At this point in the Capital Expertise series you should be able to recognise certain elements of capital expertise if I give you a new case, without telling you what I think about it. Let’s put that to the test, and return to a story earlier in Dell’s history.
In 1986, Lee Walker was settling into his role as PC’s Limited’s new president and CFO. He had secured a lending facility with Texas Commerce Bank and stabilised the company’s capital needs; the crisis with both the Boltons and with Mbank was behind them. The company’s name change to Dell Computer Corporation still lay in its future; Walker broached the topic shortly after joining the company, but Michael Dell appeared rather attached to ‘PC’s Limited’, so Walker let it drop.
There were bigger problems. At the time, IBM, Apple, and Compaq had 50% share of the personal computer market. It was clear to most watchers that the PC industry’s boom times were over: sure, the initial craze had resulted in a wave of smaller, successful PC vendors — PC’s Limited amongst them — but the industry was already starting to consolidate. Walker was mostly worried about surviving the shakeout. IBM, Apple and Compaq were all formidable competitors: IBM and Apple were pioneers, and Compaq had raised an eye-popping amount of venture capital (at least for the time). And all of them were ‘slowly turning the thumbscrews of price and quality’ — driving smaller competitors out of business. In his autobiography, Walker talked up his worries about becoming the next Osborne Computer — which had hit more than $100 million in sales before imploding, and thus ‘disappearing into the PC history books alongside scores of dead and dying companies’.
The Quality Problem
Not everything was terrible, though. Walker and Dell scored a major win early on in their partnership. When he started as president Walker was hit with two, seemingly unrelated problems. The first issue was that their computers were blowing up — literally.
It was the fall of 1986, and I had been president of PC’s Limited for only a few months. A customer from Des Moines, Iowa, had just called, warning me that he was on his speaker phone recording our conversation and that he had his attorneys with him in close attention.
“Are you the president of this so-called computer company?” he shouted.
Without waiting for a reply, he began yelling that one of our computers had exploded in his bedroom, starting a fire that was contained only by the grace of God and the Des Moines fire department. As a result of this “explosion and conflagration,” his home-based business had been irreparably harmed, and he intended to sue both PC’s Limited and me. Moreover, he was in touch with the Better Business Bureau of Austin, where he had learned that he was but one of many who had been damaged by our “lousy computer firebombs.”
“I’m going to break you bastards in two,” he screamed. “Treble damages for interstate fraud,” he shrieked.
So we had issues.
The second problem was a man named Bob Swem, plant manager of Tandem Computer’s Austin division. Tandem Computer was ostensibly a competitor of PC’s Limited, but in reality served a completely different market. Unlike PC’s Limited’s consumer focus, Tandem made no-nonsense, highly available, fault-tolerant computers for banks, stock exchanges, telephone-switching systems and other commercial transaction processing systems.
Swem sent a letter addressed to the president of PC’s Limited, asserting that PC’s Limited was poaching Tandem personnel, and threatening legal action if they didn’t stop immediately. This did have a kernel of truth, as Austin’s economy had fallen off a cliff — destroyed by a horrendous real estate bust, on top of the devastation from plummeting crude oil prices. As if this wasn’t bad enough, the consolidation that Walker mentioned earlier was clear to everyone in the PC industry, and “employees of Data General Corporation, Digital Equipment Corporation, and even Tandem were beginning to look around and see issues with their companies’ business models.” PC’s Limited, on the other hand, kept growing like a weed; no wonder Tandem employees were reaching out.
To be clear, Walker regarded both the angry customer and Bob Swem as equally dangerous threats. The company may have had sufficient working capital for the moment, but only by the thinnest of margins. Walker knew PC’s Limited would die if hit with an onslaught of lawsuits. And while one irate customer from Des Moines was bad, the fact that a company computer had caught fire was a leading indicator that product quality was a ticking time bomb.
“We needed some fundamental fixes,” Walker wrote, “and we needed them soon. Our problem was that fundamental fixes took time and money. We had neither.”
Walker was well connected in the Austin business scene; he began asking around about Bob Swem. And what he learnt intrigued him.
