, it’s doubtful that other developers would have invested their time in contributing to the Asterisk source code. Or that ADTRAN would have treated him as anything other than a typical co-op who comes for a semester and then leaves to follow a different path. Instead, he developed a following of people who believed in his ideas and were willing to invest time and considerable amounts of money in helping those ideas flourish.
Philosophically, I gravitate toward Mark’s sensibilities on the role of luck in business success. People whom we might deem successful tend to self-administer hardy backslaps when telling their own stories of conquest. They focus too much on the preparedness angle, lavishing credit on their skills, their courage, or their brilliance while downplaying luck. (Though interestingly, luck tends to become a popular topic when describing their failures…as in it came from “bad” luck, not a lack of preparedness.)
Mark’s take on his success is refreshing if nothing else, though I find plenty of places to disagree with him. The track record he has amassed does not come from lucky event compounding upon lucky event. Lady luck played a role, no doubt, but only because Mark’s preparation provided her a nest when she was prepared to roost.