Salesforce.com service hiccups again

15.02.2006
In the wake of several service outages, Salesforce.com Inc. plans to establish a Web site that the public can access to get information about service performance.

Last week, online CRM service provider Salesforce.com's system again suffered downtime. It was the latest of of several service disruptions over the past several months to affect the San Francisco-based company, which now has some 350,000 subscribers.

The outages have vexed customers and have given the company a considerable amount of unwanted media attention.

In response, Salesforce.com has been working to go online with Mirrorforce, a load-balancing system made up of three separate data centers that will have fail-over capabilities to prevent further outages. The Mirrorforce system is expected to go live sometime this quarter.

The most recent service crash was an 81-minute disruption last Thursday, CEO Marc Benioff confirmed in an e-mail yesterday. He also stated Salesforce.com intended to establish soon a public Web page called Trust.Salesforce.com that would detail performance and throughput statistics.

Four customers confirmed that they had lost service last week. One customer that was "severely impacted" for nearly an hour and a half was systems integrator Applied Computer Solutions Inc. in Huntington Beach, Calif. During the downtime, workers were unable to log case notes, access the technical documentation warehoused in the Salesforce.com system, or access customer contact data.

"The recent outages have caused my company to start looking for alternative CRM solutions unless Salesforce.com is able to stabilize their systems, and fast," said Jason Ellis, a customer support engineer at Applied Computer Solutions.

The company relies on the Salesforce.com service for both sales and customer support. "Being down a few minutes can be tolerated, but outages that can last 30 or more minutes can seriously impact our ability to perform our job functions," he said. "This can have a serious impact on our customer's perceptions of our company and our ability to provide them support," Ellis said.

"Clearly something is not right in Salesforce land," said Tom Kramer, president of Bella Pictures, a San Francisco-based wedding photography company that relies heavily on the service. Kramer seemed to be somewhat upset over the outage last Thursday, saying his company suffered from reduced productivity. Additionally, he said, Salesforce.com never contacted Bella Pictures to inform him of the outage. Kramer said the company seemed to be facing problems beyond anything connected with its switchover to the new Mirrorforce infrastructure.

Another customer, meanwhile, had a different experience. For staff at Phoenix Technologies Ltd., the Thursday outage "was not a significant problem," said Clifford Bell, the company's CIO. Milpitas, Calif.-based Phoenix makes applications and tools for the core systems of PCs and other devices, and it runs Salesforce.com for sales and marketing operations. After the crash, Bell said he spoke to Salesforce.com executives and received a written response directly from Benioff that detailed future plans to ensure that the outages would cease.

"Funny, but I never have gotten an e-mail from Larry Ellison for any Oracle issues or from Bill Gates when there are Microsoft issues," he said. "Believe me, I do have bugs from Oracle and Microsoft." Salesforce.com is working hard to correct the problems, he said.

"You have to wonder why this is happening," said analyst David Dobrin at consultancy B2B Analysts Inc. in Cambridge, Mass. "Is it an artifact of their changeover to a more robust, dual-data-center model? Is it a size problem? Are they having the same problems with their hardware/software vendors that other people frequently have?"

He noted that Salesforce.com's constituency is made up of customers that don't want to have to trouble themselves with information technology. These customers, therefore, will be more aggravated by downtime and more willing to express their dissatisfaction than those who rely on internal IT resources, Dobrin said.