People working at airlines don't like to talk about crashes, not even those affecting their computers.How do you argue with that logic? IT departments don't exist to keep Sun, IBM, Oracle or Microsoft in the black. They exist to manage corporate information technology and enable the business to function.So it might seem strange to see a major airline turn to Windows, a product much-maligned as crash prone, as the only way to run a successful business. Yet for Jeff Cohen, chief information officer of JetBlue Airways, Windows is the only way to run a successful business--in fact, it's a nearly 100 percent Microsoft software operation. [...]
JetBlue, which operates one of just a few profitable airlines, saw its operating revenue jump 96 percent year over year in the fourth quarter and more than 98 percent for 2002 compared with the previous year. [...]
Cohen's organization would like to take some credit for those margin gains. By standardizing on one operating system and using other Microsoft software, the JetBlue CIO says he cut the company's technical staff by 50 percent. Electronic publishing of pilot manuals, which are available on a cockpit laptop, also helps cut costs. For Cohen, standardization of information technology, like the approach to the planes, is helping to keep JetBlue in the black when other airlines fly in the red.
-- Helping JetBlue see black
(emphasis added)
Microsoft is the only IT vendor that aims to offer a standardized solution for the entire enterprise -- from the palmtop and the laptop to the backoffice server clusters. (Don't even think about telling me that Linux hits all of those targets -- we all know that Linux is anything but standardized.) In a sense, they're also the only IT vendor who sees this problem and aims to solve it. Sun? Java? They only address aspects of the problem -- commoditize the OS until it's a non-issue. But the CIO needs to worry about managing every OS installed in his enterprise...
Do the gains from a single standard platform outway the costs of using better technology? Sure, there are risks to homogeneity, and yes, there are more security issues than average when dealing with Microsoft software. But none of those risks will ever go away; they can only be hedged. So if you're willing to pay all of those costs, do you still come out ahead by standardizing everything on a single platform?
Re:more like Aesop's Fables...
dws on 2003-04-24T06:08:01
Especially when you have to pay an unpredictable yearly license fee on the basket.
Naively, I thought my NPOs would be interested in OSS ("Linux" in their minds, primarily) for cost reasons. Not so: they get big breaks from MS and sometimes even sets of free licenses.
The for-profit people, some of whom are keenly aware of OSS availability and its possible benefits, need to keep streamlined and not get distracted from their busniness goals; they must avoid the overhead and riskiness of trying to marshall disparate semicompatible systems and applications. And often there is a groundswell of OSS lovin' going on amongst the rank and file so training and expertise isn't really the issue; it's maintaining efficiency and focus.
These are not the organizations for whom cutting from 50 NT admins to 10 Unix admins might be cost effective (if you overlook all the user retraining for the non-admins), these are representatives of the huge number of small and medium size organizations that absolutely cannot afford the increased TCO of OSS software. Yes the TCO might be lower if they didn't already have/know/interoperate-with MS products, but that's not the real world.
I would positively love to push/pull/help my clients towards OSS, but there are very few for whom it actually makes sense. Even I end up having to use MSOffice event to read some clients' documents, OpenOffice is just not compatible enough with it. It is getting much, much better, but I can't cut over to it all the way, I launch it about 20% of the time instead of using Word, but I cannot author or edit in it without frequently ruining the
Sigh. Enough negativity for one day. Back to work...
I think these guys should be talking about their project management rather than their windows standardisation which usually only really offers benefits to a handful of middle management and purchasing suits.
Linux offers benefits closer to the shop floor that these people often don't notice. I know this because I have seen it in several different companies varying from global corporations to cowboy MCP shops as well as a variety of companies in between.
Don't take the CIO at his word here. Note that Southwest Airlines, which has a similar business model to Jet Blue, and has been doing it a hell of a lot longer, has always turned a profit also. I have no idea what Southwest uses in their IT department.
The CIO has to say something good about his decision to standardize on MS, right? It was his decision, after all.
The fact that so much of the Web is run by Unix/Apache, so many palmtops are Palm OS, so many corporate databases are Oracle says something about economies of scale to be had with non-MS products.
Anyone thinking long term will realize that if you completely standardize on one vendor, that one vendor could double their price to you tomorrow and you'd have to pay it. Anyone following MS pricing, which goes up in a recession, realizes what can happen.
You can standardize on MS clients, Unix or IBM Mainframes in the back office and Palm on the palmtops and have similar cost savings. After all, you really wouldn't want the same people supporting clients, big databases and palmtops anyway, so you might as well put the best product in where it makes sense.
