John Chamber’s epiphany – driven by a stock price swoon – that Cisco should stick to its knitting has another logical conclusion: dump the benighted Universal Computing System. But is the damage already so bad that Cisco is damned if it does and damned if it doesn’t?
Cutting the Flip camcorder loose is easy: fire the employees; dump the inventory; take the write-down. Is an angry Flip mob going to storm Tasman with torches and pitchforks?
But the case for dropping UCS is more compelling. First and foremost UCS margins are much lower than Cisco’s big switches.
Dress it up however you want and the result is the same: servers are lucky to get 30% margins, a trifle against 60-70% switch margins. The more successful UCS is the worse for Cisco’s margins and stock price.
For this we need a high-priced CEO? A 1st year MBA could figure this out.
Shane, come back
The bigger problem is the damage to Cisco’s reseller relationships. HP has invested billions to become a network vendor – would they stop now if Cisco canned UCS?
Of course they would. Not because of anything Apotheker says, but because if the HP sales force could sell Cisco instead of 3Com they would. HP could keep investing in 3Com and other network suppliers – as would IBM, Dell and Oracle – but they could slow down and focus on more profitable opportunities.
Things will never be as cozy between Cisco and its big resellers as it was before UCS. But it would be a lot better than it is now.
Strategic retreat
How does Cisco get out of the UCS business without alienating their very best customers? By selling it to HP, IBM or Oracle.
Get a bidding war going and they might even make a profit. Great way to cement a new reseller agreement.
The StorageMojo take
The bigger issue here is Chambers himself. Yes, he’s done a good job with Cisco, but that was years ago. In the last 3 years he’s made a number of bad decisions: buying Flip; not selling Flip; launching UCS; and NOT buying EMC or NetApp when Cisco’s stock price was high and theirs was much lower.
Routers have 60%+ margins and so does enterprise storage. Networks and storage are fungible in a way that networks and servers aren’t. Private label/reseller agreements are much more common, so Cisco could have moved into storage without enraging its big resellers.
But no. Like many a CEO before him – and he did well longer than most – Chambers no longer understands his industry.
It is time – past time – for him to hand the reins over to a new CEO. He’s made a pot of money, built a great company and, if he wants to preserve his legacy, should move on. Soon.
Courteous comments welcome, of course. David Pogue reports that Flip was about to launch a new camera that could stream direct to the Internet over Wifi. A boon to insurgents everywhere.
Robin,
I see where you might be going with this, but what about VCE? They just announced a new HQ in Texas. UCS is one of the key foundations of that company. I can’t see it happening unless VCE takes over the UCS assets and IP, but how exactly that would happen, I don’t know.
-Chris
Hi Robin,
While Chambers and Cisco did a lot of marketing around Flip, UCS is a very different initiative. Not only is UCS deeply tied in development to the switching product lines (Nexus), but Cisco’s field and partner organizations are spending a lot of resources to support the initiative. Flip sells through websites and consumer stores, UCS has enabled some Cisco VARs to spin up new data center initiatives with new revenue streams. When it comes to HP, of course there is the acquisition of 3COM and the very public war between the companies, but in reality, HP is still one of the largest resellers of Cisco switching. Cisco faces a lot of challenges – most of the server vendors are trying to push them out of the rack (IBM’s BNT products, HP’s ProCurve, Dell has numerous options. There is also tougher competition in the core switching from Juniper, HP, Arista and others. I agree with your point that it’s too late to turn back the clock and not sell servers, but the big question is – does Cisco’s vision of convergence align with service providers and enterprise users for the next round of purchases? The market will vote with wallets. Cisco is a very large company with broad portfolio and it’s natural that there will be rocky points in the road. I would have preferred to see Flip sold off, but it seems to me that the pundits are having too much fun skewering Chambers when he has such a long and successful run.
Why not sell it to EMC?
Hi Robin
My observation is that network devices (switches and also load balancing devices and firewalls) are increasingly being virtualised and run on a virtualised computing infrastructure. If Cisco didn’t have a computing platform within their portfolio they would risk being relegated to a software only play as this occurred and and gaining zero margin from the hardware infrastructure required to support their software products’ functionality. Surely a worse place to be that the 30% margin that you suggest they may make on their UCS product family?
Stuart,
Good points. Cisco’s current investments in UCS aren’t relevant – they are sunk costs – and the real question for Cisco’s next CEO – or current one, if he can stand to look clearly at the mess he’s made – is “are new investments in the low margin server business economically justified by the possible return?” I don’t see how the answer could be “yes” unless Cisco is completely out of new business ideas. Given how out of touch Cisco’s enterprise marketing group is with the rest of the world that may be the case. Which is why a new CEO is the right way to go.
