CRN Interview: IT Pundit Sees On-Demand Software Gaining Traction

IT market analyst Arnie Berman, chief technology strategist for New York-based investment firm SG Cowen and Co., often appears on CNBC and has been quoted in the publications such as The Wall Street Journal, Money, Forbes and BusinessWeek. He recently shared his views on on-demand software, managed services, technology spending in 2006, the impact of Windows Vista and other topics in an interview with John Roberts, 's director of editorial research.

CRN: What basic strategies will businesses follow this year for technology investment?

BERMAN: When I talk about the tech sector, I talk about seasonal recovery, cyclical recovery and sometimes secular recovery. I am now telling a secular recovery story, but I began to tell a cyclical recovery in 2002. And the key thing about a cyclical recovery then was that it was going to be a "last to be cut, first to be re-funded" recovery.

Most corporate technology budgets can be thought of as two broad slices. There's the "run the business spending," and then there's the "change the business" spending. When budgets started to get cut in earnest near the end of 2000, what went first was the wild, far-flung "change the business" spending on new projects--all the guys from Razorfish implementing Web project software from vendors run by 28-year-olds. That's what was cut first. What was cut last was basic infrastructure, the "run the business" spending. Soon I was hearing a lot about PCs that were breaking and not being replaced, storage utilization rates that were beginning to red-line and network bandwidth internally in the corporate premises beginning to run into latency issues that weren't being addressed.

It became clear that this would eventually threaten the ability of businesses to keep the doors open. So under this "last to be cut, first to be re-funded" scenario, it meant that hardware spending would come back before software spending, which has happened. It certainly meant that enterprise spending would come back before telecom spending. And within telecom spending, it certainly meant that wireless spending would come back before wireline and broadband spending because that's where the capacity constraint issues were going to appear first.

The reason I recount all this is that in 2006, the cyclical recovery story is still relevant. But it is maturing to a level where I feel now that the more interesting rate of change dynamic is in the spending that was going to recover last, the "change the business" spending. In 2005, new project spending began to outgrow overall IT spending--for the first time since the tech spending bubble--and I think that trend will become more pronounced in 2006.

CRN: What other themes will characterize business IT spending this year?

BERMAN: Overall [IT] budget growth will not be wild, but there will clearly be a shift in emphasis. I think software will enjoy a better rate of spending growth than hardware. This goes contrary to a lot of Wall Street forecasts, which assume that the software industry can never outgrow the hardware industry ever again. But Wall Street tends to do a lot of its analysis with a ruler, assuming that the future is going to look like the recent past.

Another important trend for this year is that for the first time since the spending bubble ended, company CEOs have stopped feeling antipathy toward their CIOs. Maybe it is because most of the CIOs that were there in the bubble period are now gone and all the hardware and software purchased during the bubble period became obsolete and has been replaced. But whatever the reason, with the antipathy gone, CIOs are now willing to act less conservatively when it comes to spending requests because they know they will not be immediately shunned.

I think another thing that started to change in 2005 and will become more relevant in 2006 is that when it comes to technology vendors, we are moving from consolidation back to more fragmentation. It's not just broad sweep. Solutions that matter and best-of-breed solutions are important again.

CRN: This is basically a reversal of the consolidation process that took place after the bubble popped.

BERMAN: Right. And frankly, a lot of this process is part and parcel to the trends I mentioned earlier, the fact that new project spending is going up once again. I think the most interesting new-project spending dynamic is that on-demand software has become a very relevant category and that this is the beginning of a really important trend.

When people talk about software as a service, or especially when they talk about Web services or service-oriented architectures, they have in mind something that's more comprehensive than what Salesforce.com is doing for you. More and more, software will be run off premises, in effect on demand. There will be more new categories, and you can take an architecture that had previously been highly distributed and make it super-distributed. CRN: It sounds like you view this as a paradigm shift in the technology industry.

BERMAN: I think 2006 has a lot in common with 1991. As part of that, I'd say Salesforce.com and other on-demand vendors are to 2006 what Novell was to 1991.

In 1991, users knew the shift to client-server was coming, meaning a more horizontally integrated, best-of-breed approach to computing. But you had only three choices in the beginning of this process. You could buy dinosaur technology, meaning buy more from IBM. You could buy the "not ready for prime time" stuff, [i.e.] technology from companies that you hadn't heard of and that had not been proven would scale. Or you could wait until you saw which direction the technology was moving and figure out what others were doing and follow their lead.

