Skip to Content

Invention Is a Flower, Innovation Is a Weed

The inventor of Ethernet and founder of 3Com shares some lessons with young innovators.
November 1, 1999

Bob Metcalfe’s career traces the trajectory of innovation. He started in the academy, as an undergraduate at MIT and a graduate student at Harvard. In his doctoral dissertation he laid the theoretical foundations for a novel method of boosting the power of personal computers: network them. At Xerox PARC, he turned that theory into something called Ethernet. Xerox wasn’t particularly successful at exploiting Ethernet commercially, so Metcalfe decided to try himself, founding 3Com to do the job. After many incarnations at 3Com, he cashed in his chips and became, in his words, a “technology pundit,” who writes a column for InfoWorld, organizes some of the information world’s best conferences, and sits on the board of Technology Review. TR asked Metcalfe to tell us what he learned as he followed the trajectory of innovation from the lab bench to the boardroom and beyond.


Why should you listen to me about innovation? Maybe you shouldn’t. (Especially if what you need is gentle encouragement.)

True, I lived for eight years in Boston’s Route 128 high-tech innovation zone, back when it was working. True, I lived and prospered for 22 years in Silicon Valley. True, I invented Ethernet, a computer networking technology that now connects more than 100 million computers to the Internet. True, 20 years ago I founded 3Com Corporation, which now does more than $5 billion in annual sales. And true, my personal fortune is a significant fraction of a milliGates.

But Silicon Valley-style high-tech entrepreneurship is certainly not the only way to innovate. It’s just that, right off, I can’t think of any others.


Before sharing a few lessons I’ve learned from inventing and innovating, I’d best disclaim a bit. Consider the fact that today we have computers fast enough to compute the trajectory of a thrown rock in real time. If you wanted to gather the equations to compute the rock’s trajectory, the last thing you would do is interview the rock.

Most successful entrepreneurs I’ve met have no idea about the reasons for their success. They were thrown-like rocks. I had the good fortune to be thrown into Silicon Valley. My trajectory was a mystery to me then, and only a little less so now.

Another disclaimer: I’m a sample of one. My experience is not statistically significant. So you’re going to have to read a lot of lessons learned by many different innovators before you can put together something that holds up. And even after you do that, keep in mind that after 40 years of tennis, I can tell you to get your racket back early, move your feet, and step into the ball. But then you’re going to have to spend a lot of time on the court practicing before you can put it all together and beat me.

Enough disclaimers. Here are some lessons I think I’ve learned.

1. Selling Matters

I have a six-story townhouse in Boston overlooking MIT on the Charles River. I often invite young engineers and would-be entrepreneurs over to schmooze. Many of them tell me my townhouse is beautiful and they hope to invent something like Ethernet that will get them such a house.

The picture they have in their heads is of me lounging around on the beanbag chairs in a conference room at Xerox PARC in 1973. They see me having this idea for a computer network and submitting it as an invention proposal to Xerox. Then they envision me putting my feet up and letting the royalties roll in until I have enough to come up with the down payment on the townhouse with the river view.

My picture-the actual picture-is different. It’s the picture of innovation rather than invention, the weed instead of the flower. In my picture it’s the dead of winter and I am in the dark in a Ramada Inn in Schenectady, New York. A telephone is ringing with my wake-up call at 6 a.m., which is 3 a.m. in California, where I flew in from last night. I don’t know yet where I am, or where that damn ringing is coming from, but within the hour I’ll be in front of hostile strangers selling them on me, my company, and its strange products, which they have no idea they need.

If I persist, selling like this for 10 years, and I do it better and better each time, and I build a team to do everything else better and better each time, then I get the townhouse. Not because of any flowery flash of genius in some academic hothouse.

Most engineers don’t understand that selling matters. They think that on the food chain of life, salespeople are below green slime. They don’t understand that nothing happens until something gets sold. The way I think about it is that there are three sets of people in the world. There is the set of people who will buy your products no matter what (think of your mother). There’s the set who will never buy your products (think of your competitors). Both are much smaller than the set of people who will buy your products if the products are competently sold to them. That vast middle set is why sales is so important, and it represents one of the key differences between invention, which comes up with a brilliant new idea, and innovation, which gets that inspiration out into the world.