I grant you that Chambers has had a long and successful run at Cisco, but as the SEC mandated warning goes “past performance is no indication of future returns.” And his performance in the last couple of years shows he’s lost his touch. Time to leave with dignity rather than drive a great company down with him.
Robin-
Full disclosure: I work for Cisco, but this is my opinion alone and not that of Cisco.
I respect your opinion but I think the direction you are headed is off base. I do agree with your point about missed acquisition opportunities.
I disagree with your fundamental premise- selling off UCS will not build bridges with IBM, HP and Oracle as you suggest, nor would it drive HP out of the networking business.
UCS is not a casual product for Cisco. It is the culmination of years of strategizing and planning and represents the largest development initiative in the history of the company. Cisco made the bet years ago that the future of the data center would be virtualization/private cloud when both IBM and HP declined its initial offer to them to participate. The UCS incorporates myriad innovations such as increased high-performance memory capabilities, unified fabric, unified management, VN Link, and stateless blades. The UCS is unparalleled in terms of quickly provisioning and scaling virtual infrastructure. Engineers rave about the product’s capabilities and customers rapidly understand how compelling it is; UCS implementations have soared from zero 18 months ago to over 4,000 customers at the end of last year. As Wells points out, UCS is a key foundation for VCE. It is also key for NetApp’s integrated stack offering, FlexPod. Data center is now fundamental to Cisco’s future, and UCS is fundamental to data center.
If part of the equation is the relative profit margins on servers vs. network gear and the way that (might be) impacting Cisco’s business model, then selling the UCS platform to storage vendor like EMC might not be a good match since they’d face similar challenges.
Plus, if Cisco faces challenges selling servers b/c of their key relationships with trad. server vendors, I can’t help but guess that a storage vendor would face even greater challenges given the close ties between servers and storage.
Selling to a server vendor might present some strategic value since a server vendor can better a business model including UCS (and Cisco could then better leverage strategic partnership with the purchaser).
It might also represent a good way for trad. server vendors to enhance or accelerate their own convergence strategies w/ UCS technology.
just ramblin’…
Maybe they should kill their little fibre channel hobby next? They’ve been in the market since 2002 and have barely cracked 30% market share.
Douglas,
Probably not. 1) the margins are fine. 2) preserves the “one stop shopping” and “FCoE to the future” marketing memes. Brocade will slowly wither away without any help from Cisco.
Steve,
Despite the technology razzle-dazzle, CEOs are paid to make good business decisions, and UCS has yet to prove that it was – or could be – a good business decision for Cisco. Or EMC, as I understand that NetApp is winning the lion’s share of the UCS storage business. UCS lowers Cisco’s margins; enrages large resellers; and has no sustainable competitive advantage. Cisco can’t wish those facts away, and the stock market won’t forget them either. UCS is corporate gangrene and there’s only one cure.
User01’s suggestion that UCS be sold to a server vendor – someone who won’t take a margin dilution hit from the business – is a sound one. I should add Dell to the list of prospects.
Robin,
If UCS were a server, I might agree. It isn’t – it’s a hosting platform for virtual infrastructure. It has an enormous competitive advantage as it already has a huge technical lead and the Cisco engineers are hardly resting on their laurels. Just look at HP’s struggles to compete effectively with BladeSystem Matrix or even (at a very minimal level) with blades + VC. The #1 and #3 storage manufacturers have built their private cloud offerings around UCS because it is the only VI HW platform providing the capability to easily and cost-effectively aggregate compute, network & storage resources for enabling automated provisioning of VI. The large resellers may be enraged, but HP & IBM, at least, declined years ago to participate with Cisco in its vision of a virtual future.
Douglas,
Cisco was the market share leader in Fiber Channel until Brocade bought market with the purchase of McData.
Hi Robin,
Fair disclosure: I am a Cisco employee — although the views expressed here are my own.
Let’s put aside your criticisms of Cisco’s CEO because it’s far easier for a self-identified “analyst” play armchair CEO than to know what it’s like to run a company like Cisco.
As far as UCS goes, your post displays a fair bit of ignorance about compute platforms. One could easily make the same argument about HP and Dell divesting from their skinless server businesses because of the 7-10% margins to be had there. They don’t because there is revenue to be made there. Are you also suggesting that those businesses are gangrenes to Dell and HP?
It is simply a mistake to conflate UCS with a legacy servers — and to do so expresses a fair bit of ignorance about UCS, as Stu’s post above points out. UCS is a compute system, and while it contains Intel processors it also contains a ton of Cisco IP, including multiple innovative ASICS, integrated VM-aware networking capabilities and unrivaled hardware element management capabilities made possible by UCS Manager. It unifies compute, networking, storage access, virtualization and management unlike any other compute platform in the market — which is most definitely a sustainable advantage. To suggest that UCS should command margins in line with legacy server architectures is just wrong, and frankly doesn’t reflect to well on your knowledge of the technology in question.
