O'Hara: We're interested in acquisitions in the US and
Europe. We're very cautious. We have a significant plateful of
work to integrate the companies that we have acquired. ... When
you do five deals in a couple of quarters, people notice. It's
not just people who are interested in selling their business to
Level 3 but people who are looking for a liquidity event for
Rafuse: We've been trained to think that
entertainment should come in 30 or 60 minute blocks.
... Now people are saying two minutes is pretty good.
Content used to go to the TV set but now it's going to your PC
and increasingly your mobile. It's every piece of content to
all three of those screens
Level 3 has survived, which is something you can't say about a
lot of companies in this business. And it's changed.
There are two outstanding characteristics, says chief
operating officer Kevin O'Hara: "Better operating performance
that anybody we compete with — that's operating
performance being measured in terms of margin. For instance,
gross margin and ebitda margin are better than anybody we
There were questions about the company a few years ago. It
had a lot of debt, but he uses the charming phrase "growing
into the balance sheet" to explain what's happened. What,
precisely, does that mean? "We had to generate enough cash from
operations to cover the full debt burden. I think those
questions are largely behind us as a result of both
improvements in the core business, organic growth, as well as a
whole series of acquisitions."
Level 3 has been "pretty opportunistic" in dealing with the
balance sheet "as the capital markets allowed", he says. "As a
result we're sitting here with a pretty favourable position.
The bond market has rallied strongly around the remaining bonds
we have outstanding."
And the company "has shown strong organic growth", he notes,
"and we have acquired five companies in the last 10
Has there been a strategy behind those acquisitions? They
fall into two categories, says O'Hara: "Backbone acquisitions,
which are really about consolidation and reaping the benefits
of putting two really similar businesses together and having an
improved financial profile as a result of the synergies."
This means it's possible to put the traffic of the acquired
company onto the Level 3, eliminating much of the duplication
in network costs. "While they tend not to be growth stories,
they tend to be very position from a cashflow standpoint."
The first of these was Wiltel Communications, "a company
very similar to us in the US, where we were able to put Wiltel
and Level 3 together". The revenue from such a deal "takes a
step function", but you don't see growth incremental growth
from it, he admits. However, "you see great incremental
cashflow from it" by combining the two businesses, he
Video distribution network
The Wiltel deal, which also brought Vyvx, a company that
specialises in video distribution, was valued at around $720
million and was completed in December 2005. Since then Level 3
has bought four metropolitan operators.
More recently, in late October and after the date of this
interview, it announced a deal to buy another competitor,
Broadwing, for $740 million plus shares. That deal will not be
closed until early in 2007.
O'Hara explains the purchase of the metropolitan networks.
"Historically Level 3 has constructed a fair amount of local
fibre along with its intercity fibre. That's true in Europe and
to a greater degree in the US. The last four deals were all
about acquiring local facilities."
Level 3 has had gross margins in the 70-75% range. "We've
been able to report those margins because we had superior
facilities. We had products and services to a targeted set of
customers. The combination really allowed for our great
The acquisition has allowed Level 3 to extend that
footprint, he says. The acquisitions "are all about growth".
There are some synergies as well, he notes, but the prime
motivator is growth into new markets.
"We now have 110 markets with fibre." The company has 5,300
buildings served with its fibre, "a great footprint to continue
Let's go back for the moment to Wiltel. Did that acquisition
give Level 3 extra network, or did it close down Wiltel's
infrastructure and move all traffic over? "Both," says O'Hara.
"There were some unique routes. There were some incremental
markets but there weren't a lot."
Level 3 has been decommissioning duplicated routes since the
deal. "We'll decommission more of those routes into the
And Wiltel had "an awful lot of business from SBC", which
merged with the old AT&T to create the new AT&T a year
ago. The old AT&T was a direct competitor of Wiltel.
"Roughly $1 billion of the $1.6 billion revenue was
attributable to SBC, which will all go to AT&T."
That loss of business "was all factored into the deal", says
O'Hara, but Level 3 is not able to decommission some routes
until the traffic has moved over to AT&T. "It doesn't make
sense to roll the traffic to the Level 3 route and then
decommission it. In some cases we're waiting for the traffic to
But that's revenue that the combined company, Level 3 with
the old Wiltel, will lose to the other combined company, SBC
with AT&T. "We've told the investment community to value
that as an annuity. We paid for a gross margin commitment.