Sales may not matter in invention, but it matters-in a very big way-in innovation.

2. At a Startup, Jobs Grow Faster Than People

In 1982 I suddenly lost my job as 3Com’s CEO and became our vice president of sales and marketing. Take this as a measure of our desperation. I knew we needed feet on the street, people to actually ask our customers for orders. I didn’t have too many choices-it wasn’t a very big company at the time-so I started looking over the available candidates.

Dave didn’t initially look very promising, since he wasn’t in sales. (He was a production engineer.) But he was single, his dad was a salesman, and he could move that week, so I assigned him the entire eastern United States as his 3Com sales territory. This is one of history’s bad decisions with a good outcome.

Since his background wasn’t in sales, Dave wasn’t a good bet to succeed in the huge job I gave him. But he was smart and energetic and he learned fast. Pretty soon, orders started doubling. The volume of sales was improving so quickly that I “promoted” Dave-from head of the entire eastern United States to head of the northeastern region. Again, he succeeded, and I was able to “promote” him again, this time by giving him the Washington D.C. metro area. After another strong run of successes, Dave went up the down staircase again by taking over all private label sales in Washington.

If you counted from the top, as people tend to do in big companies, Dave was getting demoted each time. If you counted from the bottom, each reduction in territory in our rapidly growing company was a big promotion with more responsibility and higher compensation. This is an example of how in small successful startups the jobs grow faster than the people, not the other way around, the way they do in big companies.

3. Don’t Hire-Recruit

Lesson #2, that jobs grow faster than people in successful startups, implies something very important about the people you bring into such a startup.

First, don’t “hire” anyone. B people hire C people-they collect resumes and choose the person they want to honor with a job. A people recruit A people. The people you need for a growing startup already hold jobs much bigger than the ones you need to fill. You have to recruit them, beg them almost, to take the small jobs you’re offering. Those with imagination will see that the company has the potential to grow so quickly that the small job will soon be much bigger than the stable position they hold at a big company. Forget about big company notions of performance. A people can perform easily 10 times better than B people, sometimes 100 times, or 1,000. The worst thing you can do is rush to fill a job with a B or C person. That could be very costly, perhaps even fatal, to your company. Wait until you can recruit the A person who can see the future and grow with it.

4. People Have Operating Ranges

The fact that jobs are growing so explosively at startups has other important consequences. Consider operating ranges. Everybody has one.

From $0 to $1 million per month I ran sales and brought in the orders, mostly through personal selling. After that, sales became too technical for that approach to work. Mike took over and carried the company from $1 million to $3 million per month. After a brief stall, Chuck took us to $5 million per month. John took over from Chuck to get us to $25 million per month. Then Bob took the company to $400 million per month. In each of these cases, our head of sales succeeded within his operating range. After that, a ruthless change had to be made to bring the company to its next plateau.

How do you know when it’s time for a change? How can you tell when the person who did such a great job six months ago has hit the upper limit of his or her operating range? The first sign is a decline in performance-salespeople missing quotas, engineers slipping schedules. At first it looks like the plans were too ambitious; then it’s everybody else’s fault. At some point, unless things start improving, sometimes even before the proof is conclusive, changes must be made. You have to be able to say, “If you can’t do it, we’ll just have to find someone who can.” If you wait too long for the person to learn what they need to know or for conclusive proof of whose fault it is, you may bring the whole enterprise down. Better to risk the lawsuit for wrongful discharge and save the venture.

5. Don’t Listen to Your Customers

In 1982, 3Com Corporation was the sole supplier of Ethernet cards to a startup called Sun Microsystems. These cards were for Multibus-compatible computers, and so internally we called them MEs (really). Sun and its competitors were buying MEs at ever-increasing rates, and they wanted us to make a cheaper, faster next-generation card. Our salespeople were right there with our big customers: They demanded we work on what internally we called the ME II (naturally referred to internally as the “me too”).

