A strange thing happened on the way towards a mature, modern, and predictable practice of enterprise IT: The Internet broke out and changed just about everything about technology. It also upended the whole model of IT, from architecture, software development, and service delivery to business models, rates of change, and even the very mindset behind IT itself.
Or least, that's what is now happening today in most organizations and the results -- and usually the lack thereof -- have been sobering.
Enterprise IT used to be about the application of technology in support of the business, usually to automate processes, making them cheaper, faster, and better -- usually in that order -- or increasingly, to communicate, collaborate, and engage as a business with the world. IT generally wasn't about rethinking the business itself, but how to better support it.
But today, for most industries, IT has become the business. "We're all technology companies now", the mantra goes. Technology enablement is still important, but because of what today's technologies and a truly pervasive global network makes possible, the biggest value proposition of technology has shifted to the new, deeply superconnected, scalable-to-billions, always-on ways of working that digital business now represents. The upshot: Nearly everyone today is connected to everyone else in the world 24/7, with the devices in their pockets and purses.
As I've made the point many times before in my talks and research, the top priority for enterprises is now to figure out what to do about such a historic and truly profound capability. We now have essentially infinite connectedness, infinite economic capacity to tap into, and an endless pool of innovative ideas we can capture and try simply, easily, and cheaply until we hit the mark.
What will we do with this world-altering capability? Too often, largely ignore it and go about our old ways of working and thinking.
Taking a look at our digital competition, we can clearly see what Internet startups around the world are nearly always obsessed with doing: Tapping into and taking full advantage of our newfound global digital connectedness. This then is where the future is coming from today: The Internet has proven itself countless times as that uber laboratory of innovation, trying -- quite literally -- millions of new ideas in scale, continuously finding the way forward on how best to create value over networks, sustain it, and to relentlessly find new models that have the least friction, cost, the highest velocity, most agility, and best ability to tap into shared innovation.
From this we can begin to see that what we have often failed as enterprises to learn sufficiently from this vast innovation stream.
While the enterprise has steadily moved towards a model where most core IT has rolled up into a few giant suites of software from a few large vendors, the Internet has gone the other direction: An endless stream of new connected applications and services that change the game. While enterprises continue to stand up new silos, the Internet breaks them down with APIs, feeds, simple integration standards, and developer networks. While enterprises still struggle to simply translate a few key apps onto mobile devices, developers on the Internet have filled app stores with literally millions of new apps that redefine what is possible on a daily basis on these powerful new endpoints. The list goes on (and on) and you can see in the diagram above how very different and divergent these two worlds have become.
As I pointed out just recently, even simple things like search is still badly broken in most organizations, much less regular application and OS upgrades. There is usually considerable pent-up demand for more and better technology solutions, but users are already tired of IT trying to stay on top of the fundamentals. In hot areas like mobile, backlogs are rapidly piling up in an way that's truly unsustainable.
Most IT departments fully realize this highly challenging state of affairs, but are just now coming out of spending control mode for a half-decade or more since the great recession, just when technology has become the most fundamental and strategic to the business itself.
But simplistically increasingly budgets to make up for lost time is not likely to be the answer either, at least not at first, especially as budgets stay relatively flat despite the big shift to digital business. Because what is almost certainly at the root of today's challenges with absorbing and leveraging technology for the business is the inherent limits of the traditional 20th century model of IT itself. In most organizations, IT is still a highly centralized, command-and-control, hierarchical, and bureacratic department, bogged down in backlogs, operational issues, and too focused on technology to the exclusion of the business, just as both are becoming one and the same. It should be clear that this has to change, and in a significant way, before the situation becomes untenable, as it is in some organizations already today.
Nearly 10 years ago here on ZDNet, I challenged enterprises to become less sclerotic, more nimble, and agile, and less wasteful by adopting cutting edge models emerging out in the wilds of the Internet. Even then it was obvious that these "new techniques might actually be accelerating away from existing practice." But for the most part, organizations haven't widely adopted the techniques the startups use every day to create and sustain some of the world's leading products.
There is some good news in that agile methods, lean processes, and new practices like DevOps have indeed made their way into organizations -- the data says as many as 9 in 10 new projects now use them -- and made a real difference in a great many traditional enterprises. But this is happening only in hot spots, largely effected only a few areas in IT, such as software development or system operations. This needs to happen throughout the business, not just in IT.
To my mind, a big part of the problem is that technology experimentation in the startup world has a much higher expectation and tolerance of failure. The rule of thumbs is that only 1 in 10 tries will result in a significant win. However, this is largely considered an untenable failure rate in the enterprise world, at least the way IT is structured today
Or is it? Is it better to limp along with the poor outcomes we get when just about every IT project will be accepted, as long as it's not a total failure? Or is it better to fail fast and fail cheap until the best solution is created? This is a model that has been very hard to adopt for traditional IT, where most projects are expected to succeed and get the answer largely right from the outset, with a few subsequent iterations to file off the rough edges.
Instead, as I look at IT organizations around the world, I see broad divergence from the highly successful models being carried out in the world's largest technology laboratory (the Internet) and what most IT departments are doing. Examples of highly divergent models now include a common information landscape (the Internet -- for the large part -- uses a single flat and deeply linked structure, and enterprises do not), continuous competition and on-upping between services to gain users (enterprises just hand IT out, and will then revisits the decision only on occasion), a top-line emphasis on search and discovery (rarely a significant priority for most enterpries.) And the list goes on and on (again, see diagram above).
Frustratingly, none of this is particularly new information. But IT has changed very slowly. Thus, it is likely the Internet will eat the business world. And certainly, while we need to move and/or adapt much of what works so well about the Internet inside our companies, and quickly, it's still also true that the enterprise is not the Internet, nor should it be. Instead, the move to new models of IT must be about adapting external successes and powerful new models -- much better ways of creating and maanging technology solutions -- to the local needs and realities of our businesses, going well beyond 'bolt-on' digital transformation.
So why don't more organizations adapt new and more effective models? There are two main reasons that I've seen come up time and again: One, the capacity for IT to change enough to meaningfully close the gap requires a very significant additional 'impulse' to get over the hump in order to effectively remake the technology landscape with sufficient speed and focus. And 'significant' at this point is likely being a short-term 2-3x increase in investment, a non-starter in most organizations, even as technology is becoming critical and fundamental to just about every facet of business.
The second reason is that the IT industry itself has not reached overall consensus on a new, less centralized, more federated model of IT where much of the basics come from the cloud, while strategic capabilities are built into a new and properly sized digital business unit. In other words, there are few successful examples of traditional companies turning into technology firms. That's not to say there aren't success stories, such as Houghton Mifflin Harcourt or Telstra, but they are still quite limited in their scope.
We are entering the age of smart machine and the cloud which will together push both knowledge work and operations, especially for IT, outside of our organization. So too will attempts to rethink -- and then rapidly remodel -- our organizations much more around a technology core, which hundreds of organizations are now attempting to do with fast-growing new role and purview of Chief Digital Officer. This will almost certainly upend the responsibilities of the CIO and everything below it. But not if IT leaders take sufficient and effective action.
In fact, I now suspect it's likely that only a few organizations will be lucky enough and determined enough to avoid the coming creative destruction that will close the gap between where the rest of the world is going with technology, and where most enterprises today are heading. But it doesn't have to be this way.
Cloud computing: CIOs are taking back control, but is it already too late?
Research: CIO relationships and priorities remain conflicted