Business: Corporate Computing

Washington Post business columnist Steven Pearlstein was online to discuss author Nicholas Carr's prediction that companies will shift to an information technology model in which they purchase computers, software and components as a service from utility providers.

Pearlstein examines Carr's arguments in a column today .

A transcript follows.

About Pearlstein: Steven Pearlstein writes about business and the economy for The Washington Post. His journalism career includes editing roles at The Post and Inc. magazine. He was founding publisher and editor of The Boston Observer, a monthly journal of liberal opinion. He got his start in journalism reporting for two New Hampshire newspapers -- the Concord Monitor and the Foster's Daily Democrat. Pearlstein has also worked as a television news reporter and a congressional staffer.

His column archive is online here .

____________________

Paris, France: Steve,

Excellent article. A comment:

Companies have been outsourcing parts of their IT for years, and we're now seeing more and more 'total' outsourcing deals of the type Mr Carr describes. The last 2 years or so have been a watershed, with IBM, HP, EDS et al working hard to improve service delivery on 'total' outsourcing deals. Reality is catching up to the hype, fast .

To those of us in the Outsourcing industry, Mr Carr's assertion has been a self-evident for some time.

In 2003, Forrester Research tried to counter Mr Carr's 'Death of IT' article by comparing companies' IT infrastructure with an iceberg: the bit above the water still delivers strategic competitive advantage and should be kept in-house. The rest should be outsourced. In other words, Mr Carr wasn't completely right just 90% right. The same applies here.

Corporate IT Professionals: if your job can be done over a wire, start thinking about a new one.

washingtonpost.com: Economy of Scale Might Inspire Companies to Ditch IT Departments (November 30, 2005)

Steven Pearlstein: That sounds about right to me.

_______________________

Glen Burnie, Md: Do you see something "Orwellian" in a giant information utility?

Steven Pearlstein: Yes, I do. There are lots of hurdles to overcome before the utility model can really take off, and privacy/security is probably the biggest. I assume there will be some technological solution to that, after a bit of trial and error. But remember -- the question isn't whether data is perfectly protected in a utility model, only whether it is better protected than the current model, ten years out. The current is not perfect either.

_______________________

Vienna, VA: Hello Steven

Corporate IT departments have a focus on the hardware and software services mentioned in your article, but there are additional functions related to business requirements collection, strategic planning and project management.

Assuming Nicholas Carr's prediction is correct, Where do you believe the "break point" will be between outsourced and in-house IT resources? Will the requirements development function remain internal? Can a typical corporation operate without oversight of the outsourced functions by a internal project management office? Who creates the strategic vision in the technology environment? How much can be removed from direct corporate control before you lose the tech-savvy ability to protect the company's interests?

Thanks

Steven Pearlstein: I think we should be clear: Companies will still need software that furthers their corporate goals and even gives them competitive advantage (sorry, Nick). That would be purchased and developed, if necessary, in house, or on a proprietary basis. But it needs to run on the company's computer system eventually, and its perfectly possible that system/network can be "rented out" from a utility.

_______________________

Richmond, VA: After reading today's column on Outsourcing, which sounds a lot like it was written in 1998, the question that needs to be asked is...

When will people learn that IT outsourcing has NOT been a low cost solution and that various tricks which have been touted as the "next" low cost solution is just a revamp of the last proposed low cost solution? The "new" utility of software delivery has been discussed for at least 7 years. It was first proposed with the "thin" client for corporate users (How much have you seen that?).

The TRUE and real IT problem is that corporations after 30 years of having true computing still have NO clue on how to manage their IT departments effectively.

Until companies get a grasp on what they need and start to use their IT groups to improve their business instead of viewing them as a cost center, we are going to keep hearing these "new ground breaking" ideas that still do not further the development of corporations or IT in general.

(Background: As an experienced IT person whom has worked with outsourcing and off-shoring for 15 years I have heard the same thing about outsourcing being touted over and over. And I watch corporations (I have worked with several goverment, small, and Forture 500 corporations) switch stratiges, sometimes away from cost-effective to chase short term "cost saving" via outsourcing that does not exist only to switch to "insourcing" when the potential cost savings could be realized. Over and over and over again.)

Steven Pearlstein: I guess you are making the case that computing is not the same as electric power, so the analogy really isn't apt. Because if it is, then Carr is right. And to the degree it isn't (and I admit the analogy is not perfect), either some or all of the computing should remain in house. My own sense is that part of it will remain in house, but probably less than you think. Very little of this stuff is of such strategic importance.

_______________________

Columbia, MD: I think that there is an incorrect assumption that IT can be viewed as a standardized 'utility'. Many companies have customized IT to provide a competitive advantage. Would Washingtonpost.com be as popular if it came from the same provider as Cnn.com? I think the more interesting discussion is the growing awareness and implementation of more robust IT governance structures ala COBIT, ITIL, etc. Companies are starting to recognize that IT requires an organized disciplined management structure to ensure that IT investment really delivers. This is not to say that large portions of IT won't be outsourced, but rather that each company will need to make individual choices about which aspects to outsource. CIOs will balance, rather than ditch, IT departments.

Steven Pearlstein: You said it better than I did. Thanks for that.

_______________________

Alexandria, VA: Pundits are always predicting the commoditization of IT, but it's a different beastie than electricity or natural gas, because there's a competitive advantage to understanding how IT works. Plus you can add brainpower to IT and IT works better (unlike electricity). If IT was a commodity, everyone would now be buying it from Microsoft, with Microsoft hookups. Now try switching to Google brand without an IT department.

Isn't there a TV ad campaign selling 'it"?

Steven Pearlstein: Last time I looked, everyone WAS buying certain types of software from Microsoft....

_______________________

Washington, D.C.: I'm curious as to what evidence you have that "outsourcing IT may be cheaper"? Your article did not really give us any answers there. I see this talked about a lot but rarely does it work out that way. Consultants and canned solutions are still very expensive.

Steven Pearlstein: I'm no expert in these things but I assume you wouldn't have companies that have become as big as EDS, CSC, IBM and others have become if IT outsourcing didn't work well and save money. Corproate types may be irrational, and act in herd-like ways according to the latest management fashion. But even allowing for that, the success of these firms over many years suggests that there is something there. It also makes sense in theory -- these are functions that naturally have large scale economies.

_______________________

Philadelphia, PA: Carr is good at getting himself attention. To be serious, though, we have to distinguish between the IT components that are analogous to electricity and those that are not. Your infrastructure -- storage, network capacity, data security -- fits with his proposed model; there's no question that companies aren't going to continue to maintain their own, unless they have a scale and an expertise whereby they can become a service provider too (think, large FS companies).

To some extent applications are moving in this direction -- SalesForce.com is a great example of a process app as a hosted service; it provides 80 percent of what customers need, which is just fine for small and middle market businesses. BPO is a poor example -- you'd only want the most rudimentary business processes managed outside of your company, where you have extremely limited flexibility.

But the catch is that applications are becoming increasingly integrated and responsive to your business process. Through some new technologies like Business Process Management and Service Oriented Architecture, applications can achieve increasing unity with the processes they support. In this world, the ability to command and quickly change your business process by changing the underlying technology is in fact the driver for quickly entering new markets, integrating a merger, cutting costly steps, responding to customer demand, etc. Your business process is critical to your ability to differentiate yourself; the underlying technology is increasingly inseparable from the business process. You simply can't have the command you need or orchestrate the changes that will separate you from competitors if you don't own your processes or the underlying technology.

Steven Pearlstein: I had meant to put salesforce.com in the column, but somehow it never did, so thanks for mentioning it. That is a web-based software service used by many companies that allows the sales force to keep track of customers and leads and contacts, etc. As for your more general point, I think we are moving toward a consensus here, although there are still a few who dismiss the centralization idea completely.

_______________________

Doylestown, PA: As a historian of technology who's fairly familiar with the systems approach to technological innovation, I'm not all that sure that the electrification model is necessarily good for business. Large corporations certainly would benefit, but the problem is one of scale as much as anything else.

In the early years of electrification, the density of urban areas ensured that those were the customers served by the utility giants like Insull's corporation. It wasn't until the government stepped in and subsidized rural electrification in the '30s that rural communities enjoyed accessible electric power. Similar situations existed in the 19th century regarding the construction of telegraphic and rail nets.

The possibility definitely exists that it would be difficult for small businesses to benefit from these IT utilities, especially if they don't have the bandwidth -- which, like electrification, definitely lags in rural areas. Consequently, rural small businesses, the very firms who might benefit most from this utility model in theory, might find it necessary to maintain their IT departments long after larger organizations give them up.

Steven Pearlstein: I think I'd be careful not to let the problem of rural businesses drive our concept of the emerging model for most business activity, which is, after all, located in metropolitan areas.

_______________________

Washinton, D.C.: What is the difference between capacity and usage? Because I have a car that will do 120 and I don't drive 120 am I underutilizing it?

Does anyone care about the workers, or is bottom-lining it destroying democracy, who dares to speak out? Bob Cratchitt anyone.....................what hypocrisy.

Steven Pearlstein: Yes, if what you paid for an engine that could go 120 is significantly higher than the cost of an engine that would get you to 90, then you are "wasting" capacity. Its probably not a big deal in the case of your car. But it may be a big deal to a company when it comes to its computer system's capacity.

_______________________

Kensington, Maryland: Have two questions:

If corporations should move to outsourcing their IT departments, what about the IT portion that deals with the soul of the company, should that be outsourced too?

If the soul or nature of business of a company is outsourced, application software developers should have started several years ago on an effort to support that type of situation. Have you envisioned what sort of environment this developers should be offering their SW in? Is there any SW available right now? For what type of industry?

Steven Pearlstein: No outsourcing the "soul" of the company -- let's all agree to that. But most companies are more body than soul.

_______________________

Washington, DC: Time-sharing was a prior form of utility computing popular in the 70s but faded in the 80s. Why should we expect utility computing to be more than a fad this time around?

Steven Pearlstein: It is possible in theory -- and I think true in practice-- that centralization could have been the optimal solution at the beginning of the computer era and now, but not in between. And it may change back again with some other technological development. Some things move in a straight line. Others move like a pendulum.

_______________________

Excess capacity: The excess capacity argument is irrelevant. If all I'm doing is email and letters, sure, I could do it on a five-year-old PC. No excess capacity. But you can't buy a machine with those specs today -- and if you could, you have to use five year old software.

And the software companies (Microsoft) won't stand for that.

Steven Pearlstein: Obviously, you can't operate a system (any time of system) at 100 percent capacity. You need room for growth. And because there are peak times, you need surge capacity. But it is easier to reduce and manage excess capacity in larger units than smaller, especially when you have a diversity of users who have different peak periods and different growth rates. That's why the utility model is intriguing.

_______________________

McLean, VA: I like your piece today. Web 2.0 really is a new world, however it differs from previous efforts at leveraging scale by the way its architecture makes it possible to do things uniquely to your company. In other words you don't have to give up distinctive value creation. For example, take a look at the poster child for Web 2.0, Salesforce.com. You can do extensive personalization, and leverage Google

Steven Pearlstein: I'll take your word for it on that, since it goes well beyond the computer usage of a newspaper columnist who hasn't yet mastered excel.

_______________________

New York, NY: Speaking as a programmer here in the US, working on payroll and benefits systems, I wonder if you think my job would likely go overseas, leaving me out of luck, or if I would just move over to a centralized company for providing more generic business software?

Steven Pearlstein: Both, I'm afraid.

_______________________

Frederick, Md: I wonder how "open source code" will work in this paradigm?

Steven Pearlstein: Better than proprietary code. I think that is one of the predicates of the utility model -- that all sorts of software works well with all the other software.

_______________________

Clarksville, Maryland: How can one outsource key business applications essential to operations like ERP's that are so very expensive and integrate all business functions under a single application framwork when none of the current crop is really web based? In other words are not fully web capable?

Steven Pearlstein: I'm a bit over my head here, but I think the answer to your question is that this will take time (decades) before the utility model is the dominant one. Installed base makes the system very sticky.

_______________________

Columbia, MD: Ahhh, what you are talking about is often referred to as "Grid Computing" in current buzz words. While it is true that a lot of computing capacity is under used, a lot of times that capacity is there for peak usage, such as 9am Monday morning. So a lot of the potential cost savings of outsourcing the computing power (thereby achieving some sort of economy of scale and increased efficiency) is offset by the need to have the extra computing power available at peak times.

Also there are issues/concerns with availability. If your provider is not available, what would the cost to your business be?

And of course privacy concerns. So while there may be some cost savings by moving resource intensive processes to off peak times on outsourcers systems, I think that much of the touted savings would be hard to capture.

Steven Pearlstein: Good questions, some of which we've dealt with already.