There is some argument for interoperability, but be careful not to overstate the advantages. Further, with one vendor everywhere, you are in an inflexible position if you find an important technology becomes available and your vendor isn't doing it yet, or well. Think about having standardized on MS in 1995 and suddenly finding out you needed a Web presence.
Re:He had to say something
ziggy on 2003-04-23T17:30:20
Yes, and it appears that he made a sound business decision, based on a handful of objective metrics that impact the bottom line. He's bucking the common wisdom that (1) you need multiple platforms to run an enterprise and (2) an all Microsoft shop is a solution for managers who don't know any better. These were conscious decisions he took for JetBlue, and interestingly enough, it seems to be having a net positive impact on the bottom line.The CIO has to say something good about his decision to standardize on MS, right? It was his decision, after all.You seem to be missing the point here. Internally, JetBlue is in a better position if it standardizes on a single platform: A320 and Microsoft Windows being two examples. The trickledown reduction in complexity for the organization in terms of training, support and parts inventory directly impacts the bottom line. This kind of standardization allows the organization to be leaner and more frugal.The fact that so much of the Web is run by Unix/Apache, so many palmtops are Palm OS, so many corporate databases are Oracle says something about economies of scale to be had with non-MS products.Forget about Windows for a moment. The same things would be true of an organization that standardizes on RedHat + MySQL + Apache + Perl + KDE, if such a stack were comparable to everything Microsoft offers today. However, there is no single-vendor end-to-end stack available today, except from Microsoft. The issue is less about Microsoft than it is about standardizing on a single platform across the enterprise.
The function of a CIO is to keep the IT infrastructure humming so that the business can do its job. There's a factor of active planning involved, and a much larger factor of day-to-day management and budget management. As a CIO, I would be remiss in my duty to the shareholders if I architected a large, complex, costly multivendor IT infrastructure that is unnecessarily complex and does not outperform a simpler one infrastructure.Anyone thinking long term will realize that if you completely standardize on one vendor, that one vendor could double their price to you tomorrow and you'd have to pay it. Anyone following MS pricing, which goes up in a recession, realizes what can happen.Hypothetical arguments that my costs may be more in the future do not help a CIO do his job (and contain costs) today. Things move slowly in this industry, and if something comes to pass that one or more of my platforms becomes a liability (higher costs, greater insecurity), then a good CIO will address those issues if and when they come up.
The received wisdom in IT planning is that you absolutely need heterogeneity in IT. JetBlue is questioning that assertion, and it sounds like they're coming out ahead (if the nonexistant numbers are to be believed). That is, they're accepting the risks of a single vendor solution and the security risks of Microsoft products, and realizing their cost savings in reduced staffing, training and inventory.
That's just more received wisdom. This article claims that assertion, although widely believed, is wrong. Or at least not universally true.You can standardize on MS clients, Unix or IBM Mainframes in the back office and Palm on the palmtops and have similar cost savings. After all, you really wouldn't want the same people supporting clients, big databases and palmtops anyway, so you might as well put the best product in where it makes sense.Again, a CIO does not run his business on hypotheticals, but on risk management. It's not 1995 anymore. And Microsoft is arguably more interested in staying at the leading edge of the curve than, say, HPaq or Sun.There is some argument for interoperability, but be careful not to overstate the advantages. Further, with one vendor everywhere, you are in an inflexible position if you find an important technology becomes available and your vendor isn't doing it yet, or well. Think about having standardized on MS in 1995 and suddenly finding out you needed a Web presence.Re:He had to say something
jordan on 2003-04-24T01:57:45
- Yes, and it appears that he made a sound business decision, based on a handful of objective metrics that impact the bottom line.
Appearances can be deceptive. This article is not exactly bristling with objective metrics.
This CIO chose Office 2003 over XP for its XML support. XML support that is mostly a marketing checkoff as, by all accounts, it lacks interoperability and formatting information. Sounds like he's got a One Microsoft Way story to tell and he's pushing it for all it's worth.
- He's bucking the common wisdom that (1) you need multiple platforms to run an enterprise and (2) an all Microsoft shop is a solution for managers who don't know any better.
I don't know about common wisdom. The largest marketing department in the world is pushing the vision of MS everywhere, so it's hardly a brave decision on this CIO's part.
- The fact that so much of the Web is run by Unix/Apache, so many palmtops are Palm OS, so many corporate databases are Oracle says something about economies of scale to be had with non-MS products.
