One of the most -- perhaps the most -- influential books in Silicon Valley over the past two decades has been "Crossing the Chasm" by Geoffrey Moore. In it he posits the existence of a technology adoption bell curve (Figure 1) -- starting with innovators, who eagerly grasp new technologies to gain competitive advantage, through to laggards who typically wait for technology to be established as a service, thereby requiring no internal technical expertise.
The catch is that between the technology enthusiast (perhaps 15 percent of the market) and the pragmatist phases (at 85 percent, the vast majority of a market) looms the chasm. The chasm is caused by the fact that, while enthusiast buyers are willing to work with unfinished, unpolished products, pragmatists require vendors to deliver what Moore terms "whole" products -- easy to use, accompanied by training and large service partners, with local company presence for personal interaction -- in other words, convenient to adopt and use. If whole products are not available, new technologies cannot "cross the chasm" and are stranded in a small subset of the overall market for a given technology.
It's difficult to overstate this book's influence in the startup world. Countless discussions within early stage companies have focused on how to cross the chasm -- with enormous pressure on every participant to help figure out how to access the lucrative mainstream IT market. Without a strategy to cross the chasm, an early stage company is doomed to a niche opportunity.
Moreover, Moore's theory was a good model for how most enterprise IT shops worked. They slowly adopted new technologies, typically waiting for a winner to emerge in a given market segment. These organizations, in effect, stood with folded arms, impassive to the entreaties of startups asking them to give their products a try -- patiently waiting for one vendor to be ready to deliver the technology in the "enterprise" form.
Today, however, the crossing the chasm model is breaking down completely -- so much so that enterprise IT itself is crossing the chasm and heading in the direction towards innovation. This behavioral change is massively disruptive to the established order and accounts for much of the confusion and angst we see in the technology industry today. It's no understatement that there is a pitched battle going on for the very definition of enterprise IT, with one side represented by a coalition of incumbent vendors and mid-career IT personnel and the other by startups, open source companies, and IT personnel focused on leveraging new technologies.
Gartner characterizes this state of affairs as "bimodal IT" -- an uneasy melange of two very different approaches to IT.
Bimodal IT: Legacy IT and Innovative IT
The first approach: Legacy IT focuses on traditional proprietary applications, functionality and processes delivering largely standardized functionality (think ERP) from legacy vendors (think Oracle). This is the world of large, infrequently updated software packages, managed via deliberate, lengthy manual processes (think ITIL), operated in a largely static infrastructure financed via large capital investment, and often outsourced to a "your mess for less" provider.
The second approach: Innovative IT focuses on individualized functionality specific to the enterprise, commonly self-assembled from open source software (think Mongo) and products from early stage vendors (think Urban Airship). Change is frequent and managed via automated processes (think DevOps) and operated in dynamic, elastic infrastructures priced by resource use in an environment managed by a cloud computing provider (think Amazon Web Services).
Bimodal IT: Two Completely Different Worldviews
In a presentation by Gartner analyst Lydia Leong I attended at the AWS Reinvent conference, she walked through the characteristics of bimodal IT. The message I took away was that these are two completely different ways of looking at the world and, even if they live in the same organization, they are as different as chalk and cheese. Crucially, the two visions of how IT operates and what it needs to deliver are so different that the solutions of one group are considered completely inadequate by the other. Leong gave the example of private clouds -- built by traditional IT, they are viewed as uninteresting by innovative IT and typically ignored in favor of public solutions.
The mutual incomprehension of these two worldviews was on display in a Twitter exchange I had while attending Reinvent. Perusing the agenda, I noted that Coca-Cola was speaking at two sessions, and so tweeted (rather snarkily, I must admit) "Coca-Cola has two sessions at Reinvent. More evidence enterprises won't use AWS." I wrote this tweet (you can see it quoted in the middle tweet in Figure 2) because of the very common assertion I hear that "enterprises don't/won't use AWS," and this seemed very clear evidence of what I see all the time -- mainstream enterprises are adopting AWS and its "commodity" counterparts Azure and Google.
However, almost immediately, worldview one was displayed by a person (blurred in the image to avoid personalizing this discussion) who responded "What is Coke doing? Actually 'enterprise' type stuff, or basically more external facing?" The exchange continued (see Figure 3), with me responding that "external facing" is enterprise IT, and the other person asserting that enterprise IT is when "back office is there" and that, even if some companies are doing enterprise IT in the cloud, it's not "any company of size."
