It says something about cloud computing when the NY Times, not often an information technology bellwether, not known for surfing on the bleeding edge of technology, publishes on its business pages an article on how mainstream companies are slowly but surely examining and adopting cloud computing. That the NY Times Business Section has noticed cloud computing is a clear indication that The Cloud has risen, in the mind of general management, beyond awareness and interest to the level of general business management discussion and decisions.
Action, one way or another, can’t be too far ahead. What will established companies be looking for? What must cloud computing providers do?
Companies considering cloud computing range from old and established like Eli Lilly to the new and working on establishment like Arista Networks. Larger companies with their multi-layered architectures, their extended labyrinthine networks of legacy applications and data farms and user communities have many obstacles to widespread adoption of cloud services. Many concerns have been expressed including conversion paths, conversion costs, data speed, data failures, responsiveness, privacy, security, recovery, long-term costs and hidden costs. In the days of mainframe-only computing I was asked “do you know the definition of an IBM elephant?” The answer, “it’s a mouse… with features.”
To date start-ups or smaller companies have been buying cloud services either to run all their applications or, just as often, to run some of the services like e-mail or, as in the case with Netflix, the web-presence portion of IT. Greenfields IT is relatively easy. But these companies haven’t got the scale and complexity of IT infrastructure that a Fortune 500 sized company has.
There is a wide array of cloud service providers. Expected names like IBM and Google are in the mix but Amazon, on first blush a surprise, is an industry leader. There are less gigantic players like Rackspace and Terremark. There are “public” cloud providers where the services are only available outside a company’s network. To the purists this is the only cloud that is a cloud. IBM and Acadia, a joint effort by EMC, Intel and Cisco, are offering so-called “private” cloud services where the web-based, on-demand capabilities are privately held or within a privately managed and controlled bubble.
Working on Step 2
Cloud computing providers are now reaching out to major corporations in an effort to provide core computing services to them via the cloud. Their relative success will depend on their ability to discuss, offer and deliver their cloud services less as a “transformational new paradigm in information technology” and more as a solid, tested, and trusted component of effective and efficient business operations.
To the extent that cloud computing significantly penetrates mainstream computing in large businesses providers will have to convince prospects that they understand the needs and motivations of mainstream business. Said in the framework of Everett Rogers’ Diffusion of Innovations, businesses that are innovators that are excited by the next new thing are already using the cloud. So are “early adopters,” companies that through need or culture like to be on the leading edge. For cloud computing to find a major market it must “cross the chasm,” in Geoffrey Moore’s “Crossing the Chasm” terminology, and appeal to the “early majority” of main stream businesses. They buy on different signals and in different ways.
The Early Majority are pragmatists. They care about the company they are buying from, the quality of the product they are buying, the infrastructure of supporting products and system interfaces, and the reliability of the service they are going to get. Pragamatists tend to be 'vertically' oriented, meaning that they communicate more with others like themselves within their own industry than do technology. A standard question: “Where else have you done this for someone just like me?”
There are four fundamental characteristics of visionaries that alienate pragmatists. When planning and executing a marketing strategy, when engaging prospects in sales efforts, cloud computing services should avoid all four. They include:
1. Lack of respect for colleagues' experiences.
2. Taking greater interest in technology than in their industry.
3. Failing to recognize the importance of existing product infrastructure.
4. Overall disruptiveness.
I have worked in multiple very innovative software companies and came to dread the meetings where the founder/chief architect/head visionary is on a sales call and forgets to listen to the prospects. Instead, they frequently listen for a bit and lapse into how all encompassing the vision is and how smart we were to provide it to solve your problems, whatever they are. Implicitly we were also saying how behind the times the prospects were if they didn’t get on board. We would get smiles and encouraging signals of further consideration. The near universal back-channel reaction was that the prospects were plenty smart and plenty experienced too. They have a number of issues and we didn’t take the time to grasp and deal with them. We should come back when we’re prepared. Similarly, cloud providers who are in a race to establish market share may very well win that race through patience and listening.
A difficulty about providing technology is that evaluation discussions frequently become about the technology. It is important to know and measure scalability, be able to show how and if you use the Rete Alogrithm or not (for 50 bonus points, without using the web, what is THAT about?), and have a discussion of file size compression for rich media content. This class of issue has its place, and, since sales to the mainstream involve consensus across the prospects functions, they will be covered. But, these issues describe how value is delivered, not what the value is. In the end the winning providers in mainstream competitions will understand and communicate about in-this-industry net value. By net value I mean the value net of vendor-risk (can they support us, are they strong enough to stay with us, if something goes off the rails can they recover?).
Mainstream corporate IT has a ton of legacy, as well it should. The legacy systems and underlying infrastructures represent the collective, and sometimes forgotten, knowledge and wisdom of thousands of people over many decades. Legacy applications are rarely pretty, in IT terms. Within an enterprise they yield multiple systems architectures, multiple sets of underlying operating and data management software, multiple hardware platforms and configurations. Successful mainstream cloud efforts will work with these realities, align with them, adjust and change them where possible over time. Watch out for the cloud provider that attempts the large-scale, 100% conversion in an accelerated timeframe. It will look good for the business case, certainly look good on their income statement, and almost certainly fail under the pressure and complications.
Vendors of cloud services are in a difficult place. They need to introduce a new way of delivering, consuming and paying for computing to the mainstream market that buys consistency and reliability and distrusts change. The apparent economic advantage of cloud computing, it is generally priced on the margin, is the carrot. The winners will be able to communicate and deliver what the buyers view as risk-adjusted superior value. My personal expectation is that they will but only to the extent they put on their customers’ shoes and walk around in them.