You seem to be missing the point here. Internally, JetBlue is in a better position if it standardizes on a single platform: A320 and Microsoft Windows being two examples. The trickledown reduction in complexity for the organization in terms of training, support and parts inventory directly impacts the bottom line. This kind of standardization allows the organization to be leaner and more frugal.
Standardizing has efficiencies, sure. You can standardize on a given aircraft to save on training and maintenance costs, a model that Southwest pioneered a long time ago with its fleet of 737s. You can standardize on database software, or client software, or servers, etc.
I just don't see the big synergies on standardizing everywhere on MS technologies. Computers are used for widely different functions, from palmtop to back office. Standardizing on all MS everywhere would be like Jet Blue standardizing on a single 'transportation' platform, using taxiing jets to deliver maintenance equipment.
If Linux/Apache delivers better price/performance for Web-based apps, as even a Microsoft funded study shows
,other non-Microsoft funded research shows a tremendous savings. I just don't see where the big synergies are by having ALL MS in your shop. Seems like you could partition off a place in your IT infrastructure for Web servers running Linux or BSD.
It's pretty telling that Yahoo and Google run almost no Windows, I think. Clearly, when you really need performance, which at some level translates to money, you don't go with Windows.
It's not really standardizing on one platform anyway. In the back room, there are big Unisys mainframe systems running Microsoft Server 2003, on the departmental level, there're probably HP servers running Windows 2000 Server (possible 2003 Server), then HP Compaq desktops and laptops running Windows XP and Palmtops running WindowsCE. Supporting these different machines requires widely different training, certification and expertise levels.
Microsoft may sell it as all one platform, but that's just marketing spin. You are still going to have Server admins and desktop admins and Palmtop support, etc.
- Hypothetical arguments that my costs may be more in the future do not help a CIO do his job (and contain costs) today.
There's nothing hypothetical about Microsoft pricing practices. Only Microsoft raises prices in a tech downturn. It wouldn't be prudent for any executive to ignore this history or the realization that they would be faced with paying through the nose or painful migration costs in the future if Microsoft dramatically raised prices. It is a certainly part of a CIO's job to look to the future.
- Again, a CIO does not run his business on hypotheticals, but on risk management. It's not 1995 anymore. And Microsoft is arguably more interested in staying at the leading edge of the curve than, say, HPaq or Sun.
Microsoft has been mouthing this Windows everywhere strategy for more than a decade. In 1995, they were saying they were leading edge and ignoring the Web as a passing phase. Seeing Microsoft's pathetic history of lack of innovation, I could easily see them blindsided again by some new technology.
Re:He had to say something
ziggy on 2003-04-24T03:12:50
I'll grant that this article is hardly brimming with real information. Yet I still find myself intrigued with the idea that eliminating non-MS platforms from an enterprise could possibly reduce overall IT costs. Specifically, that the same economic factors that aid jetBlue and Southwest to save money by standardizing on a single model of aircraft may also be a factor in IT.Appearances can be deceptive. This article is not exactly bristling with objective metrics.Looking at costs at the micro level, you can certainly save money on a per-function basis. What this CIO is saying is that there are greater costs to be saved at the macro level by standardizing on an all-Microsoft platform from stem to stern. And that's something I've not heard anyone assert before. Usually, the justification for an all-Microsoft enterprise is much weaker, focusing on inconsequential issues like size of the developer pool, cost of the hardware, or "betting on a market leader".If Linux/Apache delivers better price/performance for Web-based apps, as even a Microsoft funded study shows, other non-Microsoft funded research shows a tremendous savings. I just don't see where the big synergies are by having ALL MS in your shop. Seems like you could partition off a place in your IT infrastructure for Web servers running Linux or BSD.I've worked in predominantly MS shops and in mixed shops. In the mostly MS shops, it is expensive to introduce the auxiliary platforms, whether it is Netware, Linux/BSD or commercial Unix. In the mixed shops, I've also seen that the number of admins and related costs do increase compared to a homogeneous environment. So, in my experience, the cost drivers this article cites are real, and I'm inclined to believe that these cost savings are possible.
Of course, the devil is in the details.
:-) Yes, and neither Yahoo nor Google are remotely similar to jetBlue. Both of those companies view technology as a core competency, not a support function. jetBlue's experiences, if repeatable, are not a blueprint for a new IT management plan for every company on the planet.It's pretty telling that Yahoo and Google run almost no Windows, I think. Clearly, when you really need performance, which at some level translates to money, you don't go with Windows.Merrill Lynch has come to similar conclusions. However, they have years of legacy systems to manage and maintain. Their current view is that a monoculture is bad for a variety of reasons, but a large number of supported platforms is tantamount to anarchy. They aim to standardize on 4-6 platforms to (1) reduce management costs, (2) provide a healthy mix of environements for a variety of uses and (3) arbitrage vendors against each other. But ML is in a different space that either Google, Yahoo! and jetBlue. You should expect that they have different needs and come to different conclusions.
