So today, when we face the question of whether to farm out e-mail or CRM or some other major application, we should not be too surprised if we call it wrong. But before we turn to the future, let's take another look at some services and providers of yesteryear. Early on, AT&T, Kodak and Xerox shared an enviable reputation for providing extremely high-quality products. The great irony was that none of these companies actually sold products; they sold services. In the case of Kodak, while the camera did bear the company's name, you had to ship the device back to Kodak to get your pictures developed. Users of the Bell System didn't own their phones; the equipment was provided by Ma Bell as part of its service. And Xerox rented you a copier; it didn't sell you one. Quality control remained strictly in the hands of the vendors.

The Blue Chips of Services

As good as the services offered by AT&T, Kodak and Xerox were, after a while, the buying public became more interested in value. If sacrificing a little in terms of absolute quality translates into significant savings, we know which way buyers will lean. And so it is with services today. Companies that attempt to retain total control over the services they sell while extracting top dollar will suffer at the hands of those that offer better value and are willing to work with customers to realize that value. But this is true only for mature services -- those that are well understood by both providers and customers. Phone service, Internet connectivity and Web hosting are examples of services that are well understood. We're not implying that it's easy to find, negotiate, implement and manage these services, but there is at least a track record for service providers in these areas.

Emerging services are fraught with many more pitfalls. To prove their worthiness, service providers in new markets often will give customers top-notch attention. That's usually good news for early adopters, as long as the service provider stays in business -- a very real concern these days. Early adopters of these new services also must understand that the services themselves will likely morph on their way to maturity. Thus, contracting out new breeds of services may amount to as much hassle as running the services in-house.

So here's the $64,000 question: Should we continue to buy products and run them with our own staff, or should we look to companies willing to take responsibility for providing us with service that's better (we hope) and more economical? History shows that the no-brainers come with a fundamental scale of economy. Wide-area data and voice services are an example. And Web hosting service is becoming a no-brainer; with the burgeoning requirements for multiple high-speed providers, content distribution, and on-site nonstop and disaster-recovery systems, in-house deployment is fundamentally flawed. The economies of scale in many of the emerging xSP models, as well as in human-intensive services such as customer support and desktop management, are less clear.

The xSP Alphabet Soup

The number of service providers after your ever-scarce IT budget has ballooned during the past year. This trend started with ISPs; even the largest corporation wouldn't dream of functioning as its own ISP, so the issue of whether such a provider is necessary is moot. The question is, what sort of ISP do you need? You can find one that will provide you with services ranging from basic domain advertising (for the self-sufficient) to complete Web hosting, e-mail and VPN services (for the fed-up). Web application hosting is an area that's expected to grow quickly, and it's one that might be worth considering. A study by the Yankee Group predicts that money spent on Web and e-commerce application hosting will grow to $5.4 billion by 2003.

The good news is that evaluating ISPs is comparatively easy because of the track record of this sector. But while evaluation is at least possible, certain concerns remain. In particular, negotiating contracts -- especially those with meaningful service-level agreements -- is still a dark art form.

ASPs (application service providers) were all the buzz last year, but, as with the whole dot-com frenzy, the ASP has gone from cure-all to modern-day snake oil. Yet the news isn't all bad on this front. A recent Phillips Group survey of both large enterprises and small to midsize businesses says 65 percent of large enterprises (defined in the Phillips survey as companies with 500 to 100,000 employees) will use ASP services for internal applications by 2004, and 72 percent will use ASPs for e-commerce. Among small to midsize companies (those with 40 to 499 employees), 44 percent are projected to use ASPs for internal applications and 53 percent for e-commerce by 2004.

Two problems afflicted the early ASP market for line-of-business applications. The first and most significant impediment was the lack of software written specifically for ASPs. Some application vendors are just now releasing their Web-enabled, ASP-friendly applications, while others are working on appropriate versions, with delivery as far out as 24 months.

Companies To Watch

Second, many of the early ASPs aimed their services at start-ups and therefore didn't face the extensive customization issues inherent in dealing with established enterprises. For the most part, these ASPs were simply unable to satisfy the needs of such companies. ASPs have realized that offering bare-bones applications with limited configuration flexibility and virtually no business-process consulting won't work for customers with long-standing and complex business practices, supply chains and customer relations. The problem then is that as ASPs beef up their services (if they're even able to do so), the costs start getting beefed up as well.

The number of early-adopter customers willing to pay blue-chip prices for blue-chip services is far smaller than most original ASP business plans accounted for--but those customers do exist. In many cases, the early adopters are turning to "business service providers" (BSPs), as the new crop of ASPs likes to call itself. BSPs are a rarer breed. Forming a BSP requires an investment in manpower, facilities and software -- not to mention the time it takes to create coherent broad services--well beyond what's typical for start-up ASPs. The chances of making it in today's start-up climate, which demands profits in the short run, are slim. The result is that consultancies and software vendors in joint ventures with hosting providers are looking good. Outsourcing of user support and desktop-maintenance services is a different matter. Beleaguered IT departments tired of fighting a losing battle against their own brain drain reach out to providers of such services in an attempt to fill their needs without the headache of implementation and maintenance. The problem is that to achieve the scale of economy necessary for profitability, these service providers all too often paint their clients with a very broad brush. Anyone who has sat on the user end of a service call to one of these providers knows the situation: What the user often gets is oversimplified, bad service. As an industry, we haven't yet hit on the formula for providing these services both adequately and cost-effectively. The requirement for an intimate knowledge of clients, their environments and their products, along with a good deal of person-to-person interaction, frequently offsets the value of the outsourcer's expert staff and facilities.

The Latest xSPs

Then there are the newest service providers, including the SSP (storage service provider) and MSP (management service provider). SSPs recognize the ever-growing need for complex, highly available storage systems, so why not offer storage as a service? Some vendors, such as USinternetworking, offer their service to customers that use the vendor's hosting facilities. Other vendors offer to store extremely large amounts of data in complex dual-level storage facilities. Minimum entry is typically around 100 GB, and terabyte stores are more the norm. Avoiding running your own storage-area network comes at a price: These vendors often charge a few cents per megabyte per month, and there's still the matter of fast connections to your data. On the other hand, if the data needs to be accessed from many locations or from outside the company, this service may make perfect sense.

If ever there was a market that needed a good shake-up, it's the network- and systems-management market. For many IT departments, particularly those in midsize companies, systems and network management has been a nightmare; the tools have been insufficient for the task, too expensive or both. When network applications were mostly internal, the management hassle was an annoyance. Now that network applications are external and driving new business, their need is critical -- and MSPs, of course, think they have the answer. According to projections made in a recent Infonetics study, MSP expenditures will grow 357 percent--from $7.2 billion to $32.9 billion--between this year and 2004. An MSP will monitor your Web site, tune your network performance, analyze your applications and more -- all as a subscription service.

Certain MSP offerings are already looking like no-brainers. Web site performance analysis is a good example: MSPs typically have established points of presence around the country, and sometimes around the world. They can provide their clients with information on perceived performance that would simply be too expensive to collect any other way. However, some offerings don't immediately make as much sense. For example, MSPs that offer to monitor and manage internal network performance either have the daunting task of attempting to remotely manage a mishmash of hardware or require that only specific hardware be used.