Steve Blank Innovation and Entrepreneurship October 2009

The Secret History of Silicon Valley 12: The Rise of “Risk Capital” Part 2

Posted on October 29, 2009 by steve blank

This post is the latest in the “Secret History Series.” They’ll make much more sense if you watch the video or read some of the earlier posts for context. See the Secret History bibliography for sources and supplemental reading.

This is the second of three posts about the rise of “risk capital” and how it came to be associated with what became Silicon Valley.

———————–

The First Valley IPO’s
Silicon Valley first caught the eyes of east coast investors in the late 1950’s when the valleys first three IPO’s happened: Varian in 1956, Hewlett Packard in 1957, and Ampex in 1958. These IPOs meant that technology companies didn’t have to get acquired to raise money or get their founders and investors liquid. Interestingly enough, Fred Terman, Dean of Stanford Engineering was tied to all three companies.

Varian made a high power microwave tube called the Klystron, invented by Terman’s students Russell and Sigurd Varian and William Hansen. In 1948 the Varian brothers along with Stanford professors Edward Ginzton and Marvin Chodorow founded Varian Corporation in Palo Alto to produce klystrons for military applications. Fred Terman and David Packard of HP joined Varian’s board.

Terman was also on the board of HP. Terman arranged for a research assistantship to bring his former student, David Packard, back from a job at General Electric in New York to collaborate with William Hewlett, another of Terman’s graduate students. Terman sat on the HP board from 1957-1973.

Ampex made the first tape recorders in the U.S (copied from captured German models,) and Terman was on its board as well. Ampex’s first customer was Bing Crosby who wanted to record his radio programs for rebroadcast (and had exclusive distribution rights.) Ampex business took off when Terman introduced Ampex founder Alex Poniatoff to Joseph and Henry McMicking. The McMicking’s bought 50% of Ampex for $365,000 (some liken this to the first VC investment in the valley.) McMicking and Terman introduced Ampex to the National Security Agency, and Ampex sales boomed when their audio and video recorders became the standard for Electronic Intelligence and telemetry signal collection recorders.

Meanwhile on the West Coast – “The Group” 1950’s
When Ampex was raising its money, in 1952, an employee of Fireman’s Fund in San Francisco, Reid Dennis, managed to put $20,000 in the deal. Five years later Dennis and a small group of angel investors who called themselves “The Group” started investing in new electronics companies being formed in the valley south of San Francisco. These angels who were all working in their day jobs at various financial institutions, would invite startup electronics companies up to San Francisco to pitch their deals and they would invest an average of $75 -$300K per deal.

The Group is worth noting for:

Investing their own private money,
Reid Dennis would found Institutional Venture Partners in 1974
First group specifically investing in the valley’s electronics industry

SBIC Act of 1958
During the cold war the launch of Sputnik-1 by the Soviet Union in 1957 both traumatized and galvanized the United States. Having the first earth satellite launched by a country that been portrayed as a third-world backwater with a bellicose foreign policy shocked the U.S. into believing it was behind the Soviet Union in innovation. In response, one of the many U.S. national initiatives (DARPA, NASA, Space Race, etc.) to spur innovation was a new government agency to fund new companies. The Small Business Investment Company (SBIC) Act in 1958 guaranteed that for every dollar a bank or financial institution invested in a new company, the U.S. government would invest three (up to $300,000.) So for every dollar that a fund invested, it would have four dollars to invest.

While SBIC’s were set up around the country, companies in Northern California including Bank of America, Firemans Fund and American Express (Reid Dennis of the Group ran theirs), began to set up SBIC funds to tap the emerging microwave and new semiconductor startups setting up shop south of San Francisco. And for the first time, private companies like Continental Capital, Pitch Johnson & Bill Draper and Sutter Hill were formed to take advantage of the government largesse from the SBA. Like all government programs, the SBIC was fond of paperwork, but it began to formalize, professionalize and standardize the way investors evaluated risk.

SBIC’s were worth noting for:

The good news – government money for startups encouraged a “risk capital” culture at large financial institutions.
The better news – government money encouraged private companies to form to invest in new startups
The bad news – the government was more interested in rules, regulations and accounting then startups (because some SBIC’s saw the government funds as a license to steal)
By 1968 over 600 SBIC funds provided 75% of all venture funding in the U.S.
In 1988 after the rise of the limited partnership that number would be 7%.

Limited Partnerships
By the end of the 1950’s there was still no clear consensus about how to best organize an investment company for risky ventures. Was it like George Doriot’s ARD venture fund – a publicly traded closed end mutual fund? Was it using government money as a private SBIC firm? Or was it some other form of organization? Many investors weren’t interested in working for a large company for a salary and bonus, and most hated the paperwork and salary limitations that the SBIC imposed. Was there some other structure?

The limited partnership offered one way to structure an investment company. The fund would have limited life. It would charge its investors annual “management fees” to pay for the firm’s salaries, building, etc. In a typical venture fund, the partners receive a 2% management fee.

But the biggest innovation was the “carried interest” (called the “carry”.) This is where the partners would make their money. They would get a share of the profits of the fund (typically 20%.) For the first time venture investors would have a very strong performance incentive.

In 1958 General William Draper, Rowan Gaither (founder of the RAND corporation) and Fred Anderson (a retired Air Force general) founded Draper, Gaither and Anderson, Silicon Valley’s (and possible the worlds) first limited partnership. The venture firm was funded by Laurance Rockefeller and Lazard Freres, but after some dispute lost to the sands of time, Rockefeller pulled his financing, and the firm was dissolved after the first fund.

