In the survey of enterprises that I did this spring, I attempted to quantify the cloud opportunity in a different way. Like others, I'd looked at how much current IT spending might be shifted to the cloud, and the number has been consistently about 24% of the total, a bit over $250 billion per year worldwide. Impressive to be sure, but is that the right perspective to take? I asked enterprises to provide some data on the applications of cloud computing that didn't represent IT hosting shifts, stuff that they couldn't do at all today but could see themselves doing in the Age of the Cloud.
This turned out to be a good question, as proved by the fact that enterprises told me they couldn't answer it within the survey timeline. They'd not considered it before, they said. They proposed to give me their answer in the fall, and they came through. Their number? Almost three times as much cloud revenue could come from new productivity-based enhancements. This assessment, if true, signals a potentially huge change in computing. It's almost like there have been two clouds and we've been watching the least interesting and rewarding one.
For the last decade, enterprise IT has been mired in a productivity slump. Never in the history of IT have we not ended one cycle of innovation (what we might call "the Internet Age" was the last one) without starting a new one. In each new cycle, new productivity benefits drive IT spending to grow much faster than GDP. In 2002, we ended a cycle of innovation at the bottom as always, and we've never gotten off it. For all the hype about clouds and tablets and changes, IT spending for the last decade has simply tracked the "average" level of post-WWII growth. At the best times of the past, we exceeded the contemporaneous track of overall growth by more than 1.6 times.
The reason for this detour into statistics is that in our survey, we found that the enterprise estimates of the "new" productivity opportunities for the cloud would be almost exactly the difference between a "good" technology innovation wave and the current trend-line spending. A lot of vendors will make a lot of money if the enterprises are right.
Another fascinating truth about these numbers is that they're based on the presumption of greenfield apps to support productivity. Infrastructure- or platform-as-a-service doesn’t create a new app model, it just creates a new place to run old apps. The "true cloud" apps will have to be a form of SaaS, which means players like Salesforce may have a major leg up on players like Amazon or even Microsoft.
There's another really tantalizing question waiting out there: What are these apps and what exactly is the implementation model for the "new cloud"? I can't answer that one because enterprises aren't very good at strategizing things that nobody is currently offering. I do think I can read a couple of facts about this brave new world into their responses, though. The ones that stand out are mobility and personalization.
Enterprises say that the biggest change in their productivity paradigm for IT comes because mobile broadband lets them push computing support to the very place and very time that the need arises. We don't have to run back to our desks to research something, we simply ask a question. That implies a lot more intelligence in the application architecture. We don't run an app (Google) and ask it something. Instead we ask something and the cloud invisibly runs stuff to answer the question. We're never aware of the process or even the processes that support it.
The personalization stuff arises because people have different skill sets and practices, and if you're going to push cloud smarts into their laps at their moment of need, you better push it in the right way. This has to be more than just mashups to make a worker-specific GUI, though. It's clear that enterprises see the worker having an app-like or Siri-like interaction with an appliance, which then delivers information. This means that the cloud has to understand context. If you think about it, the information needed to make that possible is there. Location-based services (LBS) can tell the cloud what customer is being visited, the calendar appointment can say what’s likely to be discussed, the organizational chart (and past practices) can alert the cloud to who the worker might be expected to collaborate with to get an answer or move the ball on a sales call...you get the picture.
Another interesting point about this is that the whole new wave of cloud-ness could be an overlay on traditional IT apps, even ones running in the data center. The enterprise may have shown us how to create a cloud vision that can jump over the pain of trying to induce transition of current apps--something that will be done only if extravagant savings can be proven. The "old cloud" actually lowers IT spending by shifting it to a cheaper platform. The new cloud raises it dramatically by creating new benefits.
Everyone is wondering whether 2012 will be the Year of the Cloud. I’m wondering which cloud it will be the year of.