We're going to meet that gross margin commitment — SBC
will. Assume it's going to zero over time. It shouldn't
Rationalisation has been different when it comes to staff of
the two companies. It's been said in the market that Wiltel
salaries were lower than Level 3 salaries, so that has affected
the way Level 3 has determined redundancies. Is that true?
"It depends," says O'Hara. A long pause: "We said at the
time of the acquisition that we thought Wiltel had some very
strong capabilities. They had some strengths in areas that
Level 3 historically did not. Two very specific areas are the
video broadcasting and distribution business that came with
Vyvx, where Level 3 had no capabilities. And they had very
strong voice capabilities, mainly because of what they had
stood up for SBC." Level 3 "was not a particularly strong
player" in traditional voice services, he adds.
The challenge was to put together "a company that had the
best of both organisations", he says. "That's what we said.
That's what we meant."
Of the reduction in staff — a mixture of
redundancies and natural attrition — since the deal,
"roughly 50% came from Level 3 people and 50% from legacy
Wiltel". Cost "was a factor but it was much more about the
right set of skills, the right capabilities — more so
than absolute cost".
Back to the metro networks that Level 3 has acquired.
They're exclusively in the US, says O'Hara. "There were four
companies, and the combined consideration for those four was
just under $2 billion. Without exception, they were very
healthy ebitda margins, generating cash and without exception
reporting strong organic growth."
Those acquisitions were for growth, of companies with solid
financial performance. "We could take further advantage of
Level 3 "is careful about what we purchase", he adds. The
company prefers "not to have to fix something" —
though clearly Wiltel, as it was losing the SBC revenue, would
have needed fixing if Level 3 hadn't bought it. "At the right
price it may be worth it. The worst of all worlds is somebody
that is broken that wants a high price."
There are "some companies out there that have some pretty
hefty valuations" and clearly he doesn't feel that they would
be worth it.
Some years ago Level 3 moved heavily into Europe, though
later partially withdrew. Is the company looking to build up
again outside the US?
"I'll give you my two favourite statistics about London,"
says O'Hara. "Five years ago we had approximately 1,800
employees and $20 million of revenue, and we were not
generating cash. This year we have 285 employees. We're going
to do close to $180 million revenue and we're generating cash.
I don't think we've pulled back. I think we started to treat it
as a proper business."
Those numbers apply to all of Level 3's European activities,
he explains. "We don't break it out by country. It's become a
very solid business for us. We got focussed."
The person in charge of this transformation in Europe is
Brady Rafuse, president and CEO of its European operations.
He's now been rewarded by being put in charge of Level 3's
activities in the content markets — covering media
companies, TV broadcasters and movie makers — in
addition to his European task. A former BT executive, who later
worked for its ill-fated Concert joint venture with AT&T,
Rafuse is now in charge of Level 3's Vyvx business and its IP
and collocation product lines globally.
European expansion plans
Level 3 has committed itself to expanding in Europe "to nine
new markets", says O'Hara. "We are extending our traditional
two rings up through the Nordic region —Oslo,
Copenhagen. We are extending further east." Leased facilities
to places such as Warsaw and Prague will be replaced with Level
3's own kit. "We have also approved the funding for a southern
ring and what we refer to as an Alpine ring for central
That's not new construction. "We're not digging up the
streets. We're getting third party dark fibre and putting our
own transmission kit on it, our own routing kit, and growing
our product set and the number of markets we can serve
There is a lot of fibre in places such as Paris, Brussels
and Frankfurt, says O'Hara, but that doesn't apply everywhere.
"If you want to go to Prague, Warsaw, Vienna, Milan, it gets
quite a bit tighter." There's not as much capacity available in
Rafuse adds that Spain is also short of fibre. "Spain's
really tight." There are a few providers "but they are all in
the same trench".
That doesn't meet Level 3's design criteria, adds Rafuse.
"We need to be diverse. We'll get there. We'll work it
Is Level 3 looking for other acquisitions? This interview
was just before Level 3's announcement in late October of its
plan to buy Broadwing, an exclusively US-based operator, so
clearly that was already being planned.
"We're interested in acquisitions in the US and Europe,"
says O'Hara. "We're very cautious. We have a significant
plateful of work to integrate the companies that we have
acquired. We are very conscious about meeting our integration
objectives, our financial objectives. I think the right
opportunity in North America or Europe would still be
Level 3 "is not interested in going beyond those two
markets" at the moment, because it wants "to leverage the
investments that we've already put in the ground". It has an
operating infrastructure and teams in place, "we have a network
that we can integrate on top of".