The first limited partnership that lasted for a while was formed by Davis and Rock in 1961. Arthur Rock, an investment banker at Hayden Stone in New York (who helped broker the financing of Fairchild) moved out to San Francisco in 1961 and partnered with Tommy Davis. Davis (an ex-WWII OSS agent) then a VP at the Kern Land Company got involved with investing in technology companies through Fred Terman. Davis’s first investment in 1957 was Watkins-Johnson (the maker of microwave Traveling Wave Tubes for electronic intelligence systems) where he sat on its board with Fred Terman. Rock and Davis would raise a $5M fund from east coast institutions and while they invested only $3.4 million of it by the time they dissolved their partnership in 1968 – they returned $90 million to their limited partners – a 54% compound growth rate.

Limited partnerships are worth noting for:

By the 1970’s the limited partnership would become the preferred organizational form for venture investors
The “carried interest” (the “carry”) assured that the venture partners would only make real money if their investments were successful. Aligning their interests with their limited investors and the entrepreneurs they were investing in.
The limited life of each fund; 7-10 years of which 3-5 years would be spent actively investing, focused the firms on investments that could reasonably expect to have “exits” during the life of the fund.
The limited life of each fund allowed venture firms to be flexible. They could change the split of the carry in follow on funds, add partners with carry in subsequent funds, change investing strategy and focus in follow-on funds, etc.

Silicon Innovation Collides with Risk Capital
Lacking a “risk capital” infrastructure in the 1950’s military contracts and traditional bank loans were the only options microwave startups had for capital. The first semiconductor companies couldn’t even get that – Shockley and Fairchild could only be funded through corporate partners. But by the 1960’s the tidal wave of semiconductor startups would find a valley with a growing number of SBIC backed venture firms and limited partnerships.

A wave of silicon innovation was about to meet a pile of risk capital.

More on this in Part XIII of the Secret History of Silicon Valley.

Filed under: Secret History of Silicon Valley, Venture Capital | 7 Comments »

The Secret History of Silicon Valley 11: The Rise of “Risk Capital” Part 1

Posted on October 26, 2009 by steve blank

This is the first of two posts about the rise of “risk capital” and how it came to be associated with what became Silicon Valley.
———————–

Building Blocks of Entrepreneurship
By the mid 1950’s the groundwork for a culture and environment of entrepreneurship were taking shape on the east and west coasts of the United States. Stanford and MIT were building on the technology breakthroughs of World War II and graduating a generation of engineers into a consumer and cold war economy that seemed limitless. Communication between scientists, engineers and corporations were relatively open, and ideas flowed freely. There was an emerging culture of cooperation and entrepreneurial spirit.

At Stanford, Dean of Engineering Fred Terman wanted companies outside of the university to take Stanford’s prototype microwave tubes and electronic intelligence systems and build production volumes for the military. While existing companies took some of the business, often it was a graduate student or professor who started a new company. The motivation in the mid 1950’s for these new startups was a crisis – we were in the midst of the cold war and the United States military and intelligence agencies were rearming as fast as they could.

Yet one of the most remarkable things about the boom in microwave and silicon startups occurring in the 1950’s and 60’s was that it was done without venture capital. There was none. Funding for the companies spinning out of Stanford’s engineering department in the 1950’s benefited from the tight integration and web of relationships between Fred Terman, Stanford, the U.S. military and intelligence agencies and defense contractors.

These technology startups had no risk capital – just customers/purchase orders from government/military/intelligence agencies.

This post is about the rise of “risk capital” and how it came to be associated with what became Silicon Valley.

Risk Capital via Family Money 1940’s
During the 1930’s, the heirs to U.S. family fortunes made in the late 19^th century – Rockefeller, Whitney, Bessemer – started to dabble in personal investments in new, risky ventures. Post World War II this generation recognized that:

Technology spin-offs coming out of WWII military research and development could lead to new, profitable companies
Entrepreneurs attempting to commercialize these new technologies could not get funding; (commercial and investment banks didn’t fund new companies, just the expansion of existing firms,) and existing companies would buy up entrepreneurs and their ideas, not fund them
There was no organized company to seek out and evaluate new venture ventures, manage investments in them and nurture their growth.

Several wealthy families in the U.S. set up companies to do just that – find and formalize investments in new and emerging industries.

In 1946 Jock Whitney started J.H. Whitney Company by writing a personal check for $5M and hiring Benno Schmidt as the first partner (Schmidt turned Whitney’s description of “private adventure capital” into the term “venture capital”).

Jock Whitney writes himself a check to fund J.H. Whitney Co.

That same year Laurance Rockefeller founded Rockefeller Brothers, Inc., with a check for $1.5 million. (23 years later they would rename the firm Venrock.)
Bessemer Securities, set up to invest the Phipps family fortune (Phipps was Andrew Carnegie’s partner,)

These early family money efforts are worth noting for:

They were “risk capital,” investing where others feared
They invested in a wide variety of new industries – from orange juice to airplanes
They almost exclusively focused on the East Coast
They used family money as the source of their investment funds

East Coast Venture Capital Experiments
In 1946, George Doriot, founded what is considered the first “venture capital firm” – American Research & Development (ARD). A Harvard Business School professor and early evangelist for entrepreneurs and entrepreneurship, Doriot was the Fred Terman of the East Coast. Doroit had the right idea with ARD (funding startups out of MIT and Harvard and raising money from outsiders who weren’t part of a private family) but picked the wrong model for raising capital for his firm. ARD was a publicly traded venture capital firm (raising $3.5 Million in 1946 as a closed-end mutual fund) which meant ARD was regulated by the Securities and Exchange Commission (SEC.) For reasons too numerous to mention here, this turned out to be a very bad idea. (It would be another three decades of experimentation before the majority of venture firms organized as limited partnerships.)

