Those of us who cover the world of IT never tire of referring to what has become known as the “Yegge Rant,” a revealing inside look at how Jeff Bezos, CEO of Amazon, played the role of a forcing function.
The Yegge Rant includes a leaked communication from Bezos, who demanded that everyone at Amazon implement APIs (Application Programming Interfaces, for the unwashed). APIs started as an obscure way of allowing programmers to access capabilities from other applications. But as the Internet became pervasive, APIs have become a tool for transforming a business and creating powerful new patterns of collaboration and innovation.
Bezos saw how important this was quite early and threatened in 2002 or so to fire anyone who didn’t make the assets of their department available through APIs. Likewise, Apple, Netflix, Google, Twitter, Facebook, and Amazon have all expanded their reach and tapped the power of millions of developers by publishing services accessible through APIs.
But if we look at the way this was done, we see the usual divide between the engineering rich, early adopters of Silicon Valley and the rest of the business world. Chet Kapoor, CEO of Apigee, saw this divide as well, and, since 2004, Apigee has been productizing the infrastructure that will allow the rest of the business world to adopt APIs in a secure, scalable way and reap the same benefits as the mighty, mighty early adopters.
Given the benefits that companies of all sizes in a diverse collection of industries are getting from APIs, it is clear now that more CEOs and Boards of Directors should be delivering their own version of the Yegge rant. While some companies are making clear progress, for the most part, based on the conversations I had at Apigee’s “I Love APIs” conference this year in San Jose, it seems like the leadership toward API adoption is coming from the middle and the bottom of the company. The question is why? Is the logic that was so clear to Bezos so difficult for senior management to understand? Or is the domain of APIs too technical to take hold?
Something doesn’t make sense.
At the “I Love APIs” conference this year, I caught up with Kapoor to assess the state of the market for API technology and discuss the forces driving more and more companies to adopt APIs as a way of lubricating and opening up digital channels of their business.
Apigee and the whole API space are ready for their close-up. The space has clearly been validated as crucial to the enterprise by recent product offerings from Amazon, Microsoft, and the acquisition of Mashery by Tibco, in addition to offerings from several other smaller players. The real question, however, is how many businesses are ready to take up the challenge?
IT is Still Stuck in the Cost Framework
The problem is that for APIs to be an urgent matter, you must have a vision for IT as a powerful agent of change and innovation. When IT is a commodity (remember Nicholas Carr’s article, “IT Doesn’t Matter”), expense reduction is the goal and the biggest victory is to reduce the cost of IT to zero.
Of course, these days, few CEOs and boards would admit they were that stupid. But, when they have the conversation with the CIO, what is the focus? Cost reduction. How much can we cut the IT budget this year?
It would be a mistake to apportion all the blame to the CEO and board. CIOs have shown a lack of imagination too, as evidenced by the rise in the Chief Digital Officer role. This role, as I have pointed out in several articles (“Why The Rise Of The CDO Role Represents A Power Grab”, “Don’t Let A Chief Digital Officer Steal The Best Part Of Your Job“), is filling the vacuum in promoting creative, value-producing ideas about applying new technology.
For APIs to be important, a company must see that a vital transformation is taking place. The current word for that is “digital,” which usually refers to the way that new companies are transforming the world by using the vast power of cloud computing, software infrastructure, the Internet, mobile devices, and a flood of all kinds of data to create new businesses and disrupt existing ones. (In my view, the CDO title can be evidence of weakness unless a company is using that role to lead the way to creating platforms as described below. See “Why Your Chief Digital Officer Efforts Will Fail Unless You Construct Platforms.”)
I don’t care what sort of org chart you have and what the roles are called. If you have creative visionaries in your business working together, it doesn’t matter if you have a CIO, CTO, CDO, or whatever. What matters is that you understand that both the consumer and business worlds are being transformed at a rapid pace, and you must have a strategy for making this change work for you, rather than waiting for other companies to use the power of digital transformation against you.
Kurzweilian Platforms for Exponential Growth
As we all know, adopting technology without a master plan doesn’t really work. The Bezos rant was not inspired by a love of APIs. Bezos realized that without a computing platform that exposed crucial data and functionality as APIs, progress toward his master plan would be unacceptably slow.
For Bezos, this was obvious, but he realized that it may not be obvious to the rest of the company. That’s why he threatened to fire people if they didn’t make it happen. But there is an additional part of the Yegge rant that is often overlooked. Bezos not only demanded that everyone create APIs, but also that those APIs be ready for adoption by external users.
Why was this so important to Bezos? I think it is because Bezos understood that Amazon was already a platform but needed to be an even broader platform to become a more powerful company.
I think there are two ways to understand the power of platforms. First, it is important to understand the technology logic that is driving the acceleration of innovation. If you are a student of Ray Kurzweil’s thinking about technology—and if you aren’t, you should be—you will know about his theory of the exponential progress of technology (see “The Kurzweilian Logic Of Exponential Growth In The Interconnected Era”). Kurzweil points out that in most realms of technology, progress is initially slow as basic components are built. Then those basic components are used to create more advanced components, then the process repeats itself until one act, such as executing a Google search, can unleash a immense chain of computing that results in an incredibly powerful response.