There are opportunities around, companies that are being
offered. "When you do five deals in a couple of quarters,
people notice," says O'Hara. "It's not just people who are
interested in selling their business to Level 3 but people who
are looking for a liquidity event for their company."
There's been "a tremendous amount" of that sort of activity
in the US, though Europe "is still a year or two years behind
the States in terms of investors looking for a liquidity
event", he says.
That implies that, in 2007 or 2008, Level 3 might well be
interested in acquisitions in Europe.
"The fact of the matter is that there are a lot of companies
that are not in the hands of strategic investors, through
either a bankruptcy process or a distress sale. There are
"non-natural holders" that have "ended up with a lot of
assets", he says. "They are looking for the right time to get
There "have been a lot of calls to a lot of different
companies, not just Level 3", says O'Hara. "But Europe is
lagging a little bit. Lagging is the wrong word: trailing."
Why the different timing? "The boom in Europe lagged by a
year or two. The correction lagged by a year or two. I think
the ultimate shake-out is lagging by a year or two."
Part of the reason is that "there is more financial
sponsorship in Europe for a number of the companies than there
was in the States", he adds.
In the US "a number of investors, for whatever reason,
decided that they'd had enough", he says. "In Europe some of
them haven't reached that same conclusion yet. Maybe they
won't. Certainly in the States it was pretty obvious that a
number of investors were going to look for an exit
What sort of companies does he expect to come onto the
market in Europe? "Europe isn't a market. You've got the
pan-European business and you've got a number of countries with
specific businesses and a number of local businesses."
Five or six years ago a number of incumbents had ambitions
to be dominant international communications companies. "They've
all pulled back and they've instead typically done a good job
in their home country and their aspirations outside of their
home country have waned a bit. That creates some
There are a number of local fibre providers, and
pan-European providers. "It's not obvious that over the long
run there is a way for all those companies to make money," he
says. "There will be opportunities in Europe. And Europe you
need to look at more locally."
Rafuse offers a local perspective, though with his new media
responsibilities he is increasingly based in the US.
The US metro acquisitions are individually different but
they have much in common. In Europe similar operations vary
widely, he reports: many of them are associated with local
utility companies and are often part-owned by the city
authorities. "It's just not homogenous."
Some cities have a single carrier, but others might have a
number of operators — though sometimes they share
Never buy what you wouldn't build
One rule that Level 3 applies, he says, is: "Never buy
anything that you wouldn't build." The business must be right.
"We don't tend to be opportunistic" and buy something "that
would be handy". It has to be a company "that our business
strategy supports", he says.
"Would we build it? If we would, is this an acquisition that
would help us achieve our goals quicker?"
We turn to Level 3's new enthusiasm for the content market.
O'Hara explains the background. "Wiltel had an interesting
subsidiary, Vyvx, which was very strong in fibre-based
broadcast signal origination and transmission." It wasn't
broadcasting, but was carrying the feeds for sports, news and
"It also had an ad distribution business," says O'Hara. "The
relationships it had with traditional media companies and
traditional advertising companies, and the technical
capabilities it had were very good."
Level 3 has not previously been in those markets, "but it
has a very strong set of IP products" and "an interesting set
of relationships with emerging companies".
Following the takeover of Wiltel, Rafuse has been given the
task of developing Level 3's strategy for this market, says
O'Hara, "to bring together both the traditional media and the
emerging media companies, and the capabilities of both Level 3
The company believes that "video over the network is going
to become much bigger than it is today", says O'Hara. "The Web
2.0 services you're seeing rolled out are going to continue to
grow at exponential rates. We're figuring out how we can make
the most of our capabilities and what else we should do in
order to attack that segment we've give to Brady to sort
As a result of the move Rafuse, a Londoner, is now regularly
commuting across the Atlantic. He's handed over direct
responsibility to a new managing director for Europe, Craig
Butrym, a former Genuity executive who came into the group
through Level 3's acquisition of Genuity in 2003: it's long
been one of the most acquisitive companies in the industry.
"He does that and I spend two or three weeks in the States
every month," says Rafuse. "I'm very well known on flight
BA219." That goes to Denver, Level 3's headquarters, but he's
spending an increasing amount of time in California, talking to
technology companies, and on the East Coast, where the
broadcasters are, and in Hollywood.
"It's a very exciting, interesting business. We're drawing a
line from broadcasting the SuperBowl in HD through to kids
falling off skateboards."
