John Chambers, After the Deluge

Can you live with being unprofitable or break-even as long as you're entering a promising market?

No. This is one of the things we have to learn: You always focus on profitable contribution. When you enter new markets, by definition, the initial profit contribution won't be what you'll want to see later, but you've got to keep your culture focused on profit growth. This is what got the dotcoms in trouble. It will get successful companies, including Cisco, in trouble too if people make the mistake of focusing on market-share gain and not on understanding profitability.

That's the old Penn Central Railroad case that we were taught in business school. [Penn Central's misguided expansion plans helped push it into bankruptcy in 1970.] It's amazing how industries repeat themselves again and again. Anytime we move into a new industry, we had better have a plan for boosting it to our normal profit level, or else we shouldn't be there.

You've always made the point that Cisco doesn't just sell networking equipment -- it lives on the Internet. That's supposed to make Cisco extra quick and lean when it comes to everything from filing expense accounts to closing your financial books every night online. Sounds great. But how did you end up with excess inventories and an oversized payroll this past winter?

Well, first, in terms of the system, I'd evaluate its quality and performance as an A+. In hindsight, I wish I'd been a little bit more conservative in certain areas. In the first week of December, the biggest issue facing my customers -- and the biggest issue in their levels of satisfaction -- was that I just could not deliver product to them fast enough. So in an effort to meet our customers' expectations, we continued to increase our inventory and our supply capacities, and to keep up with our rising demand.

Then we had five days in the second week of December where orders were 10% off normal. All of a sudden, we were going from 65% growth to flat or negative growth. No company in history that I'm aware of has ever faced that quick a turndown.

By December 15, we had already met as a senior management team. We said that for the next 45 to 60 days, we would cut back on all hiring and refocus on discretionary spending. From that point onward, we've been managing with an eye on the state of the economy.

In April, we announced a more than $2 billion charge against excess inventory -- a charge that reflects the recent significant and unexpected drop in customer demand. Yes, we took a conscious risk with our run rate being as aggressive as it was. Would I take that risk again? Absolutely. Cisco won't change. We will always take risks. When you see us not taking risks, that's when you ought to tell your readers to sell.

Let's look at a risk that paid off for Cisco. You managed to grow your business in Japan from almost nothing in the early 1990s to nearly $1 billion a year today. And you did it with no help from the underlying Japanese economy. What did you do right, and what kind of lessons can other people draw from that?

We realized that if we went in alone -- as American or European companies traditionally did -- we'd get the same result, which is a much lower market share in Japan than anywhere else. So we linked up with 14 Japanese partners, with the full support of the Japanese government. That was our first move anywhere toward operating in an ecosystem. We learned that you acquire small companies; you partner with large companies. And then we earned the right to share our thinking with business and government leaders. We told them what they had to do differently. Even in areas where we had some healthy disagreement, we turned out to be pretty accurate. People might not follow your recommendations the first time, but they will listen the second time.

Finally, we didn't withdraw from the market during the slowdowns. We actually increased our commitment and used it as a chance to pull away. The faster our competitors withdrew, the more aggressive we got.

As Cisco has grown, you've redefined people's jobs in a very unusual way. Instead of letting managers build up ever-larger empires, you've constantly split up their duties, in what Cisco calls "divide and grow." Why do you do that?

It's one of the hardest things to grasp, but the more successful a group is, the more you ought to split it up. In sales, that allows a more focused approach to the customer. It forces you to cover each of the bases in order to achieve your overall goals. If you have multiple accounts, you can let opportunities slip away and still have what looks like a good year. But then you're leaving room to make your competitors strong. The same thing is true with R&D or other groups.