IBM's Cloud Business: Ex-Employee Divulges Shortfalls

IBM's cloud computing revenues are smaller and less "cloud-intensive" than customers and Wall Street analysts might think. That's the claim of a former IBM employee who backed up more than a few of his/her critical assessments of the vendor's cloud prowess with a number of confidential internal documents shared with InformationWeek.

The documents put IBM's 2012 cloud-related revenue at $2.26 billion, a figure the company has declined to disclose publicly. In 2011, IBM did issue a roadmap that set forth the goal of reaching $7 billion in annual cloud revenue by 2015, so the much lower figure raises doubts about whether the company is on track.

IBM also said in 2011 that only $3 billion of that total would come from net-new business, suggesting that $4 billion would be tied to cloud-based ways of delivering its current hardware, software and services.

Noteworthy is data that shows that roughly half of current IBM cloud revenues are tied to hardware, in many cases systems used to run customers' private clouds or partner clouds. "This is not what your readers would think of as cloud," said the former IBM employee, who reached out to InformationWeek after reading a column I wrote challenging IBMto be more transparent about its cloud revenue. "They will think of Amazon EC2, Salesforce.com and IBM SmartCloud as real cloud. Not stuff sitting on their data center floor."

[ Will U.S. cloud companies lose customers because of NSA spying? Read NSA's Prism Could Cost U.S. Cloud Companies $45 Billion. ]

It's unclear whether IBM's re-characterization of certain hardware sales is a factor in a current Securities and Exchange Commission investigationof the company's cloud revenue, as divulged in IBM's second-quarter financial filings.

IBM declined to comment on the ongoing SEC investigation, but speaking on background, an IBM executive said the company "established the whole category of private cloud five years ago. We told the marketplace, and Gartner and IDC came to agree, that a lot of very large companies have gigantic data centers and a lot of stuff they want to control from a privacy or security perspective. Therefore, the first major opportunity from our client base was going to be us helping them build cloud delivery."

Where cloud compute capacity is concerned, the embrace of x86-based systems and standards is cannibalizing IBM's higher-margin mainframe and Power server businesses, says Kulbinder Garcha, an analyst with Credit Suisse, which downgraded its rating on IBM's stock on Aug. 6.

Meantime, emerging platform-as-a-service (PaaS) and infrastructure-as-a-service (IaaS) offerings are "obviating the need for IBM's traditional middleware stack," Garcha wrote in a research report. "The shift to cloud continues to present risks given IBM's technology positioning."

As for IBM's ability to compete with its own IaaS offerings, internal documents supplied by the former employee detail the formidable competition IBM's SmartCloud Enterprise (SCE) faces from Amazon Web Services. One document shows SCE to be generally less expensive than AWS offerings at low levels of service and capacity utilization, but the cost advantages shift to AWS at higher service and utilization levels.

For example, IBM's 32-bit Copper-, Bronze-, Silver- and Gold-level services are all less expensive than Amazon AWS at 25% utilization levels of one-year and three-year reserved capacity, according to this internal assessment. But move up to 64-bit services and 75% or higher utilization and the tables turn, with IBM's prices becoming 20% to 40% higher than AWS's and more than 40% higher in the case of 64-bit Platinum services.

These comparisons could be outdated, as the document is from March 2013, but the competitive pattern described is consistent with the reputations of both vendors, according to InformationWeek cloud computing expert Charles Babcock. "IBM is inclined to get higher margins once it has you hooked, while Amazon rather artfully has made it cheaper to become a big user, with the rate per hour going down the more you are committed to long-term use," Babcock says, noting that no cloud competitor has claimed to beat Amazon's three-year, Reserved Instance price.

Yet another internal IBM document outlines gaps in the company's cloud functionality compared with AWS's. The functional gaps include a lack of options comparable to Amazon Elastic Bean Stalk application deployment and management capabilities, Amazon Elastic MapReduce data processing capabilities, the DynamoDB cloud-based NoSQL database and the Amazon Redshift cloud-based data warehousing service.

Those are just five of the more than 20 gaps cited in the document, most of which exist because "the nature of Amazon's business and infrastructure required it to lead in cloud innovation. IBM's didn't," Babcock observes. While Amazon was steadily building AWS over the last six years, he says, "IBM did not extensively develop middleware for the cloud that aids application deployment and management. It did not develop a native NoSQL approach to data management. And IBM clearly missed the boat on cloud-based data warehousing, a spot where it could have excelled."

These differences might well explain why Amazon won a four-year, $600 million cloud contract with the Central Intelligence Agency (CIA) in February after a bakeoff against IBM. Amazon's bid was $54 million higher than IBM's, a point that led to an IBM protest and a General Accounting Office review that's still underway. Given these gaps, it's easy to imagine that Amazon's bid included a range of ready-to-use services that were just not available from IBM and that would have to be developed or sourced from a third party if required.

