Bob Metcalfe: Co-inventor of Ethernet. Tech industry legend and Polymath
Episode Summary
Bob Metcalfe has lived in the future since the late 1950s. And lucky for the rest of us, he has brought us along for the ride. In this interview, I talk to him about the origins of networked computing as well as the birth of "Metcalfe's Law." Listen here.
Transcript
Bob Metcalfe:
So, you have to buy more of our products and we promise it will be valuable. The question arises at that point when we said that, were we lying? And I claim no. And the reason I claim no is because I had a time machine. Xerox PARC was my time machine.
Mike Maples Jr.:
That's the voice of Bob Metcalfe who helped pioneer the internet starting in 1970. He co-invented ethernet, co-founded 3Com Corporation and formulated Metcalfe's Law, which describes the exponential forces that drive the power of network effects. Bob's journey as an academic entrepreneur and builder is the amazing story of a time traveler who saw the future and brought us all along for the ride. This is Mike Maples Jr. of Floodgate, and it's go time with Bob Metcalfe.
Mike Maples Jr.:
Welcome to Starting Greatness, a podcast dedicated to ambitious founders who want to go from nothing to awesome super fast. When you're a startup founder, you have to channel your inner James Bond, your MacGyver, your Wonder Woman. I'm going to help you win by curating the lessons of the super performers, but before they were successful. So, without further ado...
Ignition sequence, start.
Mike Maples Jr.:
Let's get started. In my mind, Bob Metcalfe is one of the most colorful and legendary people in the tech industry. He was at the ground floor of Xerox PARC, which was where desktop computers, the mouse, the graphical user interface, laser printing, and the ethernet all took roots. He also founded 3Com, which brought the ethernet out of the lab and into the real world. It's been a career highlight to get to know Bob personally, and I'm excited for others to get the chance to connect with him. Let's do this.
Mike Maples Jr.:
Bob Metcalfe, welcome to the podcast.
Bob Metcalfe:
Thank you, Mike.
Mike Maples Jr.:
Thanks for coming. So, you've been at this for a little while now. So, when did you first get intrigued by computers or technology? When was that?
Bob Metcalfe:
Well, my father didn't go to college, but he was a technician, an aerospace technician. His specialty was gyroscopic platforms, and he had a workshop in the basement and it was full of toys. Then he helped me build a train set using the toys. And then in eighth grade, Mr. Gaucher, my science teacher, encouraged me to build a computer using the parts that we had assembled for the train set.
Mike Maples Jr.:
Now, what year is this?
Bob Metcalfe:
1959.
Mike Maples Jr.:
So, what does it mean to build a computer in 1959?
Bob Metcalfe:
Well, I was afraid you'd ask. It was an adding machine.
Mike Maples Jr.:
Okay.
Bob Metcalfe:
It added any number between one, two and three to any other number between one, two and three and displayed it by lighting one of the lights between two and six, and my teacher called it a computer, but it was the grade he gave me the change in my life. He gave me an A+++, H superior.
Mike Maples Jr.:
Okay.
Bob Metcalfe:
And before all that, in fourth grade, I had to do a book report one night and I hadn't read any books, and my dad had some books in his shop and I grabbed one of them and it was an electrical engineering textbook written by an MIT professor. So, I wrote a book report. This book had its high points and its low points, but on average, it was average. And then I ended the book report sort of sucking up to my fourth grade teacher saying, "And someday, I intend to go to MIT and get a degree in electrical engineering," which I did.
Mike Maples Jr.:
And so what started to get you interested in computer networking when you were at MIT?
Bob Metcalfe:
I was interested in computers already when I arrived at MIT, but networking wasn't until I graduated and my friend, Pits, and I, may he rest in peace, we both went to MIT looking for a job. We were in grad school, but looking for a job. And MIT had two projects, [inaudible 00:03:57] conductor memory and this ARPANET internet thing. It was a network for connecting computers together, and the job there was to connect some computers together. ARPA, the Advanced Research Projects Agency's Department of Defense was the principle sponsor of computer science in those days. I mean, tens of millions of dollars. And they decided that they were tired of buying a PDP-10, an interactive timesharing system, for each of their research sites. So, they had the idea of doing resource sharing where you could, from anywhere, log into one of those computers and develop your software over there so they wouldn't have to buy you a PDP-10.