The region around Boston’s Route 128 would boom in the 1950’s-70’s with technology startups, many of them funded by ARD. ARD’s most famous investment was the $70,000 they put into Digital Equipment Corporation (DEC) in 1957 for 77% of the company that was worth hundreds of millions by its 1968 IPO. It wasn’t until the rise of the semiconductor industry and a unique startup culture in Silicon Valley that entrepreneurship became associated with the West Coast.

Georges Doriot the first VC

Doriot and American Research and Development are worth noting for:

Some of the very early VC’s got their venture capital education at Harvard as Doriot’s students (Arthur Rock, Peter Crisp, Charles Waite.)
ARD was almost exclusively focused on the East Coast
ARD proved that institutional investors, not just family money had an appetite for investing into venture capital firms.

Corporate Finance
One of the ironies in Silicon Valley is that the two companies which gave birth to its entire semiconductor industry weren’t funded by venture capital. Since neither of these startups were yet doing any business with the military—and venture capital as we know it today did not exist, they had to look elsewhere for funding. Instead, in 1956/57, Shockley Semiconductor Laboratory and Fairchild Semiconductor were both funded by corporate partners — Shockley by Beckman Instruments, Fairchild by Fairchild Camera and Instrument.

More on the rise of SBIC’s, Limited Partnerships and the venture capital industry as we know it today in Part XII of the Secret History of Silicon Valley here.

Filed under: Secret History of Silicon Valley, Venture Capital | 6 Comments »

Ardent War Story 6: Listen more, talk less

Posted on October 22, 2009 by steve blank

At Ardent we assembled an amazing group of talented engineers to build personal supercomputers to sell to scientists and engineers. (Context here.) The company failed.

Getting Out of the Building Wasn’t Entertainment – Discovery and Validation
Now that I was the master of the “facts” about customer needs in these specialized vertical markets, and with my team of vertical marketers, I thought I had achieved absolution and redemption. Opinions had been eliminated as part of marketing’s dialog inside the company; we had achieved “fact nirvana.” But there was one fatal flaw. As I enjoyed my post-graduate vertical marketing education, I had forgotten the real purpose of spending time in the field.

While understanding how customer’s do their work was one key part of Customer Discovery, I neglected the other key component – Customer Validation – to understand whether there were sufficient number of customers who had a problem that needed to be solved – and would pay to solve it. I had needed to ask customers four simple questions.

Did the customers know they had a problem?
If so, did they want to change the way they were doing things to solve that problem?
If so, how much would they pay to solve the problem?
Would they write us a Purchase Order now before our supercomputer was even complete, to be the first to solve their problems?

In hindsight, these questions seem blindingly apparent yet not asking them led to the ultimate demise of Ardent. I just assumed that since customers were talking to me and spending time with me, it must mean that they agreed with our new company’s vision and would spend piles of money with us. At this point in my career I didn’t understand that the goal of getting outside the building was not only finding markets with potential customers to sell to but also confirming the company’s vision, business model and product/market fit.

I had done a good job of Customer Discovery but failed at Customer Validation.

Ignoring the Red Flags
While I had lots of people willing to talk to me, we never really pushed hard to see if any customers were willing to buy and pay for the product before it shipped.

Early startup customers are visionaries just like the founders selling to them. If your startup’s vision is compelling enough, these early customers want to buy into the dream of what could be, and they want to get in early. They will put up with an unfinished system that barely works to get a competitive advantage outside their company (or sometimes a political one inside their company.) They will count on your startup to listen to their needs for subsequent releases or follow-on systems that actually deliver on the initial promise.

All industries, markets and segments have these visionary, early adopters. It is one of the wonderful intersections between human nature, capitalism, and startups. Not finding a sufficient set of these early visionaries is the biggest red flag a company can encounter. Ignoring these warning signs is fatal.

Product/Market Fit
Getting out of the building is not to collect feature lists from prospective customers nor run tons of focus groups (I had passed this test.) Instead it was to validate the product/market fit by discovering if their were enough customers who would buy our product as spec’d. This was where I had failed at Ardent. Once we had found our target customers we spent our meetings describing our new personal supercomputer and what it could do for these researchers instead of listening and truly understanding whether what we were offering was a “nice to have it” or “got to have it.”

If I had had actually been asking “Were we solving a problem these scientists and engineers felt they had?” I would have gotten a half-hearted “maybe.” If I had followed that up with a “If our personal supercomputer delivers as promised, would you write me a check now, before it ships?” I would have seen that no one was falling over themselves to be the first to buy our product. Another clue: lots of people said, “We’d try it if you give it to us.” That answer is always a dead give-away that you don’t yet have a product compelling enough to build a business.

As often happens in a startup, we confused our own vision and passion with the passion of our potential customers.

I had talked too much and listened too little.

What did the company do when we heard customer input that contradicted our business plan and assumptions? More in the next post.

Lessons learned:

We had “discovered” Ardent’s initial markets and customers

We spent too much time selling our vision and not enough time validating whether customers would actually buy

A lack of early, eager purchasers is a red-flag – time to revisit your business model

Filed under: Ardent, Customer Development | 12 Comments »

Ardent War Story 5: The Best Marketers Are Engineers

Posted on October 19, 2009 by steve blank

At Ardent we were building personal supercomputers to sell to scientists and engineers. (Context here.) While the last post was titled “You Know You’re Getting Close to Your Customers When They Offer You a Job“, this post should probably be titled, “You Know You’re Getting Close to Your Customers When You Offer Them a Job.”

I would discover that there was a more effective alternative in building a marketing department than hiring traditional marketers with MBA’s.

