Vendors of any application or service delivered via the cloud realize that they are often held to a higher standard than on-premise software providers, but in the wake of recent outages at Intuit, the issue of what vendors in the space must do to address downtime events - planned or otherwise - has been pushed into the spotlight.
As included in cloud-vendor service-level agreements (SLAs), "uptime" is a critical metric and there are few vendors, if any, who promise less than 98 percent uptime. Should those levels dip or in the event of any unplanned or sudden downtime or outages, there's usually some level of compensation on the part of the vendor - monetary or otherwise.
This is usually standard practice, since most vendors offer a subscription model for hosted or pure cloud-based services.
Most cloud vendors will admit that 100 percent uptime, while ideal, is not likely for the foreseeable future, as most schedule several hours of downtime over the course of a year for periodic maintenance and upgrades.
In the case of QuickBooks parent Intuit, the company experienced unplanned service outages in June that made headlines, and another in July that was more short-lived. For two days in June, Intuit.com and some of its major Web sites, including QuickBooks Online, Intuit Online Payroll, Quicken and QuickBase, were down for two days following a power outage.
Mountain View, Calif.-based Intuit has more than 300,000 customers using its online network of small-business applications, and has been moving more of its products online as part of its Connected Services strategy.
The company explained that the outage occurred during routine maintenance. An accidental power failure at that time affected Intuit's primary and backup systems, taking a number of Intuit corporate Web sites and services offline.
Intuit president and chief executive Brad Smith apologized on the company's blog for the outage, saying, "There is simply no excuse for having such a negative impact on you."
In July the company experienced another, albeit shorter, outage that knocked out service for a few hours.
During that time, company spokesman Rich Walker said that customer communication was key. "Over the past month, we've communicated the news of our outages, updated our customers when we restored services and re-assured them that their data was not lost. Details have been available through various communications from our executives and on our community sites," he said. "The outages were not related, and we have learned that we need to accelerate our continuous journey to re-architect our products, modernize our technology to host them, and build or lease state-of-the-art data centers to house our online services."
As to their efforts to help prevent such downtime events going forward, Smith stated: "Efforts are already underway. Many of our new online products are architected and designed with 24-7, high-availability and disaster recovery in mind. We've already invested more than $300 million in new facilities, including two state-of-the-art data centers in Quincy, Wash., and Las Vegas. We are aggressively migrating our existing applications to these data centers."
PLAN OF ACTION
Other online providers to the accounting space realize that they too are not immune to outages, but between proper communication and detailed SLAs, they are confident that they can retain customer loyalty and faith in their offerings.
Also in June, San Jose, Calif.-based cloud accounting and financial management service Intacct Corp. experienced a hardware error that brought down one of its databases. The downed database housed demo accounts, not paying accounts, but a portion of customers were offline for 24 minutes.
In October, there was a systemwide outage that severely hampered access for 26 minutes, according to Haris Hadjiioannou, Intacct's vice president of operations. He said that customers were informed in both cases and they were back up and running in under 30 minutes.
Hadjiioannou said that in the company's 10-plus-year tenure as an Internet-based business, it's no stranger to downtime, but it has worked to limit those occurrences and is willing to make bold service promises to its customers.
A few years ago Intacct enhanced its SLA, and titled it the "Intacct Buy With Confidence Agreement." The agreement states that the company is "so confident in our IBM-operated data center that we offer subscription credits if availability falls below 99.8 percent." What's more, "For every percentage point that service availability falls below 99.8 percent, they will receive up to a maximum of 50 percent of the applicable subscription fees for that month."
Some Internet service vendors don't even offer an SLA, let alone one as "bold" as Intacct's, explained Hadjiioannou.
He stressed that "clear, honest communication" with the over 3,300 companies that currently use Intacct was just as important as trying to ensure the shortest amount of time for unplanned outages.
Intacct has a "status" page where the company allows all customers, as well as the public, to access its uptime. They also have a customer portal to inform users of any planned downtime events, as well as status updates when service goes down.
Other vendors targeting the accounting community, and CPAs in particular, such as cloud-based workflow management software provider XCM Solutions, agreed that keeping their customer base informed about planned, and during unplanned, outages is the key to customer retention.
In April, during the height of tax season, the Boston-based company experienced an outage for approximately 30 minutes. Although a relatively short amount of time, it was "a significant event," according to XCM chief executive officer Mark Albrecht. "This was the first unscheduled downtime for us in seven years. Were our clients impacted and inconvenienced? Yes, but we handled it well. Every company has to plan for the worst and we do; unfortunately, one thing that caused [the outage] was a thing you would never have dreamed of. Communication is the key and what we've done here is we have a primary contact with our client base and we knew of the problem in a minute so we were able to get that message out there and get in and fix it."
XCM chief technology officer Gary Spaulding explained that continued disaster planning is important for any cloud vendor, but there are currently no guarantees that an outage won't occur.
"You have to be aware of the amount of resources you have to address your expanding features and user base. We have to know when we release new features, we know the effect certain features will have on the whole system," said Spaulding. "To be fair, when sites get overwhelmed. you don't always accurately predict what the strain will be; a lot is trying to plan ahead."
For cloud accounting and ERP veteran NetSuite, over 1 million visitors access its Web site daily. The company is also no stranger to outages, especially during its development stages. However, the use of dashboard technology to allow users to monitor uptime has been a competitive advantage, according to acting chief information officer Doug Brown. "We have a number of success stories where people lost their on-premise systems due to a flood or fire and still got up and running in a short time using us, but whenever we have a downtime event, you know it almost immediately," he said. "You need an open line of communication to the user base; once the event has subsided, you understand what happens and make sure it doesn't repeat."
In February, NetSuite experienced a 20-30 minute outage that related to issues in its cache and authentication layers, but Brown said that this was immediately communicated via its status site.
"We have built things into our architecture and it's in our DNA to manage these things," said Brown. "When we did the SLA it was a big victory for our sales team; we do what we have to do to win those customers and we stand behind our service as well as our software. All of our customers are on different servers so they have their own view of uptime and downtime, but I can say that our uptime has increased year over year."