Since you are so knowledgeable about the business models, can you please point me to your proof of your claims that UCS commands 30% margins?
Also, since you claim you are willing to do so in you bio, I’d love to know if any other server vendors are paid clients of yours.
Jesse,
As is usual practice I noted the names of all the server vendors I’ve done work for recently in the post.
Robin,
It seems to me like you’re being illogical here. Equating lower margins with “bad” ignores many things – namely, whether the high-margin product sales are actually affected and whether total sales increase enough to make the lower margins worth it.
Second, from what I’ve heard Cisco is all about vendor lock-in and high support costs. UCS simply provides them more opportunity for both of those.
Third, they’ve already sunk the R&D. They’re not going to sell the product line off now, no way. Imagine if UCS was a startup company partnered with Cisco. I would be investing heavily in said startup right now, expecting Cisco to buy them any day. It’s what HP, IBM, NetApp, EMC, Oracle, etc would do, so why not Cisco?
Taylor,
I’m a big fan of high-volume, low-margin technology that works. The people who have a problem with it are the folks who bought Cisco stock thinking that Cisco would continue to have high margins. Now, if Cisco can add enough value to their servers that they earn 60%+ margins on them, great! But I don’t think they do and because of that investors will be disappointed and the stock will bump along like Microsoft’s has for the last 10 years.
The issue of sunk costs is simple. The rule is that knowing what you know today – regardless of past investments – would you invest in this business. You have to ignore past investments because otherwise you risk throwing good money after bad. So just because Cisco has spent big on UCS is no reason to continue to if it is not generating financial results congruent with investor expectations.
Now you may be correct that the lock-in and consequent support costs are what wins the day for Cisco. What none of the Cisco people has addressed in their comments is why Cisco didn’t just add 60% margin storage instead of lower margin servers to their product mix. It was an obvious option and instead Chambers turned that business over to EMC and mostly NetApp. And in the process they really upset their key resellers and energized their competition.
Maybe in 5 years we’ll see that these decisions were pure genius. But I don’t see how that is going to happen, hence the post.
I have been recently involved in a large UCS POC.
To me the UCS feels like a technology solution seeking a problem to solve.
Yes, it has some impressive features, so what ?
Do they really reduce tco ? Do they really bring that much value to an already commoditizing market ?
I got the impression that they lack the vision, strategy & expertise to find the right niche for it. Not a mass market product..
Not only is UCS a solution looking for a problem, it does this at a completely wrong scale and price point – most of the people needing something similar (unified hand-holding way to cloud) either have already built their infrastructure or are looking to start with three servers and some storage….
Also, I can’t but feel that whole UCS gambit was built to promote fcoe: and fcoe seems like an even worse miss (at least in its early iterations) – trying to carve market away from far superb Brocade stuff by putting Ethernet in the protocol name.
All in all, it seems that Cisco is in the very similar position to Microsoft – owns its market (ethernet and some edge routing) with sub-par yet ubiquitous products, and is trying to find somewhere else to pour all the money they’ve been making. And they are failing at it, just like MS does – because they aren’t known for any great inventiveness, nor quality, but sheer current presence and brainwashing of people in the trenches.
A bit inflamatory I fear but that’s the way it seems to me.
@Steve Kaplan, interesting you should call out Matrix as having limited success as thats not what we are seeing at our company which sells both UCS and Matrix. Matrix has recieved a far warmer approach from most clients as it takes a application centric approach, and with true management tie ins to ALL the major software vendors and hypervisors it gives HP an edge over Cisco who has focussed on a network centric approach to solving a problem they say exists (which I think only exists in some really large hyperscale DC’s).
I couldn’t agree more that it seems UCS is a vehicle to push the FCoE approach Cisco is taking in the hopes of stealing the strong position Brocade has in the DC, and because they actually make good FC switches and SW to manage the fabric, unlike the MDS line.
I agree that UCS provides some INCREDIBLE features, and its approach is definitely revolutionary in a lot of ways. But a majority of the features UCS provides about 95% of the market doesn’t need, and likely won’t need. So it again begs the question “What problem are they really solving? and how much are they really reducing TCO?”…
Disclosure – NetApp Employee – though this is a completely personal set of observations
Robin,
an interesting perspective, though I’m not entirely sure that a lower Gross Margin product line in an adjacent market is necessarily a bad thing or dilutive to Net Margin in a way that should concern a stock holder, providing the pathways to market keep the other important financial metrics like Cost per Order Dollar in-line. I think this is why we’re seeing things like VCE set up almost as a custom built OEM, and the FlexPod offering leveraging the work NetApp has done in creating a successful low-touch data-center focussed integrator ecosystem. I suspect that Cisco would be quite pleased if IBM or Fujistu or perhaps even Lenovo would OEM the UCS in the longer term which could look like an attractive options if it continues to demonstrate the kinds of wins vs HP that I’ve seen over the last few months.