Building an Advisory Board
In my travels outside the building I kept my eyes out for articulate and visionary scientists and engineers who had expertise we lacked, and were willing to help in an advisory capacity. I set up an advisory board as a vehicle to get these industry experts engaged with the company and product. Some of these advisors from the academic community would work with our of VP of Engineering and help us solve specific technical problems.

Other advisors provided marketing with industry-specific advice in our initial vertical markets (computational fluid dynamics, computational chemistry, finite element analysis, and petroleum engineering). They gave us input on 1) features our system needed, 2) what applications we needed to have, and 3) how to sell to people just like them. Of course we also hoped that in listening to their advice in how to build the perfect computer for customers just like them, they would actually buy one of the first computers. Since some of these advisory board members were leaders in their fields, we knew they would tell their peers about our company. Our company’s stock was an inducement, but all of them were in it to help us build a better computer.

Velvet Painting Period – MFLOPS Poster

Engineers as Marketers
There was one other reason I was talent-spotting our advisors and potential customers. In most other companies a product-marketing department was responsible for the pricing, positioning and promotion of the product. Yet in our case the product, the machine as delivered from engineering, was a blank, featureless computer with just an operating system and compilers. The hardware held no interest for our target customers until it had become a “whole product,” – that is not until the computer had the complete suite of applications appropriate for a scientist in their specific vertical market – i.e. all the applications to run computational fluid dynamics or finite element analysis.

While I had learned a lot about our target markets in the first few months, I would never know as much as people who had spent their careers in these fields. Since the universe of people who were great marketers who also understood these esoteric applications like finite element analysis could be counted on one hand, (and were all working at Cray, the market leader) my choices were limited. I could either hire smart MBAs who were generalists and try to get them up to speed on these simulation applications, or I could hire some of the most articulate domain experts and teach them how to be marketers.

I chose to hire engineers from within each of our target markets and set up “Steve’s one month MBA course for engineers.”

At the time this was a pretty controversial decision. These hires were definitely not your standard marketing types. We hired a PhD in computational fluid dynamics from Duke who had worked on helicopter design. (Years later he would become a venture capitalist at Sequoia Capital.)

Our head of finite element analysis came from General Motor’s Chevy division, where he headed up one of their analysis groups. (He would go on to be a co-founder of two mechanical engineering software companies.) The rest of the vertical marketing recruits had similar backgrounds (and similar careers.) (I never could find one from the petroleum industry so I wore that hat along with the VP of Marketing title.)

Few of them had ever seen a data sheet or a price list let alone written one, but they were domain experts, they knew their fields, and they could communicate the benefits of owning a machine like ours to run their applications. They knew which applications were critical for their markets and which were nice-to-have. And they were responsible for helping our 3^rd party software group reach the right application providers to port their software to our computer. Since these marketers knew what publications their peers read and what conferences and trade shows they attended, they led our presence at the right shows and conferences. They knew the technology trendsetters in their fields and got us in front of them. In short order they learned how to transition from being customers on the receiving end of a sales pitch to giving one. To a person they became passionate evangelists and effective marketers.

Technical Marketing
Years later in my career I would realize I had simply reinvented what the early pioneers in Silicon Valley knew and did – hiring engineers who were domain experts who could talk as peers to customers and communicate effectively with their own company’s engineers. (Back in the 1960’s and 70’s no sane MBA’s would work for a Silicon Valley startup.) While MBA’s have a ton of useful skills, what they don’t have is what most marketing departments lack – customer insight. I found that having a senior marketer responsible for business strategy surrounded by ex-engineers and domain experts makes one heck of a powerful marketing department.

A quick diagnostic I now use for marketing departments: if you are in a startup selling to a specific set of customers and/or industry and your marketing department doesn’t have any people from that industry, your tenure as a VP of Marketing has passed its half-life.

Lessons learned:

Advisory boards with domain experts get you connected quickly to customer needs

In specialized markets, hire domain experts, and teach them to be marketers

Filed under: Ardent, Customer Development | 12 Comments »

Ardent War Story 4: You Know You’re Getting Close to Your Customers When They Offer You a Job

Posted on October 15, 2009 by steve blank

In 1985 Ardent Computer was determined to create a market niche for personal supercomputers. To understand our potential markets, we started by analyzing the marketing literature from Cray Research then crisscrossed the country talking to prospective customers – scientists and researchers in advanced corporate R&D centers and universities – to understand their needs.

A week might start with a visit to the MIT Media Lab, the next day at Princeton in the Aerospace Engineering department, then off to General Motors’ advanced research group, across to the computer science department at the University of Illinois, up to Minneapolis to meet with ETA, Control Data and Cray, and across the country to Seattle to speak with Boeing’s advanced propulsion group before returning to to the geophysics department at Stanford.

Simulation applications
After six months, we hypothesized that our most likely customers were scientists and engineers who used one of five applications: computational fluid dynamics, finite element analysis, computational chemistry and seismic data processing and reservoir simulation.

At Boeing we had learned aircraft designers needed to calculate the airflow and turbulence around wings and engines. Instead of building a new wing to test designs, numerical simulation would allow them to use a supercomputer to build a virtual model of a wing on the screen and use an application called computational fluid dynamics to watch the resulting airflow without ever flying a plane. If they didn’t like what they saw (say the wing had more drag than expected), they could change the design and rerun the simulation.

At General Motors we heard from mechanical engineers who needed to calculate the strength, breaking point and failure modes of structures – everything from piston rods to bumpers. Their interest was easy to understand. Before computer simulation, they would test real objects until they physically broke (or get sued when something important broke, blew up, or collapsed.) Now applications called finite element analysis could calculate these stresses and failure modes on a computer screen.

