Why does calculating cloud ROI remain so tough? Examine the tactics used by two cloud innovators, GE and Airbnb -- plus our exclusive survey.
and future capital costs. A third of respondents consider staff savings a top-three priority.
-- One third of respondents using infrastructure-as-a-service say it delivers better results at lower cost than in-house IT, and 20% say they're equal. The rest (47%) say IaaS costs are higher, the results worse, or both.
--For software-as-a-service, 37% of users say it delivers better results at a lower cost than in-house IT, while 26% say they're the same. That leaves 37% who believe costs are higher, results worse, or both.
--Forty-five percent of respondents whose companies use cloud computing say they have a formal policy or stated preference to evaluate cloud for any new application, while another 38% say their companies are headed in that direction.
--Fifty-two percent of companies say they're using SaaS, up from 40% in our Cloud ROI Survey two years ago. IaaS use is up to 38%, compared with 25% two years ago.
What's clear is that cloud use is growing fast, even as the ROI calculations for it remain complicated and fuzzy. What follows is a deep dive into how two companies, Airbnb and GE, are thinking about cloud computing, with additional survey data on how other companies are approaching ROI calculations.
Airbnb: Born in the cloud Since its founding in August 2008, Airbnb has gone from an idea that couldn't possibly work -- helping people rent out their apartments and vacation homes to total strangers over the Web -- to a business that executes 4 million bookings a year. Today, Airbnb, despite the occasional bad PR about a trashed apartment and looming regulatory concerns, is making it work. When soccer fans traveled to Brazil for the World Cup, for example, 120,000 of them booked rooms through Airbnb. Private room owners serving as hosts pay a 6% to 12% booking fee, plus 3% for completing the credit card transaction.
What started out as three air mattresses leased out of its founders' loft has grown into a business valued at $10 billion by its venture capital backers in April. In 2011, when Augustine started with the company, it used 20 to 30 nodes on Amazon's EC2 cloud service. Now it's somewhere north of 1,000.
The cloud's value to Airbnb is its ability to let the company scale up its capacity quickly for the next round of business expansion. Under its "Belong Anywhere" rebranding campaign begun in mid-July, Airbnb is trying to shift its image, from being an impersonal way to find a room for a trip to being the means of finding just the right host and unique lodging experience. As part of the campaign, hosts and travelers can customize and share the new Airbnb "Bélo" logo -- so a host can change the colors and use it as wall art in the room they rent, for example. The company set up a new website, Create Airbnb, for that campaign. Augustine prepared Airbnb systems for a doubling or potential tripling of traffic, and he had less than two weeks to gear up for it.
"We talk about it all the time internally -- our speed to market and how important it is," Augustine says. "It's incredible how fast our internal systems can be built out." Being able to quickly increase the number of Amazon Web Services' large C3 servers using automated scaling makes right-sizing the infrastructure quite a bit easier than it would be with in-house servers. That kind of automated scaling -- letting Amazon add servers to meet whatever demand comes -- scares some cloud customers, because it could lead to runaway costs. In the InformationWeek Cloud ROI Survey, 89% of respondents using or evaluating cloud computing say they're somewhat concerned, concerned, or very concerned about runaway cloud costs. At Airbnb, that worry takes a back seat to meeting customer demand and quickly testing new initiatives.
Airbnb has 100 software engineers creating and supporting booking, marketing, and engagement systems. For IT operations, it has only four to eight people, because it's not actually running the data center and related hardware. With only that handful of ops staff, Airbnb has expanded into Australia, Asia, and Europe, buying up a German competitor earlier this year and expanding its own European services. "If we had to build our own infrastructure, it wouldn't have happened so fast," Augustine says.
In our Cloud ROI Survey, respondents give cloud a clear advantage in speed of delivery. Only 12% of respondents using IaaS think their IT organization provisions infrastructure faster than a public cloud service, while 24% say they're comparable. That leaves 64% saying the cloud is usually faster.
Airbnb isn't oblivious to cloud costs, even if its executives aren't interested in the specific details of cloud ROI. Airbnb can take that lax attitude in part because cloud providers keep cutting their prices. In March, Amazon announced its 42nd price cut since launching its AWS cloud service, reducing the price of its M3 servers by 38% and its Simple Storage Service by 68%, a move that followed equally steep cuts by Google for its Compute Engine and Microsoft for Azure. Companies like Airbnb that had justified adopting cloud services at earlier pricing levels now have even less interest in ROI metrics.
"Forty percent is a huge number. That definitely made a big difference in our bottom line," Augustine says.
GE: Cloud gets more complicated Most business technology leaders will look at Airbnb with envy -- no legacy applications and infrastructure, and a business built around a single online function. On the opposite end of the legacy IT spectrum is GE, whose businesses span jet engines, medical devices, locomotives, power plant
Charles Babcock is an editor-at-large for InformationWeek, having joined the publication in 2003. He is the former editor-in-chief of Digital News, former software editor of Computerworld and former technology editor of Interactive Week. He is a graduate of Syracuse ... View Full Bio
Multicloud Infrastructure & Application ManagementEnterprise cloud adoption has evolved to the point where hybrid public/private cloud designs and use of multiple providers is common. Who among us has mastered provisioning resources in different clouds; allocating the right resources to each application; assigning applications to the "best" cloud provider based on performance or reliability requirements.