Mike Maples Jr.:
So, you're working on this, you're still at MIT or have you headed over to Harvard yet?
Bob Metcalfe:
Well, I was a Harvard... I became a Harvard student in the fall of '69. This is before I became the MIT guy on the ARPANET. I went to Harvard and they said, "The promising area for graduate study is the ARPANET. So, if you want to be a grad student here," which I did and so I entered the PhD program, "You should work on the ARPANET." So, I said, "Well, I just graduated from MIT and I know I took all the courses in digital electronics. Harvard, I can build your connection to the ARPANET." We called it the ARPANET then.
Mike Maples Jr.:
Great idea.
Bob Metcalfe:
No, they didn't. Harvard said, "No, no, this is too important for grad student." So, I turned around, I walked back to MIT and they hired me. They gave me a job as a member of the research staff to build a connection between MIT and the ARPANET.
Bob Metcalfe:
I got MIT to agree that I could build two of them and give the second one to Harvard. Harvard refused. So, I built it. It was a high speed network interface between a PDP-6 and a PDP-10 to connect it to the packet switch, the ARPANET packet switch number six. That was on the ninth floor of Tech Square at MIT.
Mike Maples Jr.:
Okay. So, some people are probably... A lot of people listen to this podcast know exactly what packet switching is, but some might not. What was the breakthrough idea with packet switching?
Bob Metcalfe:
Packet switching comes from queuing theory. A packet is a bundle of bits with a destination address on the front and typically a source address and some error checking and a few protocol bits inside. So, when you have something to send, you put it in the packet and you put the address on the front and you launch it.
Bob Metcalfe:
So, the idea was to share these PDP-10s among all the research, but how to connect them together and the architecture will be packet switching, and that came from the queuing theory that said if you're carrying traffic with the ratio of peak to average is high, high peak to average, then you want to do statistical multiplexing. You don't want to dedicate bandwidth to any one connection. You want to share it among many and be able to give it all to one or not, depending on whether they were currently at their peak of their average or none. There was an efficient way of multiplexing traffic.
Mike Maples Jr.:
And what was the effect of being able to do that? What could you do with computers by virtue of doing that that you couldn't do before?
Bob Metcalfe:
The big invention was email.
Mike Maples Jr.:
Really?
Bob Metcalfe:
So, there was hardly any email. I mean, there was email within a time sharing system, but no one was sending email across the United States in those days. But this ARPANET thing, once it got connected, it was easy to write a few email clients and start sending emails around. The pioneering work was done at Bolt Beranek and Newman by Ray Tomlinson. He's the guy who invented "At" sign on the addressing schemes. So, collaboration among researchers got ramped up because we could talk to each other through email.
Mike Maples Jr.:
So, why Xerox PARC? Why go West to Xerox PARC?
Bob Metcalfe:
The main answer to that question is The Beach Boys, because I wish they could all be California girls and the surf and the... So, there's this thing that you want to go to California. A few MIT people had been at Xerox research ahead of me. Jerry Elkind was a MIT guy who went there and he's the one who suggested to me I look at going there. So, I went on the PhD speaking tour. I got nine job offers because the ARPANET was... It was the hot stuff. So, if you were a grad, a new PhD, working on the ARPANET, you were valuable properties.
Mike Maples Jr.:
It would be like being at the cutting edge of data science today or something like that.
Bob Metcalfe:
Just like that. And of course, my number one choice was to be a professor at MIT and I did not get that offer, but I did get the offer from Xerox, which by the way was twice the money, no fundraising, no classes to teach, do whatever the hell you want.
Mike Maples Jr.:
No real politics.
Bob Metcalfe:
But I'm telling you, had MIT offered me the assistant professorship, I would have taken it.
Mike Maples Jr.:
You might be there today.
Bob Metcalfe:
I might be there today, but I didn't. Instead, we decided to go to Palo Alto.
Mike Maples Jr.:
Probably not everybody knows what Xerox PARC was all about and what it was like at the time. So, what was up with Xerox PARC in the early seventies?