Wiltel was in charge of the SuperBowl feed "for the last 17
years", says O'Hara. "Showing the kids falling off skateboards
on YouTube: that's what Level 3 was doing."
Services for YouTube
YouTube is one of Level 3's big customers, says Rafuse. "It
took us five years in Europe to grow our IP business to 100
gigs. We think we're the biggest, but say we're only the fifth
biggest. That was five years: YouTube grew that in six months.
YouTube is one of the biggest."
Other Level 3 customers include Yahoo!, Google, Microsoft,
PhotoBucket and Flickr. There are application hosters such as
Windows Live. "I had an operating system update the other day
which was 128 megabits," says Rafuse. "Two years ago I'd have
been disappointed if it had arrived in a box and been 128
He continues the list of customers: Second Life, World of
Warcraft. "There's a lot of really interesting opportunity for
us there and we look at the customers on one axis, the
traditional media customers, the big search guys, Web 2.0, if
that isn't a bit passé. It's a business that is
generating an incredible amounts of traffic and our raison
d'être is always to bring scale to — for want of
a better term — existing problems."
And it's exciting. No longer are people talking of "cutting
a lot more people", he says. "You drive from San Francisco to
San Jose now and it's like it used to be. People say it was the
bubble, but if you drive down that road it doesn't feel like a
bubble, because Yahoo!'s still here, and Google's still here
and eBay's still here and Amazon's still here."
And behind that are new companies such as YouTube and
FaceBook, as well as MySpace, bought by News Corporation in
2005. "Everybody says about MySpace, what an incredible
purchase. Why would you spend that money? They made more money
from making Google the preferred search engine in MySpace than
it cost them to buy it," he says.
"Whether these numbers make sense, I don't know. What I do
know is that communities are important, and you're not seeing
the hub-and-spoke communications of the past."
More than half of content viewed by young people on the
internet now "is produced by someone that they know", says
Rafuse. "That's really different. When we went to FaceBook it
was very clear what they said. It was very clear they had a
community that was incredibly valuable."
Vyvx was "by far the strongest company" in its section of
the market, says O'Hara. "But Vyvx and the traditional carriers
weren't really looking at the MySpaces and the YouTubes. That
was a different set of companies — the IP-centric
But run the maths, he adds. "Unless you own your facilities
you're not going to be able to meet the needs of those
companies over time. From a Level 3 perspective it's going to
be hard to think on a continuum which are the traditional media
companies and which are the emerging media. As best we can
tell, you don't have anyone looking holistically at the media
space — and we thought we were the logical company.
Because of what Vyvx had and our IP expertise."
If many of the traditional telecoms operators are missing
the opportunity, what about the traditional broadcasters? Are
they spotting the way the market is changing and are they
starting to move into these new areas?
"They do," says O'Hara. "In America the broadcasters still
show the shows on particular day and at a particular time. Most
of the stations carrying those signals are affiliates of those
main broadcasters." That's the way the US television industry
works. "The affiliates make their revenue through advertising,
but the affiliates don't own the content."
Bypass the business model
But increasingly people want to watch at any time, not when
the local affiliate says. The broadcasters would like to put
the programmes live on a server, "but that would bypass the
The industry all has to make a transition, but until that
happens "increasingly broadcasters are putting their content on
the server the next day". And that means viewers can miss
programmes, knowing they'll be able to catch up with them later
"We're starting to see that transition occur. We're at the
very, very early stage of that."
That all has implications for the bandwidth. A standard
definition TV show requires "roughly 450 times the bandwidth of
a web page", says O'Hara. "If you want to watch it in high-def
that's roughly 2,700 times the bandwidth. There is no carrier
in the world that has the kind of capacity available for
network-based video distribution."
Look at the economics of the network providers. "We think we
have the most cost-effective network and the lowest incremental
cost. If anybody is going to do that cost-effectively we think
we have a better chance," says O'Hara.
Rafuse's challenge, he notes, is to talk to the content
owners, broadcast companies and media companies to find ways
that "Level 3 can play a role in their supply chain".
Some content will be distributed to strategic locations
around the market. "If you are going to see a high frequency of
hits", then that content will be "pushed to the edge" of the
network. There will be regional distribution points for
material that will be in moderate demand. "There is some stuff
you're going to get only occasional hits on and that you'll
keep in a more centralised way. So you're really going to see
pressure on all elements of the network — the last
mile, the regional and the core."
