Marketing Memos

Linux will dominate the server room in large enterprises, and Windows will hold the lead for SMBs.

How’s that for blatant prognostication?

Evans Data reports that almost half of enterprises will likely replace Windows servers with Linux, while they are rapidly replacing UNIX servers as well. Peerstone Research estimates that 70 of enterprises using Windows to host ERP systems will migrate to Linux.

But this drift to Linux is not always reflected in SMBs. Though the survey data is less robust in smaller firms, there are indicators that where Windows is installed, it will stay installed.

Why is this, and what does it mean in the long run? Let’s answer the second question first.

Over time, Linux will become the dominate server-side technology for enterprises. These organizations have massive investments in UNIX, and staffs trained to maintain mission-critical applications on UNIX servers. As commodity hardware becomes more powerful, and the cost points for LinTel servers drops even further, the switching costs from UNIX to Linux becomes very low (and perhaps negative when amortized over several years).

Installed Windows servers in enterprises will fall victim as well. CxOs have clearly shown a preference to reducing the number of in-house technologies, and are centralizing on Linux as the key infrastructure component. Since Windows has not made significant inroads in hosting large-scale applications in enterprise, CxOs will be motivated to switch Windows servers to Linux wherever possible to rid themselves of one more technology set, and one more onerous licensing black hole.

So, in high-end shops with existing UNIX expertise, the drive to reduce cost and technologies will drive Windows off of the raised floor.

But not out of the SMB computer closets.

SMBs typically have IT staffs with more limited depth of expertise. Windows is installed on all desktops, and Windows Server has a high degree of preference due to ease of administration. Given this, SMBs rarely have UNIX gurus, and thus the switching cost for SMBs (switching from Windows to Linux) is much higher.

At present, Linux vendors need not do much. Despite improved scalability options for Windows, enterprises will prefer to switch to Linux, or at worst, stay on UNIX. Microsoft has run into a Great Wall and has no ladder.

But most swords cut both ways, and Linux has found a Great Wall surrounding the SMB market. Windows Server is “good enough” for the average SMB, and unless Linux creates huge gains in manageability, cost savings and/or ISV support, Windows will continue to dominate this market.

Stalemate? For now. But over the long run Linux has the potential of creating the ease of use and ISV support necessary to erode the bricks in Microsoft’s SMB wall. Microsoft does not have the same luxury within the enterprise.

Bookmark to:

Branding gurus (like me) tend to lean toward creating one very strong and well positioned brand. Indeed, I once consulted with 3AM Labs, who had split their technology positioning between two brands – one for the end user and one for the IT professional who lived in fear of renegade end users. The synergy between a unified brand is almost always preferable to the inevitable brand confusion created by multiple public faces, and 3AM Labs made one brand out of two.

But one brand is not enough for Dell, and Dell is right.

Dell has long been about low cost computers. Their painfully well documented history has revolved around optimizing production in ways that reduced build expense, fatten margins while under price competitors. This was a great strategy for an industry that was by and large a commodity business.

But Dell ran into a problem with the segment of the PC market that wants more of everything. High-end PC buyers – radical gamers, power mad executives, those that wanted nothing but the newest and most expensive top-shelf PCs, people with money to spend on services - they were not shopping Dell first. Despite selling some mighty PCs, Dell was not the 1st thought in the minds of buyers of those machines.

So when Dell decided to launch a premium line of PCs, they did not want these machines to be perceived as Dells, which are perceived as inexpensive. Dell thus decided to launch these blazing boxes under a different brand.

What is the lesson for marketing pros in these two examples?

First, your brand communicates the value you deliver to the buyer. That’s why you spend so much time and money on building and refining a brand — because it becomes the all-encompassing statement about why people should buy from you.

Second, keeping one brand is a smart strategy if your different product lines share a mission, a value proposition, or compatibility. In the case of 3AM Labs, they were selling complimentary products to end users and the IT gurus who had to keep these same end users from creating IT havoc and security problems. Better to give the world a single view about how your products interact than to have users and IT guys think you offer incompatible solutions.

Third, if any product or line of business does not reinforce the central brand, it should have its own brand. BMW’s brand is summarized in their slogan, “The Ultimate Driving Machine.” Do you think the cute and petite Mini Cooper is an Ultimate Driving Machine? Of course not, but it is still a BMW product. This incompatibility of product line and brand is why Mini’s have their own brand – so they can avoid polluting the BMW brand, and so the BMW brand does not turn away buyers who are looking for the qualities found in Mini Coopers.

Dell has chosen wisely. Now, if only HP would catch on . . .

Bookmark to:

I recently helped Novell’s Retail division with some market messages. Naturally, these were all based around Novell Linux for Point of Service (NLPOS) and IBM Retail Environment for SUSE LINUX (IRES), the latter being a superset of the former.

During this effort, Microsoft – the seemingly endless nemesis of all things Linux – released Windows Embedded for Point of Service, or WEPOS (which unfortunately is pronounced “we pause”, which is a better explanation of their Longhorn development efforts). What is instructive is how Linux and Windows are looking upon the POS market, and how Linux will own the space ( are you listening Martin? ).

The critical factor for Point of Service terminals (formerly Point of Sale, but always POS) is openness and flexibility. Retailers have long suffered the strategic problem of POS selection, knowing that they have to live with their decision for five or more years. The reason they have such a long ROI period has to do with the multiplicity of POS devices.

Take an outfit like Circuit City, where I once worked and where the IT department is gleefully abandoning their roll-out of Windows for POS. With 600 stores, and 5-10 POS terminals per store, they have a roll-out device count of approximately 4,500 terminals. Not only is the cost of rolling out new devices huge, so is the tech travel time, training, tech support, and other logistical nightmares.