Bob Metcalfe:
Xerox was a monopoly making copiers and it was a hot stuff company with huge gross margins, and so they could afford to have a Palo Alto research center. The goal of the computer science department was to work on the office of the future, the paperless office of the future. Of course, we immediately built the world's first laser printer and produced more paper than ever before. And a lot of good personnel decisions were made and the leadership was very cool, and they recruited from the ARPA research community.
Mike Maples Jr.:
Yeah. So, my recollection of it was that Xerox PARC was basically the intellectual center of gravity of where a lot of modern computers as we know them exist; the mouse, the bitmap display, wizzy wig applications, local area networks. I guess you were living in the future when you were at Xerox PARC. Did you realize that at the time or were you just cranking on stuff that was interesting to you?
Bob Metcalfe:
Well, no, we knew it. We had some general notion of the... I don't know if we called it the time machine. Maybe I've just done that since then, but we knew that you had to assume things were going to change in a certain way and act as if they had already changed. So, a computer on every desk probably wouldn't fly at 30K per computer, but we knew that eventually... I guess we knew everything was going to get cheaper, so we pretended. So, we put these $30,000 PCs, one each... Ethernet, as I remember, cost a thousand dollars. Funny money, research funny money you'd put in to get an ethernet card in your PC, and then started writing software applications. And it was good. I mean, people saw the mouse and they saw the bitmap display and they saw the network.
Bob Metcalfe:
People would say, "Why should I spend a thousand dollars for this card when I can just take a diskette and carry it over to the printer and print my document?" And the answer was, "No, just hit..." And I think we invented Command P. You're looking at your document, you want it printed, you hit Command P and then it would send it over the ethernet to the printer. This was a cool printer. It was a page per second, 500 dots per inch. So, there was this gorgeous printer that everybody wanted to print on, and there was no other way to print on it other than sending a file over the ethernet.
Mike Maples Jr.:
So, you were at Xerox PARC for a while.
Bob Metcalfe:
Eight years.
Mike Maples Jr.:
Yeah. So, did you start 3Com because... What was it? I mean, starting a company is a pretty brave and crazy and impossible thing to do.
Bob Metcalfe:
So, I gave Xerox seven months notice. I wrote a memo and in it, I said I would like to leave to pursue entrepreneurial ambitions, but had no clue as to what it would be. And then I left Xerox in January of '79 and started consulting. And in February of that year, I went to Gordon Bell's office at Digital Equipment Corporation, then the second largest computer company, and Gordon admired ethernet. So, he asked me to design a network like ethernet for DEC, and I told Gordon I couldn't do it. One, I just felt a certain loyalty to Xerox, and two is I had just finished designing the best network I knew how, so his network wouldn't be half as good. Why don't we propose to work with Xerox? So, I drafted a letter saying, one, why don't we work together so that our computers could talk to your printers?
Bob Metcalfe:
So, they decided to start working together. And then I found Intel Corporation. I found them at the National Bureau of Standards, and Intel wanted to make chips to sell the DEC and to sell to Xerox. Xerox wanted to sell printers, DAC wanted to sell computers, and then we discovered that we couldn't meet because we would be in violation of antitrust law. So, I called up Howard Charney, who was my fraternity brother at MIT, who I knew was a lawyer, and he'd been involved in a lawsuit against IBM on antitrust. So, I call Howey and I asked him, "Gee, how can I get these meetings started again?" And he gave me a list of five things. [inaudible 00:13:07] remembered all five of them. "You have to do these five things and you can meet," and one of them is no marketing people because they will conspire on prices. You must have as a goal an open standard.
Bob Metcalfe:
So, out of that, we decided to make ethernet a standard. So, we created a standards body that went to the IEEE, the Institute, Electrical and Electronics Engineers and created IEE Project 802 for the purposes of standardizing ethernet as a local area network standard. And then the companies started meeting because they could now, and they wrote a spec, the blue book ethernet spec, which was finished in September of 1980. And on that same day, I had the first draft of my business plan.
Mike Maples Jr.:
And what did 3Com stand for?
Bob Metcalfe:
Computer Communication Compatibility. The idea is we would use industry standards to interconnect computers. What you probably don't remember is that in those days, every computer vendor had their own networking system.
Mike Maples Jr.:
Sure.
Bob Metcalfe:
And so you couldn't connect a DEC to an IBM to a Wang, to a Data General.
Mike Maples Jr.:
So, speaking of a blue, where's IBM in this?
Bob Metcalfe:
So, Project Data 2 got created, the blue book spec I'd submitted in September of 1980. And in 81', IBM showed up and said, "No, no, no, no, no, no. We have a network that's better than ethernet. Ethernet is just junk and fragile and non-deterministic." It had all sorts of names they threw at it. "It can't possibly work, but we have a thing called the IBM Token Ring and we'd like to submit that." By the way, General Motors did the same thing. So, pretty soon, we had a flock of local area networks including, those three, the IBM Token Ring and the General Motors token bus and ethernet. And then we fought for two years in committee and then the IEEE did a courageous thing. It decided to standardize all three of them. 802.3 became ethernet, 802.4 Became the token bus, and 802.5 became the IBM Token Ring.
Mike Maples Jr.:
And I remember inside of IBM, I'd have summer jobs there because my dad worked there, and they had their own language to describe everything. A hard drive was a Winchester Disk or they called it DASD for Data Access Storage Device. Obviously, everything was going to be on Token Ring, but it was kind of like its own country, almost, of standards and...
Bob Metcalfe:
So, IBM invited me twice, once to Research Triangle Park with the communications division and once to the office systems division, which I think was in Franklin Lakes, New Jersey, twice, to go pitch them on ethernet and whether they should adopt it or not. And I failed twice because I approached the whole thing stupidly. One of the stupid things I did is they were so proud of their new network architecture, which they called SNA, Systems Network Architecture. And in my talk, I insisted on calling it "Sna," and they did not... I could see them squirm every time I said, "Sna." They thought I was saying it accidentally.
Mike Maples Jr.:
Okay.
Bob Metcalfe:
I was saying it just to make fun of them. That's all.
Mike Maples Jr.:
Yeah. That was just an awesome way to sell is to make fun of it.
Bob Metcalfe:
Yeah, that's exactly right. If you want to sell, you make... But since I had been working on local area networks for six years by the time I got to those two, I won the argument because they were a bunch of IBMers who were all 32-70 dumb terminal batch processing people. They didn't know what a LAN was, Local Area Network. So, I won the argument but I did not get the order. Had I to do it over again, I would not use the "Sna" joke again.
Mike Maples Jr.:
But then when you started 3Com, you were also recruited by a pretty famous guy by the name of Steve jobs. What were your interactions with him like?
Bob Metcalfe:
So, I left Xerox in January of '79 and I got an apartment in Boston. So, I had an apartment in Palo Alto and an apartment in Boston. I was consulting. And one lonely night, I remember I was feeling lonely that night because I was sitting alone in this Boston apartment in the dark and my phone rang. I picked it up and it was a guy named Steve calling from a company called Apple, and I had not heard of Steve and I had not heard of Apple. I had not even heard of a city called Cupertino. But Steve, of course, immediately flattered me and said, "By the way, I've heard you're a real networking genius and I'd love to talk to you about your helping me network my products. I make personal computers." I said, "Tell you what, I'll come visit you in Cupertino next week."
Bob Metcalfe:
And so we went to an organic food place on Stevens Creek Boulevard, the two of us. I had started 3Com two weeks before, so I was determined to sell Steve a network and he was determined to recruit me. But eventually, it became clear to him that I had started my own company and was determined to stick with it. So, then something that I understand in retrospect was unusual, Steve did not try to kill me. He didn't hate me, didn't disparage me. He helped. He said, "Well, there's a guy you got to meet," and he introduced me to Regis McKenna, who was his advertising and PR guy, the reigning genius of Silicon Valley. And he was my personal publicists for two years because Steve asked him to, and then Steve introduced me to Pagemill Partners, which was the Angel Group.
Bob Metcalfe:
So, he introduced me to that crowd and they were later the lead investors in RA round. So, Steve was very helpful. We were lightly friends. We double dated for awhile with respective lady friends. He came to my wedding in 1980. By the way, the trouble with having Steve Jobs come to your wedding is no one remembers anything about our wedding except Steve jobs was there. He would absorb all the oxygen in the room. Fabulous guy. So, he was very helpful in starting my company and I'm forever grateful. And he could be an asshole, but I don't care. I like him anyway.