Data centre expansion
But this material is all going to be on huge servers that
are themselves distributed widely. Internet companies "are
talking of data centres that are literally millions of square
feet", says O'Hara. "The amount of bandwidth that you need to
connect those up is massive." Clearly companies such as Google,
Microsoft and Yahoo! are talking to Level 3 and its competitors
to ensure that networks are in the right place and have the
right capacity for these data centre plans.
A report in the New York Times in June 2006 suggested that
Google has 450,000 servers in at least 25 locations across the
world. Its capex for the first quarter of this year was $345
million — which one observer has suggested could buy
it at least 100,000 more servers. One of Google's biggest
projects is in a place called The Dalles in Oregon —
where there is cheap hydroelectric power to feed servers on a
Microsoft is reported to have 200,000 servers worldwide,
said the New York Times, and is expected to have 800,000 by
2011. "We're talking about huge numbers," says O'Hara. "And
this would be no value if it is not connected with huge amounts
It's happening in the US and across the world. "How do you
physically move information?" says O'Hara.
As a result Level 3 is spending hundreds of millions on
boosting its network capacity. It's no longer digging up roads,
but it's lighting dark fibres and, when necessary, pulling new
fibres through existing ducts.
To improve capacity it's introducing photonic switching,
using kit from Infinera, and is increasingly adopting ethernet
standards. "We pay a fraction of the cost charged by the big
router guys," says O'Hara.
The nature of content is different from what was expected,
says Rafuse. People thought the industry would be about TV
shows and movies, but now personal content — YouTube
and so on — is taking over. People are exchanging
video clips and sending links to YouTube — which is
getting over 100 million hits a day for links of an average
duration of about two and a half minutes.
"We've been trained to think that entertainment should come
in 30 or 60 minute blocks. Now people are saying two minutes is
pretty good." It's changing the whole nature of distribution of
TV, PC and mobile screens
"And it used to go to the TV set but now it's going to your
PC and increasingly your mobile," adds Rafuse. "It's every
piece of content to all three of those screens. That might be
direct, which is why we have the relationship with NBC or Fox
or whoever it happens to be. Or it might go through an
aggregator, such as YouTube or iTunes."
There are many different relationships. "I'd like to say
there's a very clear value chain but it's not like that."
Level 3 "didn't have the relationships that Vyvx has", he
says. "Vyvx has incredible relationships." With the
broadcasters and the big sports organisations, he means. "Those
relationships match ours with the big search guys, the Web 2.0
guys. It's a compelling proposition."
Are there similar trends in the rest of the world? "Personal
opinion, here," says Rafuse. "I think Mark Thompson is a
genius." He's been the director-general of the BBC, the biggest
broadcaster in the UK, since 2004, but previously headed the
successful independent operation Channel 4. "He's one of the
most enlightened broadcasters in the world."
The BBC has a very different model to what applies in the
US, "but I think Mark Thompson has a very clear vision of how
being a broadcaster in the 21st century is very different. He
gets a lot of this stuff."
O'Hara adds: "In many ways the BBC is ahead of the States,
the broadcasters there."
What is Level 3's view of those telcos that want to expand
to become media companies, to compete with cable TV services
and provide video on demand, rather than just making their
network available to other companies offering video.
"In the US Verizon and AT&T are the 800-pound gorillas,"
he says. "The cable TV operators are typically much stronger
than those in the rest of the world. They got ahead of the
telephone companies in broadband, and then started to put voice
over the top. As a result the consumer marketplace is a pitched
battle." The telcos had little alternative, he says.
"For the consumer, two companies are now competing for
services. It's better than where we were." Time will tell who
Telephone companies have typically been very vertically
integrated, he adds. "We have said from the beginning of Level
3 that the world is moving this quickly on so many different
planes, vertical integration tends to fall apart. We still
The business model is horizontal, he says, and that's how it
will continue — even though the telcos are trying to
rebuild a vertical model.
Rafuse is clearly dubious personally about integrated
packages: he recently moved house in the UK and turned down an
offer of six months free cable "because I didn't want the guys
to dig up my drive", he says. He kept his high-definition Sky
satellite service. "Someone could come free with television
services and I'd struggle to think that I'd pull Sky HD out of
my house. I'm happy with what I've got."
And he has "four or five voice clients on my PC —
the integrated package isn't something I'm interested in". That
doesn't make him a typical consumer.
"You're increasingly marketing to a market of one in this
world. For Verizon or BT maybe an IPTV offer is a great idea
but they're not going to sell it to me."