And herein lies where IBM and Novell have thought-out the market better than Microsoft. Novell and their IBM partner recognized several basic issues:

1. Cash registers have morphed into service terminals.
2. Application for customer service will constantly change.
3. Thus the OS on the POS terminal must be upgradeable and serviceable
4. The cost of POS terminal technology is an important factor . . .
5. But not as important as survivability and expandability

So why does this leave Microsoft paddling for breath? Cost and survivability go hand-in-hand. Embedded Windows makes switching costs higher, and tempts vendor lock-in in the age of commodity hardware. Retailers, who on average have POS terminals over seven years old and who are ready to switch, want to avoid any vendor lock-in because they have suffered badly from it in the past.

By choosing Linux, and by choosing a non-embedded OS, retailers eliminate two significant points of vendor lock-in, and cut costs in the same action. That’s one of the reasons Circuit City is dumping Windows and switching to Linux for POS. And, according to an IT buddy of mine still employed at Circuit City, the maintenance nightmare they experienced in patching and updating Windows on POS was enough to not only ditch the product, but to set fire to Bill Gate’s Porsche for revenge purposes.

The other Microsoft mistake is that they are now releasing customized versions of Windows Embedded for industry verticals, namely retail and hospitality. “WEPOS is the first time we have developed an OS specifically for an industry vertical,” said Jason Demeny, the product manager for WEPOS at Microsoft. This will lead to inevitable problems with Microsoft releases, patching, security, and more as Microsoft begins to support variants of variants (Windows Embedded is a variant of Windows, and WEPOS is a variant of Windows Embedded – who has ownership of this mess? ).

The Novell approach, on the other hand, is to have Linux on each POS terminal, and then boot different personalities based on centralized policies. Each POS terminal can run text applications, GUI apps, or web enabled apps. But each POS terminal has the same software, the same deployment process, the same patching process, and the same patches. More consistent, more stable, and less cash register down time (and if you want to see a Circuit City CEO get angry, tell them that 4,500 associates can not sell TVs because of a Microsoft bug).

Perhaps the only thing more interesting than Microsoft’s misunderstanding of the market is IBM’s utter understanding of the market. See, IBM asked Novell to develop NLPOS as the base for IRES. They heard the demand from the retail sector to provide Linux on the POS terminal because retailers already had plans on putting Linux on in-store servers, and back office servers. One operating system from head-to-toe in retail.

But, this was not universal, and some Windows users wanted WEPOS. So IBM, being the pragmatic outfit they are, said “We’ll support whatever you want” and is now supporting WEPOS on the same hardware that IRES uses. Gotta love a vendor with a plan.

Bookmark to:

Since everyone (including dear old mom) has bashed Sun’s acquisition of StorageTek, I’ll avoid adding to the din. But a quote from Sun’s leader regarding this event is worth noting:

“We have, with our ID management directory, Java card, access manager and Web services stack, the ability — from develop, create, capture manage and store and archive — to handle the entire lifecycle of data securely with a better chance of the appropriate levels of privacy and availability of any body else out there in the market,” McNealy said on a conference call.

And no mention of servers . . . .

Perhaps little else has shown how the commoditization of server technology is forcing vendors to seek margins elsewhere. This is direct result of Linux, of the Intel/AMD interoperability, and of IBM/HP/Dell pushing both into the laps of waiting customers.

However, this deal is likely for naught. Sun has an ongoing affliction which, politely classified, would be called Mumble-Mumble Disorder. Sun is amazingly inarticulate about everything except the artful bashing of competitors. I love McNealy’s bombastic proclamations as much as the next geek, but that doesn’t sell technology.

During the tech downturn, every vendor in Silicon Valley claimed they were going “back to basics”, and many kept their promises . . . well, if mass firings can be called a basic strategy. One of the basics of technology marketing is the art of communication: telling the customer why your wares are valuable to them. No CxO or techie will invest time or money unless there is a value proposition and a brand to push it. Sun little of the first, less of the second, and no ability to communicate either.

Sun once had talent for this, but no more. I fully expect them to mumble their way out of the storage boom as well. And to think, for $4.1 billion dollars they could have hired me to fix their communications problems.

“We have, with our ID management directory, Java card, access manager and Web services stack, the ability — from develop, create, capture manage and store and archive — to handle the entire lifecycle of data securely with a better chance of the appropriate levels of privacy and availability of any body else out there in the market,” McNealy said on a conference call.

And no mention of servers . . . .

Perhaps little else has shown how the commoditization of server technology is forcing vendors to seek margins elsewhere. This is direct result of Linux, of the Intel/AMD interoperability, and of IBM/HP/Dell pushing both into the laps of waiting customers.

However, this deal is likely for naught. Sun has an ongoing affliction which, politely classified, would be called Mumble-Mumble Disorder. Sun is amazingly inarticulate about everything except the artful bashing of competitors. I love McNealy’s bombastic proclamations as much as the next geek, but that doesn’t sell technology.

During the tech downturn, every vendor in Silicon Valley claimed they were going “back to basics”, and many kept their promises . . . well, if mass firings can be called a basic strategy. One of the basics of technology marketing is the art of communication: telling the customer why your wares are valuable to them. No CxO or techie will invest time or money unless there is a value proposition and a brand to push it. Sun little of the first, less of the second, and no ability to communicate either.

Sun once had talent for this, but no more. I fully expect them to mumble their way out of the storage boom as well. And to think, for $4.1 billion dollars they could have hired me to fix their communications problems.

Bookmark to: