On the Edge: An Interview with Akamai's George Conrades

by Nicholas G. Carr

Akamai has its origins in a set of mathematical algorithms developed by a group of professors and graduate students at the Massachusetts Institute of Technology (MIT). Their work, and the company built around it, make it possible to store Web content closer to customers, wherever they are, and deliver it via the least congested available route. (See What Akamai Does for a brief explanation of how it works.)

"Better technology attracts more content," he told Harvard Business Review's Nicholas Carr. "More content attracts more networks. More networks means better technology. And around we go."

Q: When a customer agrees to let you store and deliver its content, it's entrusting you with the heart of its online business—which in many cases is its entire business. How do you, as a third party, build that kind of trust with companies, particularly ones that are used to controlling their own content?

A: A number of capabilities are required to meet that challenge, and we've organized ourselves around those capabilities. First is technical know-how. Everyone who has contact with a customer, particularly in the early stages of a relationship, has to be comfortable explaining what our technology does and how it can solve a customer's problems. In the first meeting with a potential client, our salespeople explain the technology in depth. Usually at that point, the customers' eyes open up. They react the way I did when I was first introduced to Akamai's technology: "Wow, what a great idea. If that works, that's great."

Then you have to convince them that it does indeed work. That's where our technical consulting staff comes in. They assign an integration team to each customer, and the team's job is to gain a deep understanding of the customer's technical requirements and then match our services to them. One of the things they do that's essential for gaining trust is to actually show the customer how things work through a trial. They'll say, "Let's get your stuff up on the network, and then let's monitor what happens." We use independent measurement companies to demonstrate that their content is delivered faster and more reliably using Akamai's technology. And we compare the performance with what they would have achieved if they used a competing technology. We openly benchmark all the time. We love it.

So the performance of the technology itself, demonstrated in real-world conditions, establishes trust. Then the challenge becomes maintaining that trust, and that's where two sets of people come in: our account management people and our customer care people. Account management is responsible for the business relationship. They're the people who say to a new customer, "We're the ones you can call around the clock, 24/7. You got a billing problem? You got a new application? We'll be here, and we'll work with you on it. We'll do what needs to be done."

Customer care people are responsible for solving any technical problems. They take the day-to-day phone calls whenever a customer has a question or needs some help.

There are two things that are distinctive in the way we manage customer care. First is that even though customer care resides within our sales and support unit, it's tightly wedded to our research and development unit. In fact, we often rotate R&D people through customer care. That creates two big benefits. It brings deep technical knowledge into customer care, which enables us to answer our customers' questions quickly and confidently. And it exposes our R&D people directly to customers, so they can draw on that customer knowledge to refine our technology.

The second distinctive thing about our customer care is that we treat every call as a high priority. Most customer call centers employ a form of triage—an escalation process. A call comes in, and it's answered by a fairly low-level employee who has minimal expertise. If it's a simple question, that employee may be able to answer it. If it's more complicated, the caller gets bumped up to a second level to an employee who has a slightly higher level of knowledge. If that person can't answer the question, the caller gets bumped up to the top level, where he finally gets to speak to a real expert.

That kind of customer service system may make sense if your goal is to minimize costs. It's a lot cheaper, after all, to hire a novice than an expert. But from the customer's point of view, it just makes it harder and more time-consuming to get the information you need. You have to wait for a callback, or you have to make another call yourself when you realize that the information you received was wrong or insufficient.

We looked at that conventional way of handling customer inquiries, and we immediately said, "No way." We need to maintain trust, and we need to build our brand—and in the Internet space, you don't do that by making the customer wait for an answer. On the Internet, a technical problem is a business problem. So everyone in our customer care unit is an expert, and they all share a simple objective: to answer every question on the first call or the first e-mail. Period. It doesn't matter whether the question concerns something in our network or on the customer's Web site or in some third-party application; it's our responsibility to get the right answer and to get it fast. It's a complex world out there, and the one thing we never want to say is, "Well, that's not our problem. We don't do that." We say, "Whatever it is, we'll help." In fact, we like questions that aren't about our network because they help us learn more about the customer's needs and experiences—all that information is incredibly valuable to our R&D efforts.

I can't say enough about customer care, because it really is the place where we earn the customer's trust every minute of the day. We've got people who just eat, sleep, and drink customer care. They love it. They love the challenge of solving customers' problems. And they love the technical side of it; they cherish the links to R&D. You know, I sometimes review the e-mails that go back and forth between customers and Akamai, and it's just amazing. You'll see e-mails with a time stamp of three in the morning that are just totally engaged with solving the customer's problem. It never stops.

And remember, all those interactions are captured in our database, stored in our institutional memory. So the next time the customer calls, or another customer calls with a similar problem, the information is all there.

Q: Given Akamai's academic and scientific origins, was it hard to get the organization focused on customer service?

A: Not at all. And I have to admit, that came as something of a surprise to me. My background is in marketing and sales, so when I arrived here, I kind of assumed that technologists wouldn't have a strong focus on customers. I'm pleased to say I was dead wrong. Just as a commitment to communication was built into this company's culture from the get-go, so was an intense customer focus. Five years ago, when Tom Leighton and Danny Lewin were developing the algorithms at the core of our technology, one of the first things they did was go out and talk to the big content providers. They sat down with the top experts at Yahoo! and CNN.com, for example, and said, "We've got this idea about pushing servers to the edge of the Internet, and we even think we could coordinate those servers using mathematics, without any central point of control." And of course, the initial reaction was, "Yeah, sure." So Tom and Danny said, "Well, just pretend we could. What would your criteria be for using that service?" The content providers then went through their lists of must-haves—we need to see every hit, we have to be certain we never serve stale data, and so on—and those requirements were built into the technology from the start.

From that moment forward, Akamai has been customer focused. Every step of the way, the customer has called the shots. It's one of the great surprises and great joys of my career to see such highly technical people care so much about customers. I mean, it's really been a blessing to me as CEO. I didn't have to do a bunch of reengineering to get everybody to understand the importance of customers. It was already part of the culture.

Excerpted with permission from the article "On the Edge: An Interview with Akamai's George Conrades," Harvard Business Review, Vol. 78. No. 3. May-June 2000.

[ Order the full article ]

Nicholas G. Carr is an executive editor at HBR.