Mike Maples Jr.:
Were you ever exposed to his reality distortion field or all the stuff that people talk about or were you more just talking about just getting stuff done together and he was just being helpful?
Bob Metcalfe:
No. He And Butler Lampson taught me something at different times. They taught me that if I lose an argument, I am not obligated to change my mind because the two of them could convince you to do anything because they were so smart, so fast, so much factory call, just raw intelligence. You would fall into their reality distortion field and you would lose the argument every single time, and what I noticed in both cases, they won the argument whether they were right or not. So, Steve was like that. He really did have a reality distortion field, and part of it was intimidation. He'd make you feel that if you disagree with him ,that you were an idiot and you were going to die soon, and Bill Gates has the same quality, which is they both create this field in which you are worried that something bad's going to happen to you if you do not agree with them.
Mike Maples Jr.:
But then when did you raise venture capital for 3Com and what was that like?
Bob Metcalfe:
On September 30th, 1980, Xerox, DEC, and Intel issued the blue book specification for ethernet and submitted it to the IEEE. And on that day, I had the first draft of my business plan. And by the way, we didn't have PowerPoint. I later was on the board of the company that developed PowerPoint and we sold it to Microsoft in 1987 for $14-million. This is 1979, and so I had a spiral bound... I had a spiral binder. So, each day, I would bind the new version of my deck and then I would take my deck as paper and thumb through the spiral bound, but I had it ready on the same day as the spec, the theory being that the spec was sufficient evidence that there was going to be a standard ethernet that my company could support.
Mike Maples Jr.:
And the VCs at the time say, "Well, if it's open, well, why can't Xerox do it? Why can't DEC do it? Why can't all these other big companies do it?"
Bob Metcalfe:
That's right. There were patents. There was an ethernet patent. And by the way, the IEEE made Xerox put that patent essentially into the public domain in order to make ethernet a standard. So, there was no patent protection for ethernet, and the various smartest MBAs who inhabit VC firms all pointed this out to me as if I had never noticed it before. And I said, "Well, we already have the products. We had taken on Exxon as a customer and they paid us $750,000 to develop our first four products, which were then working." So, we were raising money to go manufacture and sell these standard products. So, we had a headstart. And we knew exactly what to do and we were really fast. Those are the answers that I gave.
Mike Maples Jr.:
Was it difficult to negotiate with VCs or were they pretty much like, "Okay, Bob Metcalfe, Mr. Ethernet, let's rock and roll?"
Bob Metcalfe:
No, it wasn't, "Let's rock and roll." So, I had the headstart, I had the 750K, I had the products, and I was giving a talk at... I was a professor at Stanford. I had been for seven or eight years. And I was giving a talk on ethernet and there was this guy sitting in the back and it was Wally Davis, Wally Davis who was with Mayfield, and he was the first one who said, had you thought of starting a company? I said, yeah, absolutely. And by then, I had met every venture capitalist on earth because my office was on 3000 Stanton Hill Road. We had lunch at the Sundeck and eventually Jack Melcor, who is running an angel group called the Pagemill Partners, he was the lead investor. Dick Kramlich of NEA and Wally Davis of Mayfield were the... We raised $1.1 million. So, the lead put in 400K and the other two put in 300K and then there was another guy, Fred Frank, who was an investment banker. He put in 100K.
Mike Maples Jr.:
But then looking back on that, Bob, had you ever... Up until the point that you were doing 3Com, had you ever had to really sell before?
Bob Metcalfe:
Only in a personal sense, get a job type selling. But no, I never sold a product. But at 3Com in 1982, I suddenly became the vice president of sales and marketing of 3Com Corporation and there was no sales force. So, I recruited six people and we went out to sell and we learned... The first thing we learned about selling was that you have to ask for the order. So, as soon as we had salespeople, we actually started getting orders. That was a breakthrough, but then I quickly became a salesman. I learned how to sell.
Mike Maples Jr.:
How did you get better at sales?