I also think that UCS was a necessary defensive move for Cisco, not just against HP, but also to defend against the way virtualisation led consolidation massively reduces the number of L2 switch ports, especially when you throw in the implications of 10 and 40Gbit Ethernet, the number of switch ports required in a data-center could well be reduced by a factor of 10 and stay there. I’m not sure how important the L2 switching market is to Cisco, but I suspect that an unchallenged 10x shrinkage in an important segment of that market would be far more damaging to Cisco’s stock price than an additional product line with a lower gross margin.
Rather than fight against this trend, or pretend to ignore it, I think Cisco did something brave and admirable by embracing this change. From casual observation, it would appear that in a highly consolidated and virtualised environment, UCS eprovides a platform to continue to sell Cisco’s value add in the form of things like virtual switch port management (Nexus 1000v etc) and security services and makes environments with UCS a fertile ground to continue selling Cisco’s value added network management.
I suspect that even if UCS was sold as a “loss leader”, I think it would continue tob e good for Cisco’s “Core” networking business. Having said that thanks to, what is in my opinion, a significantly superior technology base, Cisco and it’s resellers should continue to build a healthy business without having to resorting to the “unnatural pricing acts” that I’ve seen from other vendors trying to protect their server turf.
Regards
John
Agree with what John Martin wrote. It is not just virtualization, a shop like ours comprising of HP bladecenters has also slower growth in L2 ports. Cisco sees the network moving from the core to the server /blade infrastructure and had no option but to come up with UCS.
Suresh
You’re way off the mark and clearly don’t undestand the datacenter strategy if you think it can be simply sold off to HP or IBM.
Do your homework before writing such uninformed journalism
I have to agree with Joe, the author clearly doesn’t have a good grasp of the UCS offering to write such an article. I’m glad I didn’t have to pay to read this.
Joe & Rick,
I was briefed twice by Cisco on their UCS strategy. It didn’t make sense then and it doesn’t make sense now. So I’m not surprised that Cisco supporters can’t offer a crisp business justification.
John Martin’s comment came closest – and he works for NetApp – to what is likely the truth: Cisco no longer had growth opportunities in its core switch and router market; nimbler competitors like Arista Networks and some open source efforts were threatening real competition; so Cisco had to move into either storage or CPU to keep growing. Chambers could have bought EMC and spun off VMware to pay for it, but decided to go with a home-grown product in the lower-margin end of the business. Bad call.
This isn’t about the UCS technology. It is about whether UCS was and is a good business decision for Cisco. And here’s the business world’s answer: stock is down 40% in a year and Morgan Stanley advocates breaking the company into 3 parts, none of which includes UCS.
Chambers lost his way several years ago. The world has changed and, like many a once-great CEO, he hasn’t changed with it. Too bad for Cisco shareholders.
Robin, great article. I totally agree, the low margin business that UCS brings to CISCO will do nothing but reduce overall gross margins, and further delute the stock.
Given that Cisco initially approached IBM and HP to participate, was the intent to build new revenue sources OR to corner the datacenter stack (compute/data network/SAN network pieces BUT not storage) to break the stranglehold (at least that is what it appeared like back then) Brocade had on SAN networking?
People said similar things about enterprise voice about 10 years ago…. Take a look at your desk, there is about an 80% chance or better that phone has a little bridge on it in the upper left corner….:)
Disclosure – I work for Cisco…
C’mon Mojo, I believe we are due an new Blog article now in light of HP’s recent announcements… “HP PC’s RIP…Is [Insert favorite HP Product Here] next?”
Only seems fair (and balanced)
Jeff, there’s a key difference: HP quit a low-margin, slow-growth business to focus on higher-margin, higher growth business. Cisco entered a lower-margin business where it faces strong, entrenched competitors. There’s a reason Cisco’s stock price is down so much.
Following your line of thinking, I would think HP’s stock price would have risen since they were dumping a low-margin business. But it did quite the opposite – biggest drop since Black Monday in 1987. I guess investors disagreed with the strategy. By contrast, Cisco’s stock price rose after announcing restructuring and killing Flip.
Jeff, HP also laid out $10 billion for a UK software company at the same time – at a valuation many considered exceptionally rich – which given the other downdrafts hitting the market may have been the deciding factor. We’ll know in a year or 2.
True, we’ll know in a year or two (maybe less). It took a little less than that to see what a success Cisco was turning UCS into, so I’ll give you that.