Understanding Service-Level Agreements

In the emerging digital economy, thecost of a technology failure can be enormous. And when an outside provider hostsan application or service, the complications can multiply. As a result,companies choosing an application service provider are increasingly taking aclose, hard look at the service-level agreement, also known as an SLA.

    Although the conceptis hardly new – for years organizations have created detailed SLAs fortelecommunications and consulting – the emergence of Web sites, intranets, andhosted software and services has clearly added a new wrinkle to the equation.

    “As a greaternumber of companies turn to outsourcing key systems and software, there’s agreater degree of risk,” notes Kneko Burney, director of e-businessinfrastructure and services at Cahners In-Stat Group, a Scottsdale, Arizona,market research and consulting firm. “If a service fails, it can provedevastating to the business. It can cost a company a lot of money and endangerthe relationship with business partners and customers.”

    The service-levelagreement can protect both sides in what is usually a complex business deal.“The reality is that there’s no such thing as 100 percent availabilitybecause no fail-safe system exists. But a company must know with a high degreeof certainty that it can conduct business online and that it isn’t going toincur constant interruptions or performance problems,” says Greg Blatnick,vice president at market research firm Zona Research, Redwood City, California.

    A September 2000survey conducted by Zona Research found that ASP service-level agreementstypically focus on guaranteed uptime, guaranteed bandwidth, guaranteedapplication performance, SLA indemnification, and interapplication communicationguarantees.

    A companycontracting with an outside provider must pay close attention to the details andintricacies of the business arrangement, says Burney. For example, what appearsto be a guarantee of 99.9 percent uptime might not be. “The provider mightguarantee that their service will be running 99.9 percent of the time, but ifthe backbone goes down or a company’s ISP fails, they might be off thehook.”

    According to Burney,it’s crucial to define terms and conditions up front. Using exact measures anddictating specific remedies and actions can prevent conflicts and confusionlater on. In most cases, the service provider should guarantee backboneperformance and bandwidth; if it experiences a problem, it should addressremediation with its own providers rather than leave a customer marooned.

    Most SLAs dictatethat the ASP, ISP, or hosting service cover only the direct costs associatedwith downtime or performance lags. A few offer escalating penalties that aretacked on to the credit for the monthly fee. Of course, all this necessitates asystem in place to actually measure performance.

    “At the end of theday, it’s the customer that feels the burden and who feels the pain if an ASPor other outsourced service fails to meet requirements,” says Dave Boulanger,service director for consulting firm AMR Research. Nevertheless, he believesthat organizations must understand their own needs and acknowledge thatoutsourcing technology really comes down to a partnership, where both companiesmust work together to solve problems. “It’s not as simple as a traditionaloutsourced agreement,” he notes. “There are many more variables and issuesto deal with.”

Workforce, February 2001, Vol80, No 2, p. 52  Subscribe Now!