Bob Swem was a fifty-year-old manager of the Tandem Computers plant in North Austin. Before his job at Tandem, Bob had worked at Data General. For him, Data General had been sheer hell, like working in a penitentiary. He had hated his time there. What a mistake it had been for him to join a company that discouraged employees from hanging out together, taking too much time for lunch. It was a company that, in his opinion, went beyond frugality to soul-killing cheapness.
However, Tandem Computers was an almost perfect match for Bob. He managed the factory that made its terminals. Tandem terminals were the faces of its fault-tolerant computer systems, systems that were nonstop, ultrareliable. Banks, stock exchanges, and ATMs needed fault-tolerant, fail-safe computer systems of the highest quality. And Tandem Computers dominated these markets.
Bob Swem was born and raised in Wichita, Kansas, not far from where I was born in Coffeyville. He worked while going to night school at Wichita State University, majoring in physics. He crapped out when he hit quantum mechanics. “I couldn’t figure out Schrödinger’s Equation,” he said.
Some people want to be doctors, teachers, writers. Bob’s ultimate aspiration had always been to be a plant manager. To Bob, being a plant manager was like being the head coach of a place that made something. And the place couldn’t just be anywhere. It had to be a plant somewhere west of the Mississippi River. He often said, “There are too many people who behave like jerks east of the Mississippi.”
Bob Swem wanted every terminal made in his plant to be of the highest quality and to have its own identity. He wanted every terminal made in his plant to be the responsibility of one person, and he wanted that person’s name on the terminal. When a broker on Wall Street was making his millions using his Tandem terminal, that broker needed to see Suzi Smith’s name right up front on the terminal. Bob wanted there to be a connection between the user and the maker. He wanted to get away from the nameless, faceless way of making things that Henry Ford had established in the 1920s with his mass production philosophy.
Another aspect of Bob’s innovation was giving each individual part within each product its own unique identity. Every capacitor, every resistor, every subpart of every product was bar-coded so that every part’s life trajectory could be tracked. Bob had been refining this idea for more than ten years, even before his time with Data General, back when he was still working for NCR in Wichita, Kansas.
Bob was obsessed with being able to track an order from the moment it began its life on the factory floor to the time it was shipped and, if need be, to the time it should ever come back.
Bob knew that he could submit all of his products to continuous testing to find out what parts were the first to fail under rigorous burn-in. So if a 50-watt capacitor was the first thing to destruct, then all customers with 50-watt capacitors could be contacted, and each would receive upgrades with 100-watt capacitors so they wouldn’t fail in the field. If fail-safe, reliable, nonstop fault-tolerant systems meant the highest possible quality, it meant proactive replacement of key parts before those parts could fail in the field.
Walker and Dell began working on a creative solution. Reflecting on this episode, decades later, Walker admits that he wasn’t sure when they hatched it — most likely whilst they were driving to Tandem’s plant to meet with Swem.
Walker had a simple proposition: what if PC’s Limited made a large order of Tandem Computers to run their plant, using Tandem’s ultra-reliable, high availability computers to ensure they were never down? And what if PC’s Limited took precisely the same system that Swem had invented and applied it to their business? Walker was appealing to Swem’s ego here: he proposed that they could use PC’s Limited plant as a showcase for both Tandem’s remarkable computers, as well as Bob’s elegant shop floor system. In exchange, Walker and Dell promised to stop looking at job seekers from Tandem.
“Bob,” Walker said, hoping in his heart of hearts that Swem would agree, “I want your answer by sundown.”
Later that evening, as the sun was setting on Walker’s home on Cat Mountain, Swem called. “Lee, are you willing to make a handshake deal on what you proposed?”
Walker said yes.
So that was two problems solved, at least. With Swem’s production system, Dell and Walker could nip their quality issues in the bud. And whilst they had little idea at the time, Swem’s innovations became the foundation of Dell Computer’s legendary production system — and within just a few short years. Walker notes that as good as Swem’s system worked for Tandem, it worked even better at Dell — for Dell computers had far more variation than Tandem machines, given their built-to-order model.
Dell's Growth Plateau
But still: how to deal with the impending shakeout?
Walker thought about this for months. One common, if mainstream take was that PC’s Limited ought to scrap their direct-to-consumer approach and sell to customers through dealers in physical stores. Walker writes, in his memoirs:
The retail store puzzle tormented me. The argument for working through dealers by selling PC’s Limited Computers in their stores was seductive. At that time, people did not buy big-ticket items through the mail. If customers had to choose between buying something they could touch and see in a store versus buying from a company over the phone, a so-called mail-order firm, wouldn’t these prospective customers usually buy from a store?
But Walker resisted such thinking. At one point he was invited to Texas A&M to debate this very topic, in front of an audience. Walker went up against two consultants from McKinsey and Company, who presented their arguments with fancy graphs and figures. Walker had no data backing up his arguments. He lost.
I was careful never to mention this story to Michael. He was fierce enough in his opinion without my revealing that McKinsey and Company and a whole room full of Aggies (Texas A&M folk) believed that to be big enough to survive and thrive, you had to sell through retail storefronts.
But for the life of me I could not accept that. I would not accept that.
So much for the dealer growth strategy.
The other solution was to start selling to larger customers. If they could sell to large corporations, Walker thought, the odds were pretty good that PC’s Limited would survive the shakeout. But here Walker and Dell were caught in a bind (all emphasis mine):
Again and again as we listened on the phone to our customers, the message came through loud and clear from large companies like General Motors, General Electric, general anything. They weren’t buying from us because we had no service and support in the field.
They could call us for help, they could send their computer back to us for repairs, but that was it. We were centered in Austin and had no presence beyond Austin. We were a mail-order firm.
We were in a bind. We didn’t have the resources to create a field operation, yet we couldn’t move forward without one. We couldn’t raise money if we were strictly a mail-order firm, but we couldn’t break away from being a mail-order firm if we couldn’t raise money.
On the 26th of September 1986, Walker called for a timeout. He brought ten people — six full-time Dell employees and four external advisors — to a retreat at the Vintner Inn in Santa Rosa, California, along the Russian River. Walker had been grappling with the tricky problem of their growth plateau for too long now, he felt that they needed a ‘quiet, generative place that would help them shatter their normal mind habits.’ He needed his team to invent their way out of their predicament.
Walker has a remarkable passage about this meeting in his memoirs (all emphasis mine):
After the happy question came wishing time when we were asked to make our wishes known around this service and support question. I don’t remember much about the wishes that were made, with one exception.
One reason I remember that singular wish was because it was so mindless, so impractical, so lacking in any appreciation for our money constraints. I could only groan inwardly. It was against the rules to voice or show any criticism of anyone’s wishes. Only with great discipline did I stifle my annoyance—and embarrassment, because it was my fault. I had taken a chance and invited a twenty-one-year-old salesperson, Kim Roell, to be part of our ten-person team.
Kim was the antiparticle to Seymour’s vastness—and easily the youngest among us, except for Future Boy, of course. (Seymour referred to PC industry commentator and writer Jim Seymour, and ‘Future Boy’ was Walker’s nickname for Dell).
She said, “I wish I could tell my customers we would fix their problems right away. We would have a technician on their doorstep the next day.”
I bit my lower lip. I chided myself, thinking that next time I needed to take more time and care when deciding who gets invited to a crucial meeting. A technician. On their doorstep. The next day.
But that wasn’t all. Kim topped her wish with a cherry: “And I wish it were free.”
Free. So simple-minded and naïve. I could scarcely keep from wagging my whole body into a huge “no, no, no.” I thought, “You cannot ignore gravity, and you cannot repeal fundamental economics. We cannot afford this waste of time.”
The moment passed quickly as other wishes were offered and written on big pieces of butcher paper that were then taped up to hang around our meeting room. Part of me felt a slight growing desperation. Nothing transformative seemed to be happening. God forbid if we had traveled almost 2,000 miles, spending many multiples of our capital base, only to limp home with the news that we had no news.