A third simulation market, this one new and just emerging, allowed biologists to examine how drugs would interact by simulating them on a computer. A precursor to today’s biotech revolution, these computational chemistry applications allowed the active docking sites of potential drugs to be modeled and tested on a computer screen rather than in a test tube.

Finally, we could see that petroleum engineers at oil companies like Chevron and Exxon were using computers in exploration and extraction with seismic data processing and reservoir simulation, applications which were moving oil companies into the supercomputer age.

Traveling around the country had helped me begin to understand how these customers currently did their work, what journals they read, where they got their funding, what other software they ran on their machines, etc. I came back to the company and described the day-in-the-life of each type of customer.

This was one of the happiest times in my life as a marketer. I had known nothing about supercomputers and numerical simulation applications; now there wasn’t a day that went by that I wasn’t learning something new. As I traveled to some of the most arcane trade shows and conferences (AIAA, SPE, MSC, etc.), my hotel room was stacked with the journals and textbooks about each vertical market just to keep up with the people we were meeting. (I was a marketer, not an engineer and most of the fine points were way over my head – and probably not just the fine points. But reading their literature allowed me to discuss the problems and opportunities with customers.)

My Velvet Painting Period

You Know You’re Getting Close to Your Customers When They Offer You a Job
I believed that good marketers used their own products. I got facile enough with a few of the applications that I could even run some of them myself. I could build simple finite element models with Patran and set up a run of the Nastran analysis codes.

Later on in the company’s life I went to give a lunch-time seminar to Chevron’s La Habra research center on the use of graphics supercomputers in petroleum applications. I spoke about the state of the art in computational reservoir simulation and what could be accomplished using finite difference and finite element methods on the new class of machines that were coming from companies like ours. During the question and answer session my heart was in my throat since like any good marketer, my depth of knowledge was no more than one level away from being a complete idiot. At the end of the talk the head of the research facility came up to me and said, “That was a great talk. We’re glad your company hired a real petroleum engineer to come speak to us. We hate when the sales and marketing types come down and try to get us to buy something.”

For one of the few times in my life I was at a loss for words, and I was completely unprepared for what came next. “Here’s my card, if you ever want to consider a career in Chevron research. We’d be happy to talk to you.”

Marketing was really fun.

Lessons learned:

To sell to customers you need to understand them: how they work, what they do and what problem you will solve for them.

You can’t understand customers from inside your building.

Filed under: Ardent, Customer Development, Marketing, Vertical Markets | 7 Comments »

Ardent 3: Supercomputer Porn

Posted on October 12, 2009 by steve blank

As VP of Marketing at our new startup, the CEO literally threw me out of the building and told me not to return until I understood the market and could identify the key applications and customers for Ardent’s new personal supercomputer. (See the previous Ardent posts for context.)

Supercomputers
With the introduction in 1976 of the Cray-1, supercomputers were defined as the fastest vector-processing computer, one in which a single instruction performed operations on an array rather than a single number. Cray’s first customers were the U.S. Nuclear Weapons Laboratories which used supercompurters to run their hydrodynamics codes to simulate what went on in the first microseconds in a nuclear weapon and the National Security Agency (with Cray putting in a special population count hardware instruction) used to facilitate decryption of codes.

At first only the national laboratories and the largest companies could afford to buy supercomputers, (the Cray-1 cost ~$9 million) but over time scientists and researchers were also starting to use them. Companies wanted to run numerical simulations to model things that were too expensive, too dangerous or too time consuming to physically build. Because of the vagaries of how floating point units in computers were designed, your average IBM mainframe of the day would take forever to run a simulation application. A supercomputer could be a 100x faster.

What Markets?
At Ardent our hypothesis was that if we could build a desktop supercomputer powerful enough to run and display these numerical simulations there were enough customers to make this a big business. My job was to figure out what markets Ardent should target, who were the key customers in these markets and what applications these customers had to have.

The problem was I didn’t have a clue. And while others in our new startup came from companies like Digital Equipment Corporation (DEC) that had sold computers to automate scientific instrumentation and process control, the computers we were building at Ardent were targeted to different customers and markets.

The one thing I did know is that we were probably going to be running some of the same applications as the market leader Cray. I concluded that my first job was to understand what Cray’s markets, customers and applications. When I learned that Cray users would be giving papers at the Society of Petroleum Engineering conference in Denver the next day, I got on a plane to listen and learn.

Follow the Leader
At the conference I attended a bunch of technical sessions, and got lost when the speaker got past, “My name is xxx.” I could see that quite a few oil companies were buying or thinking of buying their own supercomputer. As I walked out of the conference hall, I ran into a small booth with salespeople from Cray. Since their computers were way too large to bring to a trade show, the Cray booth just had literature describing their machines. I grabbed one of each piece of literature, stuffed into my bag, and wandered through the exhibit hall looking at other hardware and petroleum software companies.

Later I sat down for lunch and began to leaf through the bag of data sheets and brochures I had collected. Hmm… typical booth stuff…key chains, data sheets, pens… Until I got to the material from Cray.

Intelligence
As I leafed through the Cray sales material, a glossy magazine with the headline Cray Channels jumped out of the pile. Skimming a few pages, I realized that this particular issue was all about computational fluid dynamics, one of Cray’s key markets. The articles described the applications these users depended on and featured interviews with their most important customers.

I went back and looked at the cover not quite believing what I was reading was real. Cray Channels was describing my market, applications and customers for me. Was it possible that my market research was being handed to me by the existing market incumbent?

I kept thinking, “could this be possible? Did Cray ever publish any more of these magazines?” I looked at the cover of the magazine again and almost fell off my chair. It was Volume 7 issue 2. These magazines had been published for the last seven years. Could Cray have actually been describing their markets and users for that long?