The hot vendor of that era was Novell. NetWare gave users the ability to work in proximity to each other and have them connect and share files and printers. That might seem quaint now, but I think it was that technology that ultimately paved the way for the much more robust and scalable implementation of client-server that ultimately paved the way for enterprisewide, mission-critical, SAP-style computing.

When I look at what's going on lately, it's kind of the same thing all over again. When it comes to utility computing or Web services, the choices are once again the same: buy the dinosaur stuff, buy the "not ready for prime time" stuff or wait to see what happens. Some vendors are offering products that are functional, but CEOs are not yet running their businesses on these kinds of on-demand architectures. And that makes sense because it's scary when you're not running software on premises, when you know you're putting information out on the Web.

CRN: That's what's brought about the demise of the old ASP model. Companies didn't feel comfortable about putting their information off premises. They weren't sure about the security, and the Web was too slow at the time.

BERMAN: Right. Broadband didn't exist, and the bellwether application service providers, such as USInternetwork, didn't have track records that major users could be comfortable with. They actually had another big problem in that they didn't have provision software; they had no way to measure usage. It was a little like Con Ed setting up shop but not having electric meters in people's homes. The natural providers of these kinds of data services to the corporate market are the phone companies. Companies like USInternetwork had the right idea, but they were before their time. The names are ones that you have been familiar with for a long time, like AT&T and Verizon.

There are two key implications here. The first is that I think what open systems did to hardware, on-demand will do to software. And that is massively lower barriers to entry. Many more companies will get a beachhead in the on-demand space, but the propensity to flame out will be high. The second implication is that when companies buy software on demand, it still needs to run on something. Companies will still need to buy servers and databases. But instead of buying it for on-premises, the focus of computing infrastructure will increasingly move from the corporate campus to the service provider.

CRN: VARs are concerned about disintermediation, that vendors will take away their customers and provide all the products, services and management they need. Is this a legitimate concern?

BERMAN: A lot of what end users have been doing for years is all about integration. Everybody bought all this new software, and it didn't communicate with everything else. Getting all systems and software to speak to each other and share data has been a bear of a project that is still ongoing. I frankly think that one of the things that comes with all the new technologies is a second significant wave of integration. This is going to keep a lot of people busy for a long time. Think of all the mobile devices that are out there and that are coming, like location-based services on your BlackBerry or your phone. Or think of RFID technology. Someday, you will be getting incredible amounts of data through these systems that won't be particularly well-integrated into existing systems.

CRN: So there's still plenty of room in areas such as customization and integration for solution providers that have the right skills.

BERMAN: That's right.

CRN: Managed services is another hot topic. How would you define managed services, and what are some key trends?

BERMAN: When I hear the phrase [managed services], it means so many different things to different people that I don't necessarily have a definition of my own. But I think what some of these managed-service models are attempting to do is build a business and do it in a way that makes fixed costs become variable costs and includes the services necessary to support the products that the vendors are selling. They're offering an attractive solution that the users will want to buy. Subscription-type, "pay as you go" variable-cost models are, as a proportion of IT budget activities, getting a lot bigger. And I don't think it's coming anywhere close to maxing out.

CRN: One big development on the way this year in the IT sector is the release of Windows Vista, Microsoft's next-generation operating system. Will Vista make a big difference in the marketplace, or will it be a yawner?

BERMAN: It's not going to be a yawner. Vista changes the user experience more than anything since Windows 95--search capabilities as an example. Think of having an office that has hundreds and hundreds of filing cabinets. In an ideal world, if all these files were in hand, all you would have to do is say, "You know, I'm thinking about this thing," and a drawer opens and the file appears in your hand. Vista will mirror that, and it's going to be significantly better than Google desktop search.

With the Microsoft search box, not only will you be able to search every kind of file you could conceivably have on your computer [hard drive], but also it will even search in, for example, e-mail attachments you've never opened. It will search everything. This is so appealing that if your PC is working well enough and you can convince your company to wait for Vista, I'd encourage them to wait. This is another reason why I think the PC market [growth] is going to be slower [than the software market] this year. The outlook for next year, in turn, is pretty good because I think Vista really does change the user experience.

CRN: Which vertical markets will lead the way in IT spending this year?

BERMAN: The big drivers in technology spending over time have tended to be either the financial services industry and the technology industry itself. And when it comes to some of the new projects that I have been referring to, it could be the case that these will be led by those sectors. But I think, overall, you're looking at what will be a decent but unspectacular growth year for [IT] budgets, with a less pronounced shift in how they're spending their budget in other sectors compared to the financial services and technology sectors.