Bob Metcalfe:
I went on a million sales calls. Oh, and I was... Once, I got some advice from a board member who said, "Listen, Bob, you shouldn't recruit to sales force because you have no idea what a good salesman looks like. You're an engineer. So, here's a guy who recruits a regional sales managers. You get him to help you." So, he helped me recruit six skilled salespeople, all many computer sales salespeople, and I learned selling from them because I went on sales call after sales call. And I had some competitors too, and they taught me how to sell because I remember the time I walked into this customer and the customer had a list of questions for me. The list had been prepared by my competitor and I went, "Oh, I see. I see how that works." So, I went from zero to a million a month.
Mike Maples Jr.:
Okay. Wow.
Bob Metcalfe:
In two years.
Mike Maples Jr.:
Not bad.
Bob Metcalfe:
And then we went public.
Mike Maples Jr.:
And who was the CEO at the time?
Bob Metcalfe:
Bill Krause.
Mike Maples Jr.:
Okay. And what did you learn from... So, first of all, why weren't you the CEO? Why did they bring in Bill Krause? What's up with that?
Bob Metcalfe:
I had never run a company before and so I believe I agreed. In fact, in my pitch, in my sales pitch to the venture capitalists... I guess that would be selling, wouldn't it? One of the tenants, I said that it's more important to me that my company be successful than I run it.
Mike Maples Jr.:
Okay.
Bob Metcalfe:
And they liked that.
Mike Maples Jr.:
Did you believe it when you put that on a slide?
Bob Metcalfe:
Yep. I heard it a million times and I got it and you could see it in the valley. It's a standard... Even Steve Jobs. People think Steve Jobs was the CEO of Apple. Yeah, in 1996. He founded the company in '76. It took him 20 years to make CEO. He knew about adult supervision. I'm sure he didn't call it that. Starting at the very beginning with Mike Markkula and then there was Scotty... Oh, before- [crosstalk 00:26:20].
Mike Maples Jr.:
Scully? Oh, Mike Scott, yeah.
Bob Metcalfe:
Mike Scott.
Mike Maples Jr.:
Yep.
Bob Metcalfe:
And Markkula again and then Scully. So, Steve understood adult supervision and I did too. So we went out and recruited Bill, who was running the general systems division of HP. So, he became president and COO initially and I was chairman and CEO initially. That lasted a year and then he became CEO and I was chairman, and then that was the day I became... I was then chairman and VP of sales and marketing.
Mike Maples Jr.:
Now, I've heard you tell a story about Bill that was important in your learning about sales, but also just being an exec.
Bob Metcalfe:
Well, so Bill shows up and it was our practice to have an ops meeting every Monday morning at 10:00. It was Bill and it was 10:00 on Monday morning, so the ops meeting was now his. He was the chief operating officer his first day of work, and so he starts running the meeting and I'm intensely curious about what does Bill Krause know?
Mike Maples Jr.:
Like what's he going to be like in action?
Bob Metcalfe:
Yeah. What does he know that I don't know? Why am I not running this meeting? Why is he running the meeting? What does he know that I don't know?
Mike Maples Jr.:
Yeah.
Bob Metcalfe:
And then I'm looking at him across the room and he's running the meeting in a normal way and he's taking notes furiously. And I walked around behind Bill and watched what he was writing and the whole yellow pad had "DNT, DNT, DNT, DNT" written over and over. He wasn't writing notes, he was just writing DNT over and over again. So, after the meeting, I took him aside and I said, "What is this?" He said, "Bob, I tend to talk too much, so this is my way of reminding myself do not talk. DNT." So, that's when I began learning the secret of sales, which is listening, and Bill knew this secret and he taught it to me. And so you got to listen. And in order to listen, you can't be talking. You have to do not talk, DMT. So, that was my first lesson from Bill. He taught me a lot.
Mike Maples Jr.:
Interesting. Yeah, it's funny because it kind of goes back to sales, I guess, as well. A lot of times, I find that when people want to sell, they sort of want to be the hero of the discussion. And then over time, you realize that the person you're trying to sell to is the hero, and you're trying to help them go on a hero's journey in their story, convincing them that you're the smartest person in the room. You're like Yoda teaching Luke Skywalker the force, right? But it's like if you think that you're Luke, you screw up, right? Because you want to go start swinging the lightsaber all around and acting like you're the conquering hero.