I ran back to the Cray booth and as casually as I could, asked the salesman about the magazine. He assured me that each one profiled a different market, applications and users. I could order back issues from their publications department.

I don’t remember how quickly I got to a payphone, but I’m pretty sure that every back issue of these magazines were on the way to Sunnyvale by the end of that day.

Supercomputer Porn
When the back issues of Cray Channels arrived at Ardent, I ripped open the package with the Cray return address and eagerly started to flip through them. I was excited about what I was going to learn, yet somehow felt guilty, as if I really shouldn’t be looking at them. The pictures were great but I was reading it for the articles. But I was breathing heavy and it felt like I was looking at supercomputer porn. The magazines got passed around to all the engineers until they were dog-eared and worn.

I spent the next few days building a table with three columns: markets, applications, key customers. At the same time I had found the Wall Street analyst who followed Cray (now a public company) who kept a list of where every one of Cray’s machines was installed. I could now cross-correlate the markets by company who used supercomputers.

I started sharing what I had learned about potential target markets with our engineering team and my CEO. We agreed that now we had a roadmap, it was time to hit the road, talk to Cray customers and learn about supercomputer markets and applications in detail.

Lessons learned:

If you are in an existing market or trying to resgement an existing market you need to understand the market leader

Market leaders tend to educate the market

Step one for a startup is know what the market leader knows

Filed under: Ardent, Customer Development | 11 Comments »

Fun For Hours

Posted on October 10, 2009 by steve blank

Charts like the U.S. Frequency Allocation Chart keep me amused for hours. Download it here.

Interesting to consider how many billions of dollars of business is done over the electromagnetic spectrum we didn’t know existed 150 years ago.

Filed under: Technology | 1 Comment »

Ardent 2: Get Out of My Building

Posted on October 8, 2009 by steve blank

Some of the most important business lessons are learned in the most unlikely ways. At Ardent I learned many of them with a sharp smack on the side of the head from a brilliant but abusive boss. Not a process I recommend, but one in which the lessons stuck for a lifetime. (Read the previous Ardent post for context.)

Lessons to Learn
By the time I joined Ardent I thought I was an experienced marketer, but I’ll never forget my first real lesson in what it meant to understand customers and product/market fit.

We were sitting in our conference room in our first “system-planning meeting” trying to define the specifications of our new supercomputer and make the trade-offs between what was possible to build, and what customers in this new market would actually want and need. The conversation that day would become one of my professional watermarks.

Marketing is Heard From
Engineering was discussing how sophisticated the graphics portion of our computer should be, debating cost and time-to-market tradeoffs of arcane details such as double-buffering, 24 versus 32-bits of color, alpha channels, etc. I was pleased with myself that not only did I understand the issues, but I also had an opinion about what we should build. All of a sudden I decided that I hadn’t heard the sound of my own voice in a while so I piped up: “I think our customers will want 24-bits of double-buffered graphics.”

Silence descended across the conference table. The CEO turned to me and asked “What did you say?” Thinking he was impressed with my mastery of the subject as well as my brilliant observation, I repeated myself and embellished my initial observation with all the additional reasons why I thought our customers would want this feature. I was about to get an education that would last a lifetime.

Picture the scene: the entire company (all 15 of us) are present. For this startup we had assembled some of the best and brightest hardware and software engineers in the computer industry. My boss, the CEO, had just come from a string of successes at Convergent Technologies, Intel and Digital Equipment, names that at that time carried a lot of weight. Some of us had worked together in previous companies; some of us had just started working together for the first time. I thought I was bright, aggressive and could do no wrong as a marketer. I loved my job and I was convinced I was god’s gift to marketing. Now in a voice so quiet it could be barely heard across the conference table our CEO turns to me and says, “That’s what I thought you said. I just wanted to make sure I heard it correctly.” It was the last sentence I heard before my career trajectory as a marketer was permanently changed.

Get Out of My Company
At the top of his lungs he screamed, “You don’t know a damn thing about what these customers need! You’ve never talked to anyone in this market, you don’t know who they are, you don’t know what they need, and you have no right to speak in any of these planning meetings.” I was mortified with the dressing down in front of my friends as well as new employees I barely knew. Later my friends told me my face went pale. He continued yelling, “We have a technical team assembled in this room that has more knowledge of scientific customers and scientific computers than any other startup has ever had. They’ve been talking to these customers since before you were born, and they have a right to have an opinion. You are a disgrace to the marketing profession and have made a fool of yourself and will continue to do so every time you open your mouth. Get out of this conference room, get out of this building and get out of my company; you are wasting all of our time.”

I was stunned by the verbal onslaught. At that moment I felt so small I could have walked out of a room underneath the crack in a closed door.

Facts Not Opinions
The shock quickly wore off as I processed the gist of what he told me. He was right. I personally didn’t have any facts, and if we were counting opinions, there were a bunch more educated opinions in that room than I had. All I had been doing was filling the air with marketing noises.

I was convinced that I had just been humiliatingly fired – 90 days into our new company.

Get Out of the Building
As I got up to leave the room, the CEO said, “I want you out of the building talking to customers; find out who they are, how they work, and what we need to do to sell them lots of these new computers.” Motioning to our VP of Sales, he ordered: “Go with him and get him in front of customers, and both of you don’t come back until you can tell us something we don’t know.”

And he was smiling.

My career as marketer had just begun.