We knew, however, that Sun was planning to design Ethernet connections into their computers, and that Intel, designer of the Multibus, was planning a competitive ME. Despite overwhelming customer demand, we decided not to develop the ME II. Some of our salespeople quit in disgust, because we were “not listening to our customers.”

They were right; we weren’t. Instead, we invested the time of our excellent engineers in designing an Ethernet card, called internally the IE, externally the EtherLink, for the new IBM Personal Computer. Today, there are no Multibus computers left, but 3Com ships more than 20 million EtherLinks per year.

The lesson? Well, of course I’m taking some license by saying that the lesson is you shouldn’t listen to your customers. The real lesson is that you have to choose which customers to listen to very carefully. And, even then, you cannot necessarily give them what they say they want. You have to develop products that your customers will need by the time you are able to deliver them. If you don’t, when the development cycle is finished, and you’re ready to ship, you will be offering what the customer said he wanted last year. And any salesman will tell you it’s easier to give a customer what he needs now than to sell him something you insist he said he wanted last year.

6. Plan for Success

But if you do look ahead successfully and see what the market wants, you will only create new problems. During 1983, for example, 3Com was in the tornado for exactly this reason: We had ignored our customers and salespeople and guessed right about what the market would want-Ethernet for personal computers. In the third quarter, we grew 85 percent in three months, which almost killed the company. We sold much more than we could deliver, leaving a backlog of unsatisfied customers for our competitors to unhook. We didn’t have enough people to answer all our customer support calls. Our production programs slipped behind schedule.

I remember our CEO, Bill, saying we would never grow that fast again. But I pointed out to Bill that Compaq had the year before gone from zero to 10 times our size in one year. Our problem was not that we had grown 85 percent in three months. It was that we had planned to grow only 15 percent. Which is to say that you can plan too conservatively. (I hasten to add that if you have to err on one side or the other, it is better to plan conservatively. Just don’t overdo it.)

7. Be an Entrepreneur, Not a Visionary

In 1982 my board of directors starting calling me a visionary, and I ate it up. Next thing I knew, I wasn’t CEO anymore. Turns out, nobody wants visionaries running companies. At my level of the game, being called a visionary was faint praise.

Here’s the difference between a visionary and an entrepreneur. Both have visions, which are a dime a dozen. But an entrepreneur has, in addition to visions, plans. In addition to plans, actions. You might have heard that 80 percent of winning is just showing up. Well, showing up is an action, like taking that wake-up call in the dark in the Ramada Inn in Schenectady. I wouldn’t touch a visionary with a 10-foot pole.

Being a company’s proud founder is also foolish. As they build their companies, many people walk around saying to their employees, “I am the founder and you’re not.” You want every employee of your company to be a founder. To have and therefore feel ownership.

8. Know Your Own Operating Range

In 1990, I “retired” from 3Com exactly 11 years after incorporating it. I should have gone three years earlier. Twice the board of directors chose someone other than me to be CEO of 3Com. In 1982, when I didn’t flounce out the door, but instead got sales and marketing. And again in 1990, when I retired amicably.

Putting together the board of 3Com was one of my proudest accomplishments. I built that board with the best people I could find. When they decided someone else was better qualified to be 3Com’s CEO, who was I to argue? Both times, in retrospect, they were right. By 1990, 3Com had outgrown me. Fortunately I had a board smart enough to know that I had succeeded in moving 3Com out of my own operating range. You should be so lucky.

Deep Dive


Capitalizing on machine learning with collaborative, structured enterprise tooling teams

Machine learning advances require an evolution of processes, tooling, and operations.

The race to destroy PFAS, the forever chemicals 

Scientists are showing these damaging compounds can be beat.

How scientists are being squeezed to take sides in the conflict between Israel and Palestine

Tensions over the war are flaring on social media—with real-life ramifications.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.