Bob Metcalfe:
It's very straightforward. It's a sign of insecurity. If you're insecure, you're afraid that if you stop talking, the other person will take the conversation in a direction you don't want it to go. So, you use "Mms" and "Ahs" to extend your control of the channel and you keep talking and yakking, but that's not listening and that's not selling.
Mike Maples Jr.:
Did you have any near death experiences where you thought, "I don't think we're going to survive, I don't think this is going to work?" Any of those kind of down downswings of the rollercoaster.
Bob Metcalfe:
My favorite one was while we were promoting ethernet in the early days and IBM and General Motors on top of us, we sent an expedition to the big conference in Germany, the messa. In any case, the big German show, we were there and DEC was there and everybody was there and the show Daily came out with a headline right on the top and it said, "Ethernet [German 00:30:01] super flop," and my revenue at that time, which had reached 250K per month went directly to zero, and that's basically how Bill ended up CEO. And it was, in retrospect, a great decision.
Bob Metcalfe:
So, that was a near death experience. But then there were all... When IBM marched into the IEEE and said, "Here's the Token Ring." Ah. Suddenly, IBM was the enemy and it got really ugly. IBM paid professors to write papers proving that ethernet could not work.
Mike Maples Jr.:
It was a hoax.
Bob Metcalfe:
It's a hoax. The math showed it. Dave Bogs and I had filled Xerox Corporation with personal computers and connected by ethernets with routers and printers and everything and it worked just fine, but IBM paid professors. So, professors are writing papers on purpose to kill ethernet and they failed. Dave Bogs was fond of walking around saying ethernet works only in practice, not in theory.
Mike Maples Jr.:
And then conversely, was there a time where you're just like, "Oh my gosh, we have product market fit. This is going to happen?"
Bob Metcalfe:
Yes. So, this is a great story. It's a little different. We didn't have the term product market fit, so I can't squeeze it into there. But we were making these cards that you plug into an IBM PC and they put them on ethernet. In the engineering of that card, the leader was a guy who just passed away, Ron Crane, and Ron was over in his cubicle finishing the prototype that would be given to manufacturing. And we had announced the product and we had a road show all worked out. We had early orders for it, and Ron's sitting in his cubicle working on it. So, at the ops meeting, I was deputized to go find out what the hell Ron was doing because the manufacturing was dying to get started.
Bob Metcalfe:
So, I went and just quietly visited Ron over there. I'd recruited him from Stanford. He said he was working on the lightening arrester. There's no lightning arrestor in the spec. Why are you delaying our whole company and we're running out of money and everything is going and you're playing around with a lightning arrester? And Ron just looked me in the eyes, said, "I'll do it as fast as I can." So, he finishes putting the lightening arrester in the damn thing and he turns it over to the manufacturing and it's a huge success. We sold a thousand units to a big bank in New York city, a high rise, but they were smart. They bought 2,000, 1,000 from us and 1,000 from our arch competitor. So, they had, side by side, huge ethernets in this downtown wall street type. What do you think happened next?
Mike Maples Jr.:
Oh, there was a lightning.
Bob Metcalfe:
Lightning struck this building.
Mike Maples Jr.:
You got to be kidding.
Bob Metcalfe:
No. So, this was one of those, "Oh, we have product market fit" because all of our competitors' cards were fried and their whole network, it was burning. It was smelling from the lightning and ours just kept working.
Mike Maples Jr.:
Just kept going, huh?
Bob Metcalfe:
So, we got an order for another thousand of this card. These cars were 750 bucks each, by the way. So, Ron Crane, stubborn, super genius guy, may he rest in peace. He ignored the product requirements. He knew that there was a problem here and he wanted a lightening arrester in there and God damn it, he wouldn't release it to manufacturing until we let him do that.
Mike Maples Jr.:
And from there, did it just kind of mushroom from there where you just started getting more and more orders?
Bob Metcalfe:
Well, it was a tornado. So, we went from a hundreds of month... When you think about it, 750 bucks each a hundred a month is pretty good revenue, yeah. And steadily increased till we were shipping millions per month. That is actually one of bill Krause' early contributions. I call it Krause's Law. He came in and he said, "This ethernet thing at 750 bucks" or whatever it was like that, "Is way too expensive, Bob. We're going to have trouble selling it. So, I want us to project what ethernet will cost in 10 years." So, he made us draw an exponential declining curve over 10 years from $750 to a $100 and then he said, "Now I want you to annotate this curve with what you're going to do at each of those points to achieve that price," like custom chips, one chip, two chips, three chips, transceiver on board, dah, dah, dah. So, we did this and then he made it a slide in the sales presentation.
