Last year, uber-analyst Esteban Kolsky and I did a research project to better understand cloud computing's uptake and related issues, and last week Financial Force, our sponsor, made the results public.
The findings are interesting to me because they reveal a more or less typical adoption cycle for cloud, by which I mean that some of the downstream effects are only gradually becoming apparent.
One of the appealing parts of cloud computing, aside from its favorable cost model, is its consumerized packaging -- people have the expectation (and rightly so) that they can get up and running with little fanfare and not much input from IT.
Cloud has been a liberating force throughout business for that very reason. Back in the day, words that rolled off the tongue a little too easily were, "Our system won't let us do that," but that's a thing of the past, and cloud computing is responsible in large part for the shift.
Check Your Baggage
What's less obvious though, in a world where lunch is not free, is that the cloud comes with its own baggage too. We found that the vast majority of companies surveyed had many cloud apps, and nearly half (46.3 percent) were using four or more.
With that comes some subtle issues. For instance, usually a cloud app doesn't exist in a vacuum; it needs to be integrated, and with lots of them it becomes something of a logistical challenge getting data to where it can do the most good. Also, each distinctly different app can come with a different cloud and tool set or at least user interface for maintenance.
For a long time, I've been articulating that the platform is the new competitive software battleground because it determines what works easiest with what.
Platforms provide a degree of standardization for app builders in the same way that a motherboard imposes a set of standards on what can be plugged into it. To be clear, there are many ways to get apps to communicate, but in many situations, when two apps have been built to the same platform standard, life is quite good.
Standards and Practices
Through this research I developed new empathy for the folks in IT for many reasons. Old IT was a reflection of the technology it supported. In the mainframe era, it was bureaucratic, rigid and limited, and many people associated these attributes with the people of IT rather than their jobs.
In new IT, the people responsible for making cloud computing work are quite the opposite. IT has learned to team with line-of-business units, to be supportive of business initiatives, and to do what it takes to make the business successful. Therein lies a new issue.
It seemed to me that, at least in some cases, IT has been left to clean up after departments that go ahead with an implementation without necessarily accounting for the downstream eventualities that can cripple a project, such as integration and performance.
Somehow this all comes back to platform too. One of the reasons for the proliferation of disparate clouds in some organizations is that no one considers the implication of having three clouds for three separate apps rather than one cloud in common among them. It's no one's job at the line-of-business level where many decisions are made, and in these cases some businesses were slow to promote platform standards.
The result is that at the front end, the lines of business run well with their cloud systems, but often that's because the people in the engine room are working hard, or harder than necessary, to keep all the cloud balls in the air. Add a new job to IT's growing list.
The IT department has developed some good partnerships with the lines of business, and now it's time to use some of that goodwill to help set up standards and to show the departments some best practices.
Other issues worth considering -- there are more, but you'll need to download the report -- include cost considerations and performance, and to a great degree, they come down to platform too.
Briefly, you'd think the arrival of cloud computing would have banished concerns about cost, but that's not true any longer -- if it ever was. Many organizations have capped IT spending, or nearly so, making it difficult to find thousands of dollars in a budget for new systems in order to save millions by replacing old systems. It's crazy but true.
Speed of implementation is one of those areas you don't see advertised or even written about very often, but it is becoming a real concern, at least for cloud vendors looking to generate initial revenue from a customer. Of course, platform is a big factor in deployment, especially for the time to deploy the next app on the platform.
Performance is right up there as a customer concern too, and platform is important here as well, but so is the speed of the Internet connection and the number of integrations running. While most of our study population had "more than four" cloud apps, it's not unreasonable today to see companies with a dozen or two cloud systems running the business.
At some point, every business will have to face the need to rationalize or refactor cloud deployments -- it's a natural part of cloud's success. This reconsideration should happen around your business's platform strategy. If you don't have a strategy, it's a good time to develop one.