The technical term for this attitude is "cognitive dissonance" -- refusing to allow any facts to interfere with a perceived worldview. Coca-Cola, according to MarketWatch, did $47 billion in revenue last year. Not as large as, say, General Motors, but it certainly qualifies as a large enterprise in my book. This adamant, persistently negative viewpoint regarding public cloud computing makes it clear why bimodal IT is on the rise, and why approach two is driven not by traditional IT, but by corporate executive leadership and a new breed of IT executives attuned to the need to transform enterprise IT.
t's easy to understand the urgency of making this change. The disruption exploding throughout our economy is IT-driven and changing the very nature of customers and competitors. To take just one example, look at hotels. Five years ago, if you ran Hilton, you viewed Marriott, Hyatt, and Sheraton as your primary competition. Today, thanks to Airbnb, your competition is every spare bedroom on the planet. Your competitive landscape is completely different, and an IT organization that promised to support you with a lower-cost room reservation system is hopelessly unable to respond to this new (yes, external facing) landscape.
However, many corporate and IT executives recognize the need to change IT, but are unsure of what's needed. Here are a few of the items that should be on your list:
- Understand the requirements of the new application types. Microservices, PaaS, NoSQL and open source are the foundation of the new application paradigm. Your application architectures and componentry need a new approach. You're undoubtedly going to have to increase your vendor portfolio, disagreeable as that is to the decade-long initiative to pare vendor portfolios.
- Adopt new infrastructures prepared to meet the operational needs of these new applications. Insisting that public cloud computing isn't needed because most applications don't experience erratic load ignores the irresistible force of change. It's like insisting your growing teenage boy continue squeezing himself into a size 12 shirt because it used to fit. Public cloud for enterprise applications is in your future, notwithstanding my counterpart in the Twitter exchange. This isn't just hosting 2.0, so understand and adopt it aggressively.
- Accept your polyglot future. Java is the new COBOL, and is unsuited for the new application types, especially when combined with legacy frameworks. Even LAMP is showing its age, joined (or potentially supplanted) by MEAN -- MongoDB, Express.js, Angular.js, and Node.js. This isn't to say that you'll only use MEAN; it's to say that your future is going to be a lot richer and complex.
- Incorporate new cloud-based services. In a conversation with me a year or so ago, Leong made the point that the problem for most "enterprise" cloud providers is that they're competing with the AWS of 2010, not today's. At last year's Reinvent, AWS announced Kinesis, a real-time event processing service, which I regard as likely to have as large an effect in the Internet of Things (IoT) world as EC2 did for the computing world. This year it announced Aurora, a high-performance, highly-available, MySQL-compatible database service. It promises to provide incredible scale at a low price, enabling a new range of applications. As an enterprise, you need to track these developments and evaluate how you can take advantage of them.
- Implement new organizational types. Obviously, much of traditional IT is unsuited for this new world, so create new organizations prepared to meet it. I've been struck by the growth of innovation labs started by mainstream enterprises. I used to regard them as affectations, Potemkin Villages used by companies desperate to demonstrate their hipness credentials. I was completely wrong in this; these are places that companies are founding to foster experimentation and innovation. I've had the privilege of visiting a number of them and been impressed by the caliber of people and non-traditional thinking in them -- and the creative solutions being worked on there. Of course, not every lab or initiative is going to pay off, but it's clear that traditional IT is so mired in its historic responsibilities and solutions that it can't respond to the imperatives of the Third Platform -- and your company doesn't have the luxury of waiting for things to change. It goes without saying that you'll need to staff the lab with employees skilled with the new tools and architectures, so prepare to hire a new breed of worker. It also should go without saying that you need to integrate your lab with your mainstream IT organization; otherwise, it does run the risk of being a Potemkin Village.
- Willingness to live with a more rapid pace and with more risk. One could say that traditional IT has been charged, more than anything other of its responsibilities, with risk reduction. That has inevitably, and understandably, led to caution and sluggishness. That was then, this is now. The traditional responses of IT aren't capable of meeting the challenge of the Third Platform, and it's not enough to want things to change. You have to change and be willing to live with less certainty and more risk. If your message says "change a lot" but your metrics and rewards reinforce traditional requirements, you'll get lip service but no change, because it will be clear what you really value.
I've seen many presentations containing a slide with Mark Andreessen's quote "software is eating the world" in the past month (and, in the spirit of disclosure, my presentations also have a slide with the quote). The remarkable disruption our economy is undergoing is completely changing the role of IT -- turning it from a late adopter of mature products to an innovative creator of leading-edge solutions. Instead of impassively standing on the righthand side of Moore's chasm, enterprise IT is increasingly reaching across it to incorporate emerging technologies. We are in the very early stages of this process, and it's becoming clearer and clearer that the world of enterprise IT is going to look quite different in the very near future.