Hemispheres Magazine

My Smartest Mistake: Ignoring the customer

Michael Dell, Chairman, Dell Computer Corporation

One of the most common problems in our industry is that you often have inventions looking for problems, rather than the other way around. Without customer input, the result is a misguided project that has nothing to do with real customer demand. Dell has always been a company that takes pride in—and direction from—customer feedback; the direct nature of our relationship with our customers, more than anything else, has been the key to our success. But when we started working on the Olympic project in 1988, we forgot about creating technology for the customer’s sake. In creating technology for technology’s sake, we consequently fell right into the very trap we’d been so proud of avoiding.

The Olympic family of products was what technology people call a “boil the ocean” product: It spanned a wide variety of desktop, work station and even server markets, with single and dual processors and many, many different operating systems, and it proposed to do pretty much everything. This was the kind of project that put you on the map.

It was an incredibly ambitious plan for a company that was doing about $250 million a year in sales, but it seemed like a good idea at the time. In our four years in operation, we had had a lot of success developing more and more complex products. Olympic was the largest and most complex project we’d ever undertaken, but that made it seem all the more enticing, especially since it would give us a leadership product covering a broad range of markets and would allow us to continue to grow.

Still, the stakes were big. Olympic consumed perhaps half of our R&D budget and involved lots of technology that we’d never had much experience with in the past. We hired a tremendous number of bright, young engineers. Because it was a very big deal and was so exciting, a lot of people had an emotional fervor about the project and for the product.

What we didn’t realize was that we were working on bleeding-edge, rather than leading-edge technology.

With great enthusiasm, we started introducing prototypes of what we thought were Olympic’s terrific features to our customers. They weren’t terribly impressed. “Some things are compelling,” they said. “But the whole product isn’t compelling enough.”

We could barely believe what we were hearing. The truth is, we didn’t want to believe it, so at first we denied it. We proceeded to prepare prototypes to show at the annual Comdex trade show in November 1989, and, with some fanfare, unveiled them to a broad range of customers. Some were enthralled by the idea but the majority of them said, “Thanks very much, but we don’t need that much technology.”

We knew that technologically speaking, this product line made sense. There were ideas within it that made for great inventions; in fact, the graphics and disk technology were later incorporated into highly successful products. But elegant technology alone just wasn’t enough to convince people to buy something they didn’t want. Facts are facts, and that fact was pretty overwhelming.

I often say that mistakes are learning opportunities. Olympic was a learning opportunity of, well, olympic proportions.

First of all, we learned to try to identify potential problems early—and fix them fast. We admitted we had made a mistake and basically shut the project down before any of the products ever saw the light of day.

Olympic taught us that big-bang, boil-the-ocean projects were quite dangerous. If one or more of the inventions didn’t turn out to be achievable, the whole project was at risk. Gradual, incremental improvements were a far better choice for us: They reduced our risk and allowed us to take advantage of fast technological transitions, so we could offer the best new components as they became available. In other words, we’d do better to try for lots of base hits rather than swinging for home runs and striking out.

But the most important lesson was to involve our customers early and often in the development process. We had gone ahead and created a product that was technology for technology’s sake, rather than technology for the customer’s sake. If we had consulted our customers first about what they needed—as we had been accustomed to doing—we could have saved ourselves a lot of time and aggravation.

Things changed as a result of the Olympic experience. For one thing, we started thinking and talking in terms of “relevant technology,” a term we used to describe the features that were important to our customers. All of our product plans since Olympic have been heavily driven and influenced by actual customer needs. We don’t pursue things that seem technically interesting or even technically elegant, but rather look to find things which create a lot of value for our customers.

For another, we consciously committed to a set of principles regarding “buy versus make.” At the time, most of the computer industry subscribed to a “we-have-to-develop-everything” view of the world. If you weren’t doing component assembly, you weren’t a real computer company. It was like a rite of passage. You somehow proved your manhood by placing small semiconductor chips on printed circuit boards. But we realized that while there were times it made sense to invent things ourselves, there were times when it made more sense to leave the R&D to our suppliers. Figuring out the difference helped us to focus on how best to use our engineers.

We were presented with a real challenge. We had hired a large cadre of wonderfully bright engineers, but we were a young company—we had been public for a little over a year—and were under the gun to be efficient. We couldn’t keep a lot of projects going that wouldn’t deliver a lot of value.

When we canceled Olympic, we could have easily said, “These people developed this thing that nobody wants to buy. Let’s just get rid of them and start over.” Instead, we concluded that our engineers were, in fact, as brilliant as we originally thought; they just hadn’t known what our customers wanted. If we could convince them to channel their energies into thinking about things that deliver value to the customer, as opposed to more classical engineering challenges, we trusted that they would create terrific products that our customers would love.

We encouraged our engineers to get to know the customers’ needs by spending time with sales representatives. We involved them in product planning so that they could be exposed to the logic of the decision-making process, and we tried to train them to think in terms of what their contribution meant to the overall business. Some resisted, saying, “I don’t want to do that. I just want to design silicon.” But some really thrived.

And so did the company.

In the six to nine months after we canceled Olympic, we announced in rapid succession the broadest and most robust product line yet in the company’s history. The launch included our first high-end floor-standing systems and advanced storage options, both of which incorporated technologies developed during the Olympic program. Dell’s business went from $250 million in 1989 to $2.1 billion three years later.

We’ve since formalized and institutionalized the process of gathering feedback from our customers. We know this is the grounding we need to ensure that every development effort we undertake is totally on target with customer needs and delivers a lot of value to them.

Some people look at us and say, “Wow, everything goes right for this company.” But it was mistakes like this one that led us to the kind of excellence we have today.