Calculating the cost of downtime

Although it's no secret that avoiding Web downtime is a number one priority for CTOs everywhere, many are still toiling to turn the costs associated with preventing downtime into bottom-line savings for the company.

To determine precisely how much a crippled Web site may or may not cost a company over time, a CTO must consider a variety of accounting methods. Producing accurate estimates may help justify important decisions and expenditures, including outsourcing and technology costs.

Identifying potential losses

First and foremost, when determining which metric will best quantify the consequences of online business loss, CTOs need to consider which aspects of their business are most vulnerable to lost seconds, minutes, and hours, says Sean Moriarity, vice president of technology at Pasadena, Calif.-based TicketMaster Online-Citysearch Inc.

Depending on how active a company's site is throughout the day, examples of business losses following an outage -- in many cases irretrievable losses -- include missed sales opportunities from advertisers and customers, damaged credibility, and brand name degeneration, according to Donna Scott, an Austin, Tex.-based analyst at Gartner Group Inc.

Downtime cost consideration checklist

* The greater the competitive environment, the greater the cost of downtime

* High customer loyalty lowers the probability customers will switch to competitors

* The greater the proportion of impulse buying, the greater the likelihood a direct sale will be lost due to downtime

Source: Donna Scott, Gartner Group

Downtime planning

Take these steps when determining the most cost-effective way to battle downtime.

* Estimate lost revenue and opportunity costs

* Uncover additional hidden costs

* Closely examine outsourcing agreements

"For the content sites, it creates a bit more work, because [outages are] not as clear to the bottom line," says TicketMaster's Moriarity. "It's a little different when Yahoo has an outage vs. eBay, for the same reason that [it's different] if ABC News went off the air vs. the Home Shopping Network. One is eyeballs; the other is transactions."

Moriarity says a common-sense, standard ROI approach typically serves a CTO best for prioritizing sensitive areas that could be harmed by Web site downtime. TicketMaster Online-Citysearch has set the importance of its three discrete business units in this order: ticketing, then online personals, and finally the Cityguide business. In this company's case, ticketing is the most crucial environment given that Ticketmaster Online-Citysearch is paid per ticket sold online.

"For each ticket [we] don't sell -- and we sell a lot at peak -- that hits the bottom line," he says. "When you live with these business realities everyday, generally the answers are very obvious."

Looking for hidden costs

Lost revenue may be the most obvious consequence of Web site downtime, but there are many other factors to consider, says KBKids.com CTO Shawn Davison. Other unpredictable elements that should be considered in the event of an outage or site glitch include customer confidence, investor interest, and negative publicity.

"Just because you can get to the home page, doesn't mean you can buy [an item]," says Davison. "The longer you're down, the more difficult it will be to put a cost associated with downtime."

Based in Denver, KBKids.com is a wholly owned subsidiary of Kay Bee Toys that resulted from a joint venture of Brainplay.com and Kay Bee Toys in June 1999.

Davison says KBKids.com relies on third-party Internet performance monitoring firms, such as San Mateo, Calif.-based Keynote Systems, for a "good, non-biased perspective" on performance and availability to acutely measure daily activity on the Web site and to help pinpoint factors affected by potential increments of inactivity.

Your company will incur more costs if it issues redeemable credit to customers as compensation for unavailable or slowed Web site activity. That can be a tricky proposition, depending on the business type and customer expectations. KBKids.com does not institute a credit policy.

"It's too difficult to know who's being impacted at that moment when you're experiencing downtime to set that blanket policy," Davison explains.

Quantify outsources responsibility

Outsourcing Web mission-critical applications could be a solution for companies willing to let an outside party bear the responsibility of maintaining a site's uptime and performance levels. When factoring potential downtime cost into that relationship, however, be aware that there may be negative effects.

The solution "isn't how much does an outage cost you and how much do you outsource. That implies outsourcing delivers a higher level service than you can yourself," Moriarity says. "To turn over that responsibility to a group that doesn't have the depth of talent to their business is dangerous."

He says Ticketmaster-Online Citysearch outsources very little aside from Akamai's content delivery service.

Whether involving applications, storage, or staff expertise matters, the option to implement operations in-house versus outsourcing boils down to assuming responsibility for setting and monitoring required SLAs (service-level agreements), Scott said.

"The businessperson doesn't want a thousand different service levels on each component of the Web site. They want one: availability and response time," says Scott. "[Companies must] take that top-down service level that's required and do bottom-step analysis to see if it's needed. The cost must be worth it."

Businesses need to understand application availability trade-offs, meaning high-availability applications that cost twice as much to purchase than non-high-availability applications may make sense for a Web site dependant on constant uptime.

As the CIO of SciQuest.com, an Internet-based organization in Morrisville, N.C., that generates 90 percent of its revenue from Web site transactions, Rob Fusillo says his company separates performance as quality of experience from availability as quality of service in order to quantify Web-site service loss.

That distinction helps Fusillo produce SLAs with his outsourcing partners based on availability -- not performance -- with several limitations, including loss of service resulting from failures of communications services outside Sciquest.com's control, from clients' faulty hardware/software, or from outages because of scheduled maintenance, he says. "A loss of service from one of our service providers means that our customer is not able to use SciQuest.com marketplace," Fusillo says.

"We have some form of insurance coverage, but basically service providers don't guarantee us service to any disruption above and beyond the service we pay them," Fusillo says.

CTOs must pay careful attention to SLAs with providers in order to assess how much their company is at risk to compensate for the costs associated with a downed Web site and the resulting loss of business.

Although a number of loyal customers will tolerate a company's unavailable Web site for a limited time, long-term or repeat inefficiencies could prompt frustrated users to migrate toward competitors' sites. When plotting metrics for calculating downtime costs, CTOs should ensure their business side knowledge of bottom-line impact directly reflects IT's willingness to assist in helping to make sense of those metrics.

This story, "Calculating the cost of downtime " was originally published by InfoWorld.

Copyright © 2001 IDG Communications, Inc.

It’s time to break the ChatGPT habit
Shop Tech Products at Amazon