Lessons learned:

Corporate culture is either set by fiat, by default, or by consensus. But regardless of how it gets set, it gets set early

An intelligent opinion is still a guess

The dumbest person with a fact trumps anyone with an opinion

There are no facts inside the building so get the heck outside

Filed under: Ardent | Tagged: Customer Development, Early Stage Startup | 50 Comments »

Ardent 1: Supercomputers Get Personal

Posted on October 5, 2009 by steve blank

Last month on an east coast college tour with my daughter, I found myself in North Carolina for the first time in nearly 24 years.

I had last been in Chapel Hill on a winter’s day in 1986, traveling with the VP of Sales of our new supercomputer startup, Ardent. We were on the University of North Carolina campus to meet with Fred Brooks and Henry Fuchs. We just turned on the rental car radio as we entered campus and heard the mid day BBC news – the space shuttle Challenger had just exploded.

It was the best of times, it was the worst of times
Ardent would be my third technology company as a VP of Marketing (Convergent Technologies and MIPS Computers were the other two.) It would be the company where I actually earned the title.

This is the first of a series of posts on the company.

A Phone Call
After I left MIPS Computers I was in New York tagging along with a friend (a computer architect whose products at Apple a decade later would change the shape of personal computing) who was consulting for a voice recognition startup. We were sitting in our cheap hotel room when the phone rang. It was my ex boss from Convergent Technologies, “Steve we’ve all just resigned from Convergent and we’re starting a new company. I’ve convinced the team you’d be perfect, come join us as the VP of Marketing.” My ex-boss was going to be the VP of Engineering and I would report to the CEO whose marketing acumen and sales instincts seemed at the time to be telepathic and sense of theater was legend. And so was his reputation for being verbally abusive to his direct reports. Gulp.

The culture and work ethic of Convergent had earned it the title “the Marine Corps of Silicon Valley”. (Not until I was older and wiser did I realize that this was not always meant as a compliment.)

Working with my old boss sounded like a great idea. And in the course of the phone call I put my friend on the phone and let him interview for a job. On the ride to the airport my friend asked me what our new company was going to do.

Only then did I realize we both forgot to ask.

Never mind
The first idea for our new company was a software product that looked something like Hypertext. With a bit of research it turned out that a professor at Brown University had invented something close to what we had in mind. The VP of Sales and I flew to Providence to convince Andy van Dam at Brown to join our company, or at a minimum lead our advisory board.

On a rainy day in Providence we tracked Andy down just as he was leaving for a trip to Europe. He agreed to talk to us as he packed his office, and we followed him down the street as he went to get a haircut. With me holding the umbrella our VP of Sales kept reminding him how wonderful it would be if his research could turn into commercial products- all as we all walked downtown to the barbershop. While van Dam sat in the chair getting his haircut, the VP of Sales and I flanked him on either side, with the barber trying to get his clippers in between us. We were painting a picture of hypertext on every desktop computer. I knew we almost had him convinced when our sales guy and Andy started talking to each other in Dutch.

As the conversation began to get down to how much stock and salary we could offer van Dam, we left the barber to finish his work and went to a payphone to call our CEO to confirm the deal. The response from across the country? “Glad you two called, we were trying to get a hold of you guys. Forget the Hypertext idea and come on back to California. We’re building a supercomputer.” Oops. We told Andy we’d talk further when he got back from Europe.

Supercomputers get Personal
Back in Sunnyvale my friend had not only been hired but had convinced the team that we should be building hardware – making a new class of computers not a software application. Our vision was that just as the PC was revolutionizing the business market, we were going to do the same for scientists and engineers. We were going to target scientists and researchers who were longing to do “interactive simulations,” requiring both scientific computing and visualization of real-world phenomena. We were going to invent a new product and create an entirely new market by putting a personal graphics supercomputer on every desk.

By the mid 1980’s microprocessor technology—specifically off-the-shelf RISC-based microprocessors like the one from MIPS, my previous startup– had evolved to support the speed needed to support a new class of computers for scientists and engineers. Unlike Intel chips, MIPS chip architecture also made it possible to plug in a math co-processor. By adding a vector unit to these RISC processors, we believed we could take some of the supercomputer market from Cray (at the time the maker of the most powerful scientific computers in the world) as well as from the emerging class of mini-supercomputers (Convex and Alliant.)

To do that we needed to build a supercomputer, but since the RISC processors weren’t fast enough, we decided to build a multiprocessor supercomputer, (running up to 4 processors in parallel.) We had to write our parallelizing and vectorizing compilers and build our own high-end graphics boards, and write our own 3d graphics subroutine language – and put in all in a box that could fit in an office. Oh, and since it was not code compatible with anything, we were going to have to port all the key scientific applications our customers needed (as soon as we figured out who they were.) Some of the other founders had sold minicomputers to scientists and engineers, but no one knew or understood the unique class of applications and customers of supercomputers. We were going to be guessing.

Personal supercomputers meant yet again learning something completely new; new computer architectures, new applications and customers, new markets.

I couldn’t believe they were paying me to do this job; I would have gladly done it for free.

The Streets of Palo Alto
As our company was getting formed, I happened to bump into Gordon Bell – the ex VP of Engineering of DEC (the company that defined the minicomputer) on the streets of Palo Alto. (It was Gordon who had prodded John Hennessy and the MIPS team at Stanford to start a commercial chip company.) After telling Gordon what we were doing and who was doing it, he realized that he knew most of our founding team when they all had worked at DEC. I invited him to meet the team. A few days later Gordon became a founder. (Later he would leave for a few years to start the Computing Directorate at the National Science Foundation, help spec what became the Internet and then come back and run Ardent’s engineering.)