Bob Metcalfe:
So, when we would go out, we'd say, "Yeah, it's a standard and it costs a lot today, but we promise it's going to go down and down and down." So, that curve, in retrospect, I call Krause's Law.
Mike Maples Jr.:
And speaking of exponential curves, how did that Metcalfe's Law come about?
Bob Metcalfe:
So, this card we had, the Etherlink, first shipped in September of 1982, connected IBM personal computers together on an ethernet. The problem was there weren't any IBM personal computers. They weren't shipping until fall of '81, the IBM PC. So, there were really no IBM PCs. So, we started selling a trial kit, three node network. Three cards that you put into the option slots, a cable to connect the three machines together and a diskette with software on it. The software would allow you to share a disc.
Bob Metcalfe:
So, if you put a disc on one of the machines, the other two could access it. Share a printer, you put a printer in one of the machines, the other two could print also. And there was an email client, so you could send email to anyone on that three node network. So, we started selling this for $3,000 and we sold a bunch. And then the customers all came back and said, "Works just like you promised. It's just not useful." And I was head of sales and marketing, so that fell right in my lap. So, I went over to Stanford one night where they had some Alto personal computers and I typed up this 35-millimeter slide and, of course, take a picture of it with your camera, have it developed overnight, had six 35-millimeter slides made, gave one each to the sales force. And this slide basically said the cost of a network is linear in the number of nodes.
Bob Metcalfe:
So, every time you buy one of our cards, you add another node. So, the end nodes, but the number of possible connections goes up faster than that. In fact, it goes up as in squared because every node can talk to every other node, so that became a quadratic. And as you may know, quadratics often overtake linear. So, there was this point out there called the critical mass point where the value of the network expressed as roughly in squared exceeded the costs roughly in, and then this became part of our sales pitch and we basically taught our sales force to say the reason that your network is not valuable is that it's not big enough.
Mike Maples Jr.:
You haven't bought enough.
Bob Metcalfe:
So, you have to buy more of our products and we promise you'll see it will be valuable. The question arises at that point, when we said that, were we lying? And I claim no is because I had a time machine. Xerox PARC was my time machine. So, at Xerox PARC, I went 10 years into the future, we filled Xerox Corporation with LANs and PCs and printers and routers and all that stuff, and it was good. Everyone loved it. So, I knew that if the networks got big enough, they would be useful, and they were. So, we went public a year later and that slide then, I gave a copy of it to George Gilder in 1995 and he called that slide Metcalfe's Law, that the value of a network grows as the square of the number of users or attachments or whatever.
Mike Maples Jr.:
Interesting. And so there's going to be a lot of founders listening to this. What advice would you give to founders who have ambitions of starting something really great?
Bob Metcalfe:
Go for it. I mean, it's just enormously fun. My father used to say, "Do whatever you want. Just be the best in the world." And so somewhere in there, I got the idea I wanted to do great things and not just live my life. I wanted to do great things, but to please my parents, basically. And starting a company is a great thing. It can be a non-great thing, but if you choose the goal carefully, that is if you do something that's worthy of you, I'm talking to you people who want to start that do dating apps on Sixth Street in Austin on Thursday night. That's not a problem worthy of you. Do not start that company. But my company had a very high goal, which was to build the internet and that was well chosen and very cool.
Mike Maples Jr.:
Cool. Well, thanks, Bob. Thanks for taking the time.
Bob Metcalfe:
You're very welcome. It's been fun.
Mike Maples Jr.:
Thanks for listening to Starting Greatness. You can follow me on Twitter @M2JR, and please shoot me an email with any comments or questions to greatness@floodgate.com. I hope you'll subscribe on Apple, Spotify, or wherever you get your podcasts. And if you like the show, I'd be grateful if you could leave us a review on Apple Podcasts. Never let go of your inner power to do great things in whatever matters to you. And until we meet again, remember, greatness is a decision.