Ron Spears: we control the network, because the technology we have put in place is ours

Ron Spears sits across the conference table and sips a glass of water as he explains how AT&T has learned from its mistakes over the years in providing services around the world.
Except he's not across the conference table from Global Telecoms Business. Spears, the president and CEO of AT&T Business Solutions, is in New Jersey and GTB is in AT&T's London office, just down the road from Westminster Abbey and Big Ben. We're talking through AT&T's telepresence service — a stunning demonstration of the services that are now available thanks to international broadband communications.
But back to those mistakes — Spears himself raises the subject and he's happy to talk about them: they were joint ventures with a variety of companies, and they ended uncomfortably.
It didn't work, he says. "The decision we made in 2002 was the only way we were going to become a legitimate provider of global networking for multinationals was to own and have control of that global network ourselves."
The decision gave AT&T the chance to build a global state-of-the-art network, using MPLS technology. "We deployed MPLS technology outside the US before we fully deployed it inside the US. That was driven by the fact that when that [last] joint venture broke up we had to get a network in place to handle our customers."
Was this a complete replacement with AT&T's new infrastructure, or were there places where there were partners providing the connections?
"Only on the last mile," says Spears. "We were able to deploy all of our own switching equipment, in facilities that we had control of and we put badged employees in those facilities."

Glut of fibre

It was in a way a good time to be building a new network, he concedes with a smile across the intercontinental telepresence network. Back in 2002-03, there was a glut of fibre around the world, and that enabled AT&T to connect up its many points of presence.
Connections were leased at first. "You start with a lease. As there is more and more demand for capacity, you move to an IRU," he says: an indefeasible right of use gives AT&T access to the fibre as if it were its own. "And eventually you look at a route and decide if you need to build your own capability, just because the economics make sense. That's the way we have deployed capacity outside the US today."
But whatever the precise ownership status of the fibre, "we control the network, because the technology we have put in place is ours", he says. "We believe that we provide the only seamless MPLS IP infrastructure in the world. We believe that other companies that are providing network services to multinationals are still having to do that with a multiplicity of technologies and networks. And we don't."
That's the first challenge, he says: "Get your infrastructure right, because at the end of the day what's occurred since 2003, as we have moved clients onto IP infrastructure, that requires seamless connectivity. It requires consistency, as it relates to the class of service you provide, and we have spent the last four or five years getting that done."
In 2006 AT&T took the next step, recalls Spears. "We took a hard look at our strategy and determined that what our real focus was going to be."

Controversial decision

Essentially, AT&T decided to concentrate on a certain group of global customers. It was, he admits, a controversial decision at the time. "We took a lot of criticism for it", because the company "backed away" from a larger group of customers that it had picked up around the world over the years through a number of different relationships and ventures.
"Some were very local, some regional, very few global," he says. "They are very expensive to maintain and try to make profitable."
The company narrowed its focus "to about 1,800 global clients" and over the past three years, since the decision was taken, "that number hasn't changed much though the players have changed, either through mergers or acquisition", he says. The number is still around the same, "of which 700 are US multinationals and 1,100 are multinationals outside the US".
It's not true that the company focuses on the US, he points out, though he accepts that "those 1,800 do a significant amount of business in the US". That "gives us the opportunity to take advantage of the huge infrastructure that we have in place in the US", he says: "It is a huge differentiation for a global network."
The company "has been very disciplined about staying on that strategy", says Spears. "Three years into it, it has proven its worth."
There has been negligible churn among those 1,800, he says: "Customer don't go anywhere else." The reasons for any change are mergers and acquisitions, or "some folks have emerged as global players that we hadn't seen before". Some of them previously didn't have much business in the US, "but started to expand there".

Acquisition mode

AT&T really hasn't lost international customers in that time, he insists. "Outside of the US, we have really been in acquisition mode." He lists some of the customers the company has picked up over the past three years: a rollcall of famous names, including Royal Dutch Shell, Volvo, Smiths, Siemens and others. Of course, Siemens is not what it what it was, and some of that business has moved away as the company has turned into a set of regional businesses.
But apart from such cases, "today we win that business rather than lose that business", says Spears.
Of course, AT&T has a low market share outside the US, so it is less vulnerable to competitors winning business away. "In most of the world, we're on the front end of the acquisition curve of large multinationals."
And that was the intent of the strategy when it was formulated a few years ago: to be very focussed on a low number customers "and when those customers' global networks came up for bids, we wanted to be well positioned to participate and compete for that business".
So he believes that the company's track record in that "has been very successful at winning those big global network bids".
Spears contradicts the assertion, repeated by many players in this market, that customers can assemble their networks from a variety of different carriers. Indeed, with MPLS "it really does not work economically", he says.
"Technically it can work, but when you start trying to merge MPLS networks", there are problems: first, "carriers don't do that: we don't peer with other MPLS networks, so if a customer wants to do that, they have to put the infrastructure in place themselves".