So, in an enterprise computing context, the same process should be going on. Basic components should be created and exposed via APIs, then more advanced components can be created on top of those, and so on. In a modern enterprise, you must be able to use both internal and external APIs in a reliable, secure, and scalable manner. That’s where API infrastructure provided by Apigee and other companies in the space becomes crucial.
But being able to create and use APIs and expose them internally and externally is a technology activity. What is the business logic that makes a collection of API-powered services and products disruptive and successful? The answer is the power of platforms.
Sam Ramji, currently CEO of Cloud Foundry, and for several years a senior strategist at Apigee, has been preaching the gospel of platforms for several years. In a killer presentation that he gave at the Apigee conference, Ramji assembled a trenchant narrative that explains how platform economics drives the disruption by the usual suspects of digital innovation—Amazon, Google, Facebook, and Netflix. Ramji, by enlisting thinking from MIT’s Marshall Van Alstyne and many others, explains how these companies play the role of orchestrator of a variety of services. They enable many different groups to use those services to create products that are then sold to other parts of the ecosystem. A network effect ensues that makes the pie bigger and bigger. The platform creator then collects a larger and larger portion of the pie. This idea has been with us for a long time. Think about Microsoft CTO Nathan Myrvhold’s idea of collecting a “vig” on network-enabled commerce. In Myrvohld’s vision, Microsoft was too much the king. The way modern platforms are created is facilitated by the creative energy of millions of people, each providing something and in turn reaping benefits.
How do millions of people work together to create massive amounts of products and services to sell to each other? Using APIs.
Another idea that is not often referenced is also relevant here. The platform-creating companies play a key product management role, one anticipated by John Seely Brown and John Hagel in their early work on web services. Seely Brown and Hagel anticipated the rise of service aggregators who would collect and organize large amounts of services and guarantee their security, reliability, scalability, and functional completeness to solve a problem. Seely Brown and Hagel pointed to the Chinese company Li & Fung as an example of a company that orchestrated huge amounts of business services to create a vast array of products. The modern platform innovators take this orchestration to the next level by allowing the innovation to happen without prior restraint in a way powered by the individual energy of creators and consumers. These platform innovators are service aggregators and ecosystem designers.
Chet Kapoor: Philosopher CEO for API Adoption
With these ideas in mind, we can now understand the challenge facing Kapoor as CEO of Apigee. For Apigee to succeed, the kind of Bezos thinking about the importance of APIs to empower platforms must become widespread.
In a fascinating conversation with Kapoor at the “I Love APIs” conference last week, he surveyed the issues that most business are facing related to accelerating their digital transformation. This edited Q&A summarizes Kapoor’s point of view.
Q: How would you categorize companies related to their attitude toward API adoption?
The CEOs and the boards we talk to fall in two camps. In the smaller camp are those who know that they are being disrupted or are going to be disrupted, and they want to get ahead of it.
In the larger camp are those who know that something is going to happen, but who are somewhat paralyzed and don’t know what to do about it.
The first group looks at an Uber and says, “If that can happen to taxis worldwide, it can happen to us or anyone else.” Sometimes it is clear how the disruption will happen in their industry, but more often they glean lessons from another industry.
For example, block chain (the foundational technology for Bitcoin) is an example of something that financial services executives are looking at and saying, “if you squint hard enough, you can see a similar technology changing the way we do business.”
Businesses that are not digital businesses come to the realization that what they have is valuable in the digital realm. They just have to make it relevant in the new world. When you come to the conclusion that you have some really valuable assets and need to deliver digital experiences to customers and partners, you realize that the only way to unlock those assets is to make them available through APIs, supported by the right management and security.
Q: What steps should companies be taking?
Companies need to engage with customers and developers through a platform relationship.
Companies must ask their customers how they should innovate, because the customers’ expectations have changed. Can you imagine if United had talked to all the millennials and said, “What do you expect from travel?” Every one of them would have said Wi-Fi access. And if United had provided that, Virgin America would not have taken off. During the first two years, everybody went to Virgin not because of the cool lighting, but because Virgin offered consistent Wi-Fi access. You must check with the customer about their buying habits and expectations first of all.
The second key step is to understand and apply platform thinking. CEOs and boards need to understand the network effects of digital. The crucial step is to understand how a platform is going to apply in a particular industry. The problem is that the platform economics don’t work when you’re selling a tractor, but they do work when you’re selling equipment on subscription. When you start thinking about your business as a digital business, the unit of value is not a physical good but a digital good or a digital experience, which makes network effects much easier to come by.
Q: What are the barriers to adopting platform thinking?
The most enduring barrier is the mindset of a hierarchical culture. When the culture is a cathedral, it is natural to think, “I know what’s best for customers; I will tell them what’s good for them.” Or, “I’m going to ask my direct reports”—all of whom sit in corner offices. These are not people who are thinking through how behaviors are changing and how they’re affecting everybody.
When you are not talking to the right people internally or the right people externally, how can you really know what to do? You may know something is happening, but until you change the sources of information, you won’t know your way out.