I would learn a ton from Gordon for over a decade, not only about practical heuristics for managing complex engineering projects (i.e. the “schedule fantasy factor,”) or his eleven rules of supercomputer design but also a real appreciation for how a technical visionary thinks. (I tried my best to narrow the time that I went from believing that Gordon had yet another insane idea to when I realized it was a profound insight.) It was a challenge to keep up with him (I never did) but it was fun to try.

At the same time Gordon was looking forward, he had a great appreciation of saving the past. He and his wife Gwen would found the Computer Museum, first in the lobby of DEC headquarters, then in Boston (and now as the Computer History Museum in Mountain View, California.) When our kids were little they would play with the computer artifacts (Napier bones and Pascal engines) scattered across their living room and overflowing their shelves when we stayed at their condo in Boston. My first inkling that computing had a history (with deep military connections) was looking at the SAGE air defense computer at the Boston Computer Museum.

I would be lucky in my career to work with Gordon and three other people I consider as mentors. They would all work in this one company.

Get Out of Building
Our trip to North Carolina was part of a year long effort to get out of the building to understand our market, customers and their applications. How I learned to “get out of the building” is in the next post.

Lessons learned:

Ardent’s personal supercomputer pushed at the edge of what was possible to build in technology
Our enthusiasm and passion for technology would soon intersect with our hypotheses about customers and market

Filed under: Ardent | Tagged: Early Stage Startup | 9 Comments »

Durant Versus Sloan – Part 1

Posted on October 1, 2009 by steve blank

The entrepreneur who built the largest startup in the United States is someone you probably never heard of. The guy who replaced him invented the idea of the modern corporation. Understanding the future of entrepreneurship may depend on understanding the contribution each of them made in the past.

This is the first of several posts on Durant versus Sloan.

Alfred P. Sloan
In the middle of the 20th Century Alfred P. Sloan was one of the most famous businessman in the world. Known as the “Inventor of the Modern Corporation,” Sloan was president of General Motors from 1923 to 1956 when the U.S. automotive industry grew to become one of the drivers of the global economy.

If you look around the United States it’s hard to avoid Sloan. There’s the Alfred P. Sloan Foundation, the Sloan School of Management at MIT, the Sloan program at Stanford, and the Sloan/Kettering Memorial Cancer Center in New York. Sloan’s book My Years with General Motors written 40 years ago is still a business classic.

The Modern Corporation
Sloan is rightly credited with formalizing the idea of the modern U.S. corporation, and by extension Sloan laid the foundation for America’s economic leadership in the 20th century. One guy really did all of this.

Peter Drucker wrote that Sloan was “the first to work out how to systematically organize a big company. When Sloan became president of GM in 1923 he put in place planning and strategy, measurements, and most importantly, the principles of decentralization.”

Sloan realized that the traditional centralized management structures (like General Motors had in 1920) were poor fits for the management of GM’s already diverse product lines. Top management was trying to coordinate all of the operating details (sales, manufacturing, distribution and marketing,) across all the divisions and the company almost went bankrupt that year when poor planning led to excess inventory (with unsold cars piling up at dealers and the company running out of cash.)

Sloan transferred responsibility down from corporate into each of the operating divisions (Chevrolet, Pontiac, Oldsmobile, Buick and Cadillac). Each of these GM divisions each focused on its own day-to-day operations and with each division general manager responsible for the division’s profit and loss. Sloan kept the corporate staff small and focused on policymaking, corporate finance and planning. Sloan had each of the divisions start systematic strategic planning.

Sloan put in place GM’s management accounting system (borrowed from DuPont) that for the first time allowed the company to: 1) produce an annual operating forecast that compared each division’s forecast (revenue, costs, capital requirements and return on investment) with the company’s financial goals. 2) Provide corporate management with near real-time divisional sales reports and budgets that indicated when they deviated from plan. 3) Allowed management to allocate resources and compensation among divisions based on a standard set of corporate-wide performance criteria.

Finally, Sloan transformed corporate management into a real profession, establishing the standard that the professional manager is duty-bound to put the interests of the enterprise ahead of his own.

Modern Corporation Marketing
At the same time General Motors also revolutionized automotive marketing by creating multiple brands of cars, each with its own identity targeted at a specific economic bracket of American customers. The company set the prices for each of these brands from lowest to highest (Chevrolet, Pontiac, Oldsmobile, Buick and Cadillac.) Within each brand there were several models at different price points.

The idea was to keep customers coming back to General Motors over time to upgrade to a better brand as they became wealthier. Finally, GM created the notion of perpetual demand within brands by continually obsoleting their own products yearly with new models rolled out every year. (Think of the iPod family and its yearly new models.)

When Sloan took over as president of GM in 1923, Ford and its Model T was the dominant player in the U.S. auto market with 60% of the U.S. car market. General Motors had 20%. By 1931, with the combination of superior financial management and a astute brand and product line strategy, GM had 43% market share to Ford’s 20% – a lead it never relinquished.

Thanks for the History Lesson – So What?
Well thanks for the history lesson but why should you care?

If you’re an entrepreneur you might be interested to know that when Sloan took over General Motors in 1923, it was already a $700 million dollar company (about $8.5 billion in sales in today’s dollars.)

Yet you never hear who built that company. Who founded what would become General Motors 16 years earlier in 1904? Where are the charitable foundations, business schools and hospitals named after him? What happened to him? Who was he?

The founder of what became General Motors was William (Billy) Durant. His board (led by the DuPont family) tossed him out of General Motors (for the second time) in 1920 when GM sales were $567 million (about $6 billion in today’s dollars.)

William Durant died managing a bowling alley in Flint Michigan in 1947.

From the day Durant was fired in 1920, and for the next half a century, American commerce would be led by an army of “Sloan-style managers.”

But the spirit of Billy Durant would rise again.

This time in what would become Silicon Valley.