Lucent: The Truth Hurts

In the latest Heavy Reading report – “Setting a Course to Convergence: The Incumbents' Wireline Strategies” – I looked at the way that the strategies and product portfolios for the “Big Six” incumbent equipment vendors match up with a common vision they seem to have of the emerging converged network. I chose to focus on what I think are the foundations of the converged network – the packet core and the packet edge. For all of the companies I spoke to (the six vendors, and the 50 or so service providers) the overwhelming consensus was that IP/MPLS is the “convergence technology,” although deployment timescales tended to differ, as you’d expect.

The result of all this effort is a 120-page report, and you can check the executive summary here. Unfortunately, in such an abstract there’s a need to summarize arguments that may take tens of pages to express. In this case my research and analysis of Lucent Technologies Inc. (NYSE: LU) was summed up in a single bullet:

  • Lucent is abandoning its own convergence products in favor of re-inventing itself as a service organization. It will fail.
Lucent was understandably upset at this statement and has challenged its accuracy. As the author of the report, while I might fret at the brevity and directness of the summarized bullet, I stand behind it, and am prepared to outline the research and analysis that led to this conclusion.

Before I do that, let me just clarify one thing. The report as a whole sprang out of the interviews with the six CTOs, and I do feel that one of the best and most interactive of these interviews was with Bill O’Shea, CTO of Lucent. However, the bullet conclusion had virtually nothing to do with this interview. Sure, Bill described Lucent’s vision eloquently and intelligently, but you’d expect that from a man with his experience and credentials. My subsequent research into the Lucent product line, and its recently announced partnerships, led me to conclude the three things we summarize in the bullet, namely:

    1. Lucent is closing down its own convergence product lines.
    2. Lucent is moving to a greater focus on services.
    3. Heavy Reading feels that their approach will “fail,” according to the definition I shall offer below.
Items 1 and 2 are just research. In other words they are essentially documented fact, and people like me just have to find them. Item three is analysis, which means it’s a conclusion I reached as a result of my research.

So let’s consider the documented facts behind Item 1. remembering that the products I’m looking at here are the packet core and edge devices. In Lucent’s case, these break down as follows:

Table 1: Convergence Stall-Out
Product Status Comment
CBX 500 Ascend acquired June 1999; no follow-on product in the intervening 4 years ATM edge product with IP/MPLS capability
GX 550 Ascend acquired June 1999; no follow-on product in the intervening 4 years ATM core product with IP/MPLS capability
TMX 880 Nexabit acquired in July 1999; renamed TMX 880 and launched in March 2002; product cancelled in October 2002 Designed as core IP router, extended to include MPLS and ATM functions; at launch, product was positioned as �upgrade� solution for Lucent ATM customers
SpringTide SpringTide acquired in September 2000; product cancelled November 2002 IP/MPLS services router

Lucent spokespeople argue that they're not “abandoning their own convergence products” because they are still making and shipping the old Ascend boxes. Well, they have definitely “abandoned” the other two, and two out of four isn’t too bad, but let me ask our readers this: If Lucent is really committed to these products...

    1. Why has no follow-on product been launched since these products were acquired (for around $20 billion in stock, incidentally)? After all, the GX550 is supposedly a core switch, yet only has a nominal capacity of 32 Gbit/s! Equivalent core ATM switches from Alcatel SA (NYSE: ALA; Paris: CGEP:PA), Cisco Systems Inc. (Nasdaq: CSCO), Marconi Corp. plc (Nasdaq: MRCIY; London: MONI), and Nortel Networks Corp. (NYSE/Toronto: NT) can have more than ten times this capacity.

    2. When Lucent bought these products they were market leaders (with a 37 percent market share, according to Synergy Research Group Inc.). Today they are Number 3, behind Nortel and Cisco with 23.7% share.
But wait! At the launch of the TMX 880, the product was positioned as Lucent’s next-generation core multiservice switch. So the TMX was the follow-on product to the GX 550... except they cancelled it.

If Lucent had followed through on a real next-generation GX 550 I’m pretty sure they’d still be in first place, because the kind of people who buy GX 550s like ATM. This leads me to my conclusions concerning Lucent’s relationship with Juniper Networks Inc. (Nasdaq: JNPR).