The team had returned from the retreat, and straight to the grind of day-to-day operations. But Walker couldn’t get Roell’s insane wish out of his head. As CFO he knew that Dell simply didn’t have the capital to deliver on Roell’s proposal. But by the end of the trip, nearly everyone present at Santa Rosa believed that providing a ‘free service’ was the right answer to their predicament. Sure, Dell couldn’t build a field support force, much less provide support for free. But were there other options?
Walker doesn’t say how he came up with the idea for what he eventually did — only what happened next. Five months after their retreat at Santa Rosa, he reached out to Honeywell Bull — then a mainframe manufacturer — and flew out to Bull’s headquarters with Dell counsel Kelly Guest for a two-day negotiation session. The offer: for every computer Dell sold in the US or the UK, Honeywell Bull would get a little over $30. In exchange, Honeywell Bull’s field support technicians — a national service network — would fix any Dell computer for up to one full year after they shipped it. More importantly, Dell ate the cost; they didn’t pass the $30 on to the consumer. Walker explains:
My thinking was this: If I took the point of view that we had just added $30 of cost to each computer, then my engineering/accounting brain argued for increasing our prices by at least $30 plus some markup. But I went to my nonlinear quantum physics brain, which began to hypothesize that if we said our service was free, the message our customers would hear would be our complete confidence in our computers. I thought our sales might grow exponentially and our fixed overhead for every computer would go down at a much greater rate than the $30. In other words, even if our sales doubled, our general and administrative overhead would not because it wouldn’t require us to add another president or anything like that. While this had elements of voodoo economics, I thought it was worth a try. I decided to go with the complete aspiration Kim Roell had wished for in the Russian River Valley.
All of this was possible, thanks to our manufacturing system, courtesy of Tandem’s Bob Swem. We knew precisely the individual details about each component in each computer we shipped. Each PC was bar-coded and had its own identity. We could tell the Honeywell Bull service guy exactly what parts he should take when making his service call. IBM and Compaq could not match this because they mass-produced “unpersonalized” personal computers. The whole notion of particularizing individual units was mind-blowing at that time. In our case, constraint had bred imagination. We turned out inefficiency into a breakthrough. Had we had the money to fix our manufacturing issues, I don’t know if we would have come up with such a creative solution. For us, it was very personal. We had the systems to walk our talk. The other guys didn’t.
(...) Such a silly wish made that day along the Russian River. A technician on the customer’s doorstep the next day. For free. Sweet, naïve Kim’s idea was not so dumb after all. We all finally realised that at the end of our three-day think. Her wish was the bolt of lightning needed to power up our flux capacitor, the necessary service and support end-component to our Swem manufacturing system.
Walker found a way to get the breakthrough they were searching for. In Play Nice But Win, Michael Dell explains what happened next:
We decided not to pass along the $30 to the customer — our calculation being that the guarantee of free on-site service would create enough new sales volume that the outlay would quickly pay for itself.
The calculation proved correct. This was a huge step up for us. We could now sell made-to-order PC’s Limited-branded computers to major corporations and the US government with the assurance of free on-site service — which meant that major corporations and the US government would no longer turn away our salespeople without listening to them.
And with a forty-five-year-old genuine grown-up and experienced entrepreneur as president and CFO, we now had access to all kinds of working-capital credit we couldn’t get before. Unlike the twenty-one-year-old CEO, Lee Walker could go to people like Frank Phillips at Texas Commerce Bank and say, “Look, Texaco, Exxon, Monsanto—all these companies, not to mention the US government—they all owe this company money. Give us a loan based on all these receivables.” And the bankers would say, “Okay, Lee, we don’t know about the kid, but we trust you.”
Still, mere credit wasn’t enough to take us where we wanted to go.
A year later, Dell went public.
A couple of questions to close this out:
- How is this instantiation of capital expertise different from the previous cases we’ve looked at?
- What are some surprising similarities? And what are some surprising dissimilarities?
- Business narratives are complex, which means that there are often other business concepts embedded in a story. What other concept instantiations do you observe?
(Note: these questions are motivated by Cognitive Flexibility Theory, which lays out what expertise looks like in ill-structured domains like business. You should read this essay if you’d like to understand why I’m asking these questions).
Originally published , last updated .