Re:Licensing and fixes
ziggy on 2003-04-23T17:47:35
See my (rather lengthy) response above.Jetblue may have slashed their IT staff in half but what happens when the next round of patches and licensing comes arround?I still assert that it is possible for jetBlue's CIO to have accounted for licensing issues, vulnerabilities and upgrades and still come out ahead on the bottom line through a streamlined IT infrastructure.
I don't have his numbers in front of me, nor do I have numbers from similarly sized and similarly tasked IT departments, so I will not assert that his plan is more cost effective.
Re:Licensing and fixes
gizmo_mathboy on 2003-04-23T20:33:00
It would be interesting to see a cost breakdown. The interesting bits being licensing and upgrades.
Also the "application stack" is truly the weakness in the OSS system. While a lot of the stack is there there a probably some that are vital to jetBlue's business.
If I would hazard a guess the largest outlay of money is for apps that jetBlue uses. I would think that the biggest chunk would be scheduling software unless they're using someone else's Sabre system but I would think that someone coded up an API to that (guessing Sabre is not running on Windows).
I just reiterate my wish to see the cost breakdowns and savings.Re:Licensing and fixes
ziggy on 2003-04-23T20:37:09
Hear, hear.I just reiterate my wish to see the cost breakdowns and savings.
Re:benefits of standardization
ziggy on 2003-04-25T15:22:19
To some extent, he's right.The CIO's argument hinges on the assumption that Microsoft software fits together into some grand design.First, concede the desktop. Windows, Office, VisualStudio, Outlook, etc. are the corporate standard in 99.44% of the corporate world.
Next, the department level. Standard capabilities like file/print sharing, email and web serving can be performed by anything slightly more powerful than a PalmOS device. Most companies standardize on Windows here out of ignorance; jetBlue sounds like they're standardizing here on a worse-is-better line of reasoning. Sure, it'd be better to run these services on Solaris, Linux or whathaveyou, but that introduces costly complexities into the IT space; standardizing on Windows reduces costs.
Finally, there are the "enterprise applications". This is the weakest link in the chain, where most serious businesses start to deploy Solaris if they haven't done so already. Again, this CIO is using a worse-is-better line of reasoning; better to use VisualStudio to develop and Win2K to deploy and make your developers fungible within the organization. The alternative? Two (or more) tiers of developers introducing redundancies and inefficiencies in the IT structure. Oh, and vendors provide solutions on Windows or they go with another vendor; he stated that there were instances where Oracle-on-Windows was required for some application or another.
<devils-advocate>In my experience this is simply false. If you buy all the products they put out in one calendar year there is good synergy. Three years down the road, it's a very, very different story. Microsoft in particular keeps doing this because they rush to adopt certain technology paradigms, badly, and then spend the next several years undoing their mistakes.
Vendors in general are subject to this behavior. There is no guarantee that a multi-vendor environment will prevent these missteps from occurring. For example, would swapping out Microsoft's CMS for Zope or Vignette make your website any easier to maintain in three years time? Probably not. Is the state of file-and-print services in such flux that an all Microsoft solution is significantly more risky than, say a *NIX based solution? No.
</devils-advocate>(I'm explicitly ignoring concerns like uptime, security holes, scalability and size of the admin staff to run a network like this. I presume the jetBlue CIO is doing the same.)
I think you're overstating the case here. Incompatibility isn't the only driver here; adding features (and fixing last year's mistakes) are another significant profit center.But the problem with any single-vendor shop is that they make their money by breaking the beautiful crystalline palace they sold you last year. Incompatibility is a profit center.Of course there are significant risks to a single vendor approach, because their interests are certainly not aligned with yours. Yet, with all of the arguments against single vendor shops, I'm intrigued that there's even the remote possibility that an all Microsoft shop could be more cost effective than a typical heterogeneous shop.
I'm much more familiar with their older, more fascile arguments in favor of an all Microsoft shop: easier to find (certified) staff to run your shop (everyone uses the stuff), cheap body of labor to draw from (from oversupply), cost savings through cheaper hardware, easier to develop applications through wizzy tools and wizards.BTW, I'm rather suprised you've never heard that argument before; it's *the* argument that MS promulgates, AFAICT.Maybe I've been tuning out the marketroids for the last couple of years?
:-)