On Aug. 14, however, IBM announced it had won a $1 billion, ten-year cloud contract with the Department of the Interior. In this case, IBM's strengths in SAP application hosting and Unix, now available on the IBM AIX Cloud, were cited as important factors in the win. IBM is well positioned to bring legacy systems into the cloud.

Some Transparency Of Our Own

IBM didn't provide answers to a number of questions InformationWeek presented with key facts from the hundreds of pages of documentation shared by the former employee. The depth, detail, profusion of company acronyms, use of company presentation formats and citation of company locations and executive names strongly suggest that the documents are authentic.

The former employee, who says he/she was "resource actioned," says the motivation for sharing the documents is "protecting customers and ex-colleagues by getting the truth out," and the reason for doing so anonymously is fear of losing outstanding severance payments. InformationWeek's motivation is to shed light on the state of cloud competition not only with the likes of Amazon, but also Hewlett-Packard, Microsoft, Oracle and SAP.

As InformationWeek has reported extensively, HP, Microsoft, Google and others have been playing catch-up with Amazon.

IBM's $2 billion June acquisition of SoftLayer provides fresh evidence that the cloud assets and capabilities the company had before that deal weren't enough. SoftLayer data centers, its stack and its people will become "a foundation unit" in a new IBM Cloud Services division incorporating SoftLayer and SmartCloud Enterprise.

Clearly, IBM isn't simply trying to mimic AWS offerings. With SoftLayer it's trying to cater to enterprise customers. Where AWS virtualizes servers in the public cloud, SoftLayer offers the option of dedicated servers for enterprise clients that have requirements to run on the bare metal.SaaS Comes Up Short

The depth and breadth of IBM's software portfolio are well known, but with much of that software tied to IBM's hardware and middleware, it limits the potential for software-as-a-service versions.

According to an internal document that details IBM's position in the SaaS market, based on the company's own research, its top competitor in 2012 was Oracle, with $707 million in SaaS revenue, followed by Microsoft with $430 million. IBM placed itself third with $371 million in SaaS revenue, and SAP fourth with $347 million (apparently not counting revenue from SAP's Ariba acquisition). Salesforce.com, with 2012 revenue of $3 billion, surprisingly wasn't listed as an IBM competitor, though Workday, with $130 million in 2012 revenue, was included.

IBM's goal is to generate $1.5 billion in SaaS revenue by 2015, according to this document. While most competitors were listed as growing their SaaS businesses at triple-digit rates in SaaS (fueled largely by acquisitions), IBM's SaaS revenues grew 55% in 2012. More worrying, signings of new SaaS accounts fell short of goal, with the biggest shortfalls outside of North America.

Those shortfalls are "an early warning sign," says the former IBM employee, as the softness will carry into coming quarters. What are the reasons? Most of IBM's SaaS offerings are hosted in North America and "don't always appeal to customers elsewhere," according to the former IBM employee. What's more, this source says, "IBM salespeople are not aligned to sell SaaS. While the high-level executives want to promote SaaS, many sales executives are not incented to sell."

These are cloud growing pains common with other vendors focused on on-premises software and systems. It's just surprising that IBM, so esteemed as an industry leader, didn't address them earlier.

The software businesses IBM is in -- mostly infrastructure software and middleware -- aren't generally the ones moving to SaaS first.

For example, IBM for years avoided buying into enterprise application categories where it serves as an integrator for the likes of Oracle and SAP. For every dollar IBM might make on enterprise applications, IBM senior VP Steve Mills has reasoned, IBM stood to make as much as five dollars on implementation and customization services.

Now that on-premises deployments are losing ground to the cloud, integration services dollars are starting to disappear. IBM now lists Oracle and SAP as cloud competitors. No coincidence that IBM has started to diversity its software portfolio by, among other moves, partnering with SugarCRM and acquiring HR SaaS specialist Kenexa.

Awakening The Giant

One thing that the many documents shared with InformationWeek make clear is that IBM is acutely aware of its competitive weaknesses and threats, and that it's intent on acquiring what it needs, repackaging and evolving what it has and, in some cases, changing the way it does business.

In an internal document in which IBM shared customer assessments, one customer was quoted saying "no company I can think of is more difficult to deal with for contracting." This assessment prefaced a detailed plan to offer simplified, SaaS-only contracts that offer better "clarity and transparency" on terms, conditions and policies.

Recent acquisitions, including Kenexa and SoftLayer, signal that IBM is investing in the cloud, even if most of the company's $104 billion in total revenue will continue to come from elsewhere. Cloud initiatives matter to IBM and to Wall Street because it's where the growth is, reflecting trends in customer demand.

Will IBM reach the $7 billion in cloud revenue (including $1.5 billion in SaaS revenue) it wants to achieve by 2015? It will take more acquisitions to get there. Will the cloud be the kind of high-margin business that IBM is used to? Doubtful. Competitors such as Amazon, where CEO Jeff Bezos preaches investing for the long haul, will not make it easy for IBM to do business in the cloud the way it likes to do business.