_______________________

Sterling, VA: I agree with Carr that computing as a resource will change over the next few years. If you view it as a resource, then companies can figure out how to use it to a competitive advantage, even if they are all using the same resource.

A case in point: companies can use the same network infrastructure and software, but produce dramatically different results. We have a company that offers phone/fax/Web/database integration and we have hundreds of clients using the same hardware/software to accomplish radically different solutions.

Think about it this way: the only reason that most people do not hire a lawn service to take care of their yard is cost (except for the 5% "Tim Taylors" who love to have the tools...). So if the effective cost of offering this to everyone gets cheap, why would you own your own lawnmower, trimmer, leaf blower, etc.? And everyone's yard doesn't have to look the same just because they use the same yard service.

IT will follow that same idea, I believe. And we have it in action right now, generating revenue and helping to build one company from $1MM to $17MM in revenue within 12 months.

Steven Pearlstein: Thanks. We needed some other analogies.

_______________________

Colorado Springs, CO: Back to the mainframe, centralized, dumb terminal/operator scenario? You are probably correct; however, one of the primary advantages of the PC in the office was individualization and an enhancement of the creativity of the individual.

Questions:

1. Will centralization of IT cost flexibility? Instead of getting what you need ASAP, you will have to wait until the centralized folks get around to it.

2. When they are down, you are down. Doesn't this increase the risk?

3. When your stuff goes there, what happens to it? Doesn't this lessen security?

4. Do we (this country) have an infrastructure that can handle that kind of data traffic?

Thanks.

Steven Pearlstein: All good questions that will have to be answered satisfactorily before the utility model becomes the dominant one. But ask yourself: wouldn't the electric power-dependent manufacturer in 1910 have asked similar questions?

_______________________

Bethesda, MD: Thanks for the provocative column. It seems to me that you are mixing several different ideas. Primarily, you have conflated two separate tasks that IT departments perform:

1. Production support. Keeping the production hardware and software running. This can be outsourced effectively by companies like Digex.

2. Software development. Defining, constructing, testing, and deploying the company's software. This typically requires specialized business knowledge and thus cannot be outsourced effectively.

So I can see #1 going away, but not #2.

Steven Pearlstein: Exactly.

_______________________

Takoma Park, MD: We have decades of experience with multi-user systems. If one of the main issues is the distribution of computing resources you can connect a bunch of X-terminals and keep the servers out of the hands of a third party. But even if outsourcing computing services becomes attractive, web-2.0 seems like old ideas repackaged into something made to sound easy to use. Even on windows there are already multiple ways to run GUI applications on a remote server.

Steven Pearlstein: I'll take your word for it, even though I don't have the slightest idea of what an X terminal is or a GUI application. Let me take this moment to make a plea to you techies: can we please try to use the Queen's English a bit more often.

_______________________

McLean, Va.: Steven,

The problem with using metaphors as a means of IT analysis is that while they seem to address organizational situations at a strategic, they do not adress the implementation issues that make or break the approach.

Several commentators in today's discussion have a good grasp of current organizational IT practices. I think they are correct in saying that the utility metaphor is a tarting point for an analysis of what aspects of an organization's IT architecture can be outsourced, and which are better kept in house.

To put it another way, Carr's metaphor is this month's IT buzz. Let's see what next month brings.

Steven Pearlstein: That's a good place to sum up.

A number of you have written in today about earlier columns and I want to take a moment to get to them. So here goes....

_______________________

Laurel, Md.: Since we haven't chatted in a bit, could I ask about November 18's column about growth fees in the suburbs?

Are they motivated by the inadequacy of all types of infrastructure needs, like water and sewer, or mostly by roads? The problem (well one of them) with suburban roads is that they're usually planned to serve the needs of the community in which they're located, but after a while become thoroughfares used to drive THROUGH those communities.

When I grew up in Gaithersburg in the 70s, everyone complained about how inadequate Rockville Pike was, because it had been built to serve Rockville residents but not to get people from Gaithersburg to Bethesda. The same song goes on everywhere that suburbs are still pushing outward.