Now, as you all know, pretty well everyone (apart from Cisco) sells Juniper routers, and darn fine routers they are, too. But does the term “over-distributed” spring to mind? Lucent potentially has a great differentiator in selling Juniper because they’re putting Navis signaling onto the Juniper platforms so that Lucent ATM customers can interconnect their “ATM islands” over a Juniper core. This sounds great, but if the kind of people who bought CBX 500s and GX 550s like ATM, how comfortable are they with the idea of putting a bunch of IP/MPLS boxes in their core networks? Time will tell, but let’s face it: How many of these “strategic alliances” can you name that were actually successful?

Moving onto Item 2, that Lucent is moving to place a greater emphasis on services. Well, this is just documented fact. You can go to the Lucent Website and download Pat Russo’s presentation at the Sanford Bernstein Strategic Decisions Conference and check out Slide 7: “Over time, we intend to transform Lucent into the leading ‘network integrator’ for service providers.”

Finally, Item 3: We think Lucent will fail. Now I did feel this was an emotive word, but then again I’m a bit of a boring old network techie and don’t get out as much as I used to. So let me just rephrase it: With their chosen strategy, we don’t think Lucent will succeed.

I’ve explained why I think the Juniper relationship may not pan out as well as Lucent hopes (by the way, it’s a huge win-win for Juniper because they get introduced to all of Lucent’s major accounts). And remember that, for Lucent, those Juniper boxes represent the foundation of the converged network vision. Whatever they slap on the top, the traffic eventually has to pass over some pretty tired Ascend boxes, or some non-ATM Juniper boxes.

Despite my negative take on the Lucent-Juniper partnership, I think we’ll see more of the same from Lucent. In other words, the company will begin to sign more and more reseller deals to replace the products it no longer manufactures. And in each of these deals Lucent will be “adding value” in some niche way in order to exploit its installed base in that area. While this seems like a great idea on paper, I say again that historically these relationships have never really worked. What Lucent will do is simply expose its customer base to a more nimble and opportunistic company’s sales force.

In terms of deriving more revenues from services... Well, Lucent will have to get to the back of a long line. Most of the other incumbents, and a bunch of real service companies – like Electronic Data Systems Corp. (EDS) and IBM Corp. (NYSE: IBM) – started this a while back.

The problem I see for Lucent deriving much more revenue from this approach is that much of their service organization works in the lower value-add levels of the service business. Sure, there are a few great business consultants and network designers in the Lucent fold, but there are a lot more cable pullers. Don’t get me wrong: Pulling cables is a noble task (I’ll be doing it at home this weekend), but it doesn’t bring in the big bucks. To retrain all those folks is going to take a lot of money (which cuts into the bottom line somewhat), and even after they have their CCIE certificates, that doesn’t make them an expert. Pity, then that Lucent sold off International Network Services (acquired October 1999, sold August 2002), a company that boasts “over 450 highly qualified network consultants who collectively posses the industry’s largest inventory of multi-vendor internetworking skills...” They might have come in quite handy right now.

As an analyst it’s easy to use hindsight to criticize Lucent’s strategy. They’ve clearly made a series of mistakes over the past few years – but so have all of their competitors. The problem for Lucent is that their competitors have made fewer mistakes or have at least been able to correct them more effectively. It’s customary for analysts to offer some glib advice to the people who run the companies they analyze – this is usually quite easy, as we’re not the ones who have to carry out that advice! But in Lucent’s case there’s no obvious solution to the set of problems with which they’re faced.

The only conclusion I can reach is that over the next couple of years, Lucent will continue to do less well than its major competitors. It will lose market share in the segments in which it continues to operate. Faced with a diminishing revenue base, the company will have to keep focusing on cost reduction, at a time when its competitors will probably be enjoying increasing revenues and may even be growing their employee base. The most difficult challenge for Lucent’s highly paid executive team will be to manage the expectations for the company as it migrates from manufacturing into reselling and service – in effect, as the company becomes a “manufacturer of diminished importance.”

I wish them luck in that endeavour, and I truly hope that I’m wrong about my assumptions on reseller deals, for the sake of the 36,500 people who still have jobs at Lucent, and the 240,000 former Lucent employees who depend on the company for their promised retirement and healthcare benefits.

So, to sum up. Yes, the one-bullet summary was brutal, but, you know, sometimes the truth hurts.

— Geoff Bennett, Chief Technologist, Heavy Reading

Be the first to post a comment regarding this story.
Sign In