Q: So in practice, where has the leadership on API adoption come from?
What is amazing to me is customers whose businesses are more than 100 years old are leading the way in scaling adoption. Over the past two years, the API traffic in the Apigee cloud for this group grew 5 times faster than other customers’ API traffic. I think that for these companies, making large changes is already a habit. They are more comfortable moving on from businesses that are successful before the bottom drops out because they have examples of the value of making such moves. For these firms, the leadership is driving change.
But most of the time, the transformation that we hear about from customers is being led from the bottom and the middle. If you look at customers like Walgreens, we started working with them on a mobile project. They wanted to take photo printing and make it available through a mobile device. And the API strategy behind that particular project has spread like wildfire. It wasn’t a top down initiative, but somebody at Walgreens was listening and let the people with the ideas run with them. That’s the important part.
That’s why I ended my keynote by saying, “Get started.” It doesn’t matter where you start, but just get started with a mission critical project of some sort. In the process, you’ll discover a lot about yourself and your customers’ expectations.
Q: How can you jumpstart innovation?
One of the ways the Valley produces great products and services—from Uber to Facebook to Apple—is by enabling users to create the products they want. If you’re a carpenter, you also sit. If you’re building a chair, you can tell a lot about the chair because you’re the creator and a user. That type of thing needs to happen more often.
You need to nurture and reward a network of makers either externally or throughout the enterprise, because you don’t know where the next big innovation is going to come from. This is one area where we can learn from open source. Put the right rules and constraints in place to ensure security, but make sure the APIs are available to people and that they can innovate on top of those digital assets.
The person who came up with the idea of integrating Spotify with Uber was probably not somebody sitting in the back office. It was probably a developer who said, “You know what? I want to listen to my playlist when I’m in an Uber car.”
Q: Are you worried about the new arrivals in the API management space?
The arrival of Microsoft and Amazon into this space and the investment from Tibco to buy Intel’s API management assets is validation that APIs are here to stay. Every time somebody says they’re investing in an API management product, we celebrate. It’s awesome that people think that this is the future. And quite frankly, it is, and we’ve believed it for a long time.
We fundamentally believe that the old will not get you the new. So this is not about an integration company having an add-on, the same way the cloud isn’t about a hosting company saying that they’re in the cloud business.
It is inevitable that platform as a service vendors will have some API gateway functionality that they’re obligated to provide to their developers because they’re going to make APIs available and they have to throttle traffic. Making that functionality available is great, but securing, analyzing and scaling those APIs and the resulting data through an intelligent platform is so much more. And so we see that as complementary.
Q: What is your advice about how to accelerate digital transformation at a typical company?
If you are a CEO who sees the need to become more digital, you must figure out how to break down the barriers and change the organization. While doing this, you will move out of the hierarchical cathedral, which at first will be scary. As soon as you open up an API, whether it’s external or internal, all of a sudden, you can monitor who’s using it, but you can’t really stop anybody from using it. The command and control mentality changes. If my board asks, “What risks are you going to run this way?” and I say, “I don’t know, we’ll have to manage them,” that would make anyone nervous.
The remedy is to move quickly, but cautiously. You won’t die tomorrow if you don’t do this, but you will lose ground – rather quickly (Blockbuster), or maybe over a period of time (Wal-mart). As long as your industry is not currently being disrupted, simply start moving the organization,, and start implementing APIs as an experiment with the right governance. If your industry is being disrupted, and you can see and smell the disruption, then you have to overcome risk aversion and thoughtfully start opening up. And if you’re in a sensitive industry like financial services, you may not start opening up externally, but at least start opening up internally.
By the way, financial services is a great example. The way we live is not the way we bank. And there are some companies that are changing the way they want us to bank with them, because the expectation is an Uber-like experience. It should work the way I think or the way I live.
Q: How do you handle the explosion of data?
Most CEOs know they have a data management problem but don’t know what to do about it. It is common knowledge that most companies don’t even know how much data they have. This is not going to change overnight.
My advice is to start thinking about the experience you want to deliver to the customer, and then see what data and functionality you need and make that available through APIs. Then get the right network of makers working on it to deliver the experience.
Once that cycle is working, think about all the digital assets that you have and determine which ones could be the foundation for new products. Consider which digital assets, if put into the hands of an active network of makers, have the most chance of delivering a unique experience that can provide a competitive advantage.
You should not start from the asset and say, “let me expose it.” You should think about the asset and assess the leverage you would get in delivering the experience, and then fill in the gap between the two.
The wrong approach is a boil-the-ocean data management project. The right approach—and there were four or five different sessions at our conference on this exact topic—is how to go from the UI to the API. You go from the experience to the API to the data.
Follow Dan Woods on Twitter:
Dan Woods is on a mission to help people find the technology they need to succeed. Users of technology should visit CITO Research, a publication where early adopters find technology that matters. Vendors should visit Evolved Media for advice about how to find the right buyers. See list of Dan’s clients on this page.
This article was written by Dan Woods from Forbes and was legally licensed through the NewsCred publisher network.