We have a tradition of free public roads. I suspect what some inner- to middle- suburbs really want is a road use tax for people one town out.

washingtonpost.com: Suburbs Ask How Much Growth Is Too Much (November 18, 2005)

Steven Pearlstein: That's a very good point. It is also why ordinances that require adequate public facilities to be in place before any new development are flawed. Are you going to demand that the developer of the next project in Rockville bear the entire cost of widening Rockville Pike before you approve his project? As now construed, the ordinances would prevent that development from ever getting built unless the community, on its own, decides to pay for widening Rockville Pike, which it won't, because the "slow growth" people will be out there saying that widening the Pike will only invite more traffic and development. In the end, what you get is no growth, even though that may not be the intent of the slow growth movement.

The way I see it, roads are a very interdependent network system, which is why they are a public good and ought to be financed with money collected from everyone. Its one thing to require a developer to put in the sewer and water pipes necessary to hook up to the system. But it is not fair to ask the next developer to pay the cost of expanding the sewage treatment plant by 20 percent because the current one is now used up.

_______________________

Fairfax, Va.: Outstanding article last week! I was not a regular reader of your column, but count me in -- the article was informative, funny, and left me wanting more.

How did the Native American Contracting Information achieve exception from Freedom of Information Act and Government Accountability Office review? As a former GAO employee, I found this particularly distressing].

Steven Pearlstein: This comment concerns a tongue in cheek column about selling Alaska back to the Russians. As to your question, the exceptions were written in as appropriations riders, as I understand it. But I'm not completely certain about that.

_______________________

Orlando, FL: Have you considered a proposal to give Washington, D.C. to the Russians and we can keep our most pristine land mass -- Alaska!

Steven Pearlstein: Ah, I think I hear the voice of an environmentalist. Why stop with Alaska, however. Why don't we depopulate Montana, Idaho, Washington, Colorado the California and turn them back into pristine land masses as well. You can't have enough pristine land masses--that's my view.

_______________________

Anchorage, Alaska: As an Alaskan liberal (there are 4 of us in the state), I have to save that I was offended by your comments and your sanctimonious lower 48 attitude - which generally relegates Alaska as a nice place for us rich folks to come up and visit and make jokes about refrigerators and eskimos.

By the way you were incorrect on the bridges, only one went to an island, and while they made good fodder for comic relief the simplistic coverage does not capture the complexity of doing business in Alaska or the advantages of doing the same.

The writer needs to leave his cozy Washington office and get up here and meet with folks and learn a bit about its state and its value to the nation it fought hard to join.

washingtonpost.com: Alaska Would Be More at Home in Russia (November 23, 2005)

Steven Pearlstein: I love these appeals to the "complexity of doing business in Alaska." What this amounts to is the predicate for demands for new subsidies from the rest of us because your lfiestyle and economy there is inefficient. If you want to live in remote villages or places with hostile weather and environment, that's certainly your right. But why is it you please continue to expect that the rest of us are responsible for making that economically feasible. You already get a big boost from oil and other natural resources. If that's not enough to pay for the infrastructure you say you need, maybe you should consider moving down here. No passport or work visa is required.

_______________________

McLean, Va.: Steven,

Restricting people to the Queen's English when discussing Carr's idea is like saying that the 30,000 foot view is all that is necessary to implement a new paradigm.

The poster who wrote of X-Windows Terminals is making the point that the utility model has been used within the past 15 years for local area networks. X-Windows terminals were hot within the UNIX community 15 years ago. Today, you're more likely to find a true X-Windows terminal on Antiques Roadshow than in a corporate office. Networked PCs running Linux are now used to develop software. They offer the advantages of having central servers with the flexibility of having custom workstations.

TO put it simply, you can't intelligently discuss many of the issues raised by other posters without getting into the technical weeds.

Steven Pearlstein: Back to tech.

This is one of the great tensions involved in any difficult issue. If you let the discussion be dominated by people who are only in the weeds, you won't get the right answer. ON the other hand, if you let the discussion be dominated by only those with the 10,000 foot view, you won't get the right answer, either. So we need to talk together. And to do that, we need a common language. I propose English.

_______________________

Takoma Park, MD: It isn't painless to read business-speak either.

Steven Pearlstein: I try very hard in my own writing to avoid business speak. It's not always possible. But that ought to be the goal.

_______________________

Steven Pearlstein: Thanks, folks. A good discussion. See you next week.

_______________________

Editor's Note: Washingtonpost.com moderators retain editorial control over Live Online discussions and choose the most relevant questions for guests and hosts; guests and hosts can decline to answer questions.