Network complexity

This means that customers look at the complexity of putting together an MPLS distribution network, "they more and more are looking at a single provider to do that, either globally or in some cases regionally". They can then decide to manage the regional networks themselves, he says.
"I find more and more that even the regional strategy tends to be in the minority", as customers are preferring a global approach, he says. "It's all about consistency. When you start talking about quality of service metrics on an MPLS network, I may do it one way, BT may do it another way, Telefónica may do it another way, Verizon may do it another way. You try to put those networks together and take on the responsibility to make sure that the QoS is consistent — that is not trivial, nor is it inexpensive."
So customers are more and more putting at least their global backbone into the hands of a single provider, says Spears. "That is what's different about the MPLS technology that is driving IP infrastructure today." It's different from the approach for frame relay or ATM years ago, when "you really could bolt networks together yourself", he says: "Today you really can't do that economically."
The technology is complex, he says, so more customers are coming to the conclusion that the complexity is such "that they have to make some decisions about what they're going to do well and what they're not going to do well".
It was relatively common 15 years ago for large corporations to manage their networks themselves, acquiring capacity in different parts of the world as needed and linking them together.
At the other end of the spectrum, many large companies now will hand over virtually all responsibility for networks to a provider such as AT&T. "That is the General Motors model that we're involved in today, the IBM model, the Royal Dutch Shell model, the Boeing model," he says, listing more of the company's customers. This is where AT&T sees itself in the market.

IT services

But there's something beyond that spectrum that he's illustrated: the attempt — by BT Global Services, notably — to offer IT-related services to multinational customers. That's an area that AT&T has not tackled — so what are the arguments for and against.
There's a sceptical noise delivered via telepresence over AT&T's transatlantic optical fibre: "I think BT's results are probably the answer to that question," says Spears.
But when AT&T approaches a multinational client with a proposal to deliver services, where does it draw the line and why?
"Network," says Spears simply. "Network. We are networking experts. We have deep and broad knowledge of how to build, install and manage networks, and so when you see us talk about services that we provide — and we provide a broad scope of services — they are network-based services. I stop, basically, at the data centre door."
There are companies that are good at providing IT services inside a data centre, he adds. "That is not our expertise. Nor will we make it our expertise."
He was first asked that question three years ago, he adds, when someone told him: "You guys are really behind. You clearly don't get it. You've really narrowed your scope in terms of the number of customers and the amount of work that you're going to do."
But AT&T rejected that analysis. "Our position was that we couldn't make the math work," says Spears. "You've got to be very careful about getting into markets that feel like they're adjacent, but really aren't. They're not adjacent to your expertise, or to where you've made your investment."

Owner's economics

Three years ago the pressure was also coming from the other side, he recalls, from systems integrators, which would do the IT side and include the network. "Customers have found out that that doesn't play out well over seven to ten years. They fall behind on technology. They need someone that gets owner's economics out of a network."
So, today, outsourcing has split into three, he says: a data centre stack, a networking stack, and an end-user computing stack. "You can bit where you want to bid. We tend to focus on the networking stack, and we get owner's economics out of that stack. Customers have figured out that this approach is the way to do it. It's become the generally accepted practice today."
But there is an IT area that AT&T does offer: application management in association with our global data centres. "We expanded that last year" into a sort of cloud computing and storage offer.
"We are in the collaboration space," he adds: the telepresence system he is using for the interview is "the ultimate high end of collaboration, but we are fitting it into a broader collaboration platform that we're deploying".
AT&T also offers services in security and in content distribution, and he sees a future where customers can work with mobile devices on its network. "More and more companies are going to move applications to these networks and more and more companies are going to want to allow employees or partners to access that information over these high-speed mobility networks with the very smart devices that have come into the market in the last 24 months — like the iPhone."

Cloud computing

Companies will turn "bits and pieces" of their applications — such as hosted Exchange — over to AT&T, he says, "where it's practical and not strategic". Companies will not hand over their data centres to AT&T and tell the operator "to put it all in the cloud", he says.
But they will hand over subsets of the data centre infrastructure. Over time, it's an economic argument, says Spears. "Economics." Over time there are things a network-based provider can do more economically.
It is moving towards providing IT, but in a different way, isn't it? "Using network infrastructure," says Spears. And that's the important difference. "That is the discipline that we are showing as we evolve into services." GTB