From Vertical to Horizontal Integration
The announcement of the Amazon Kindle Fire touchpad gave me quite some food for thought. It has to do with the concepts of application development and deployment that I will come back to at the end. The Internet teaches us once again valuable lessons for the enterprise.
Amazon versus Apple versus Google
I had expected Amazon to somehow ‘retaliate’ on Apple’s iCloud as it is a potential competitor to their Elastic Computing Cloud (EC2). The companies seem to be quite different but Kindle and iPad put them head to head. Apple has up to 30% gross margin (for the iPad2) on its hardware products while being fairly demanding with rules and commissions on the publishers that offer software and media for its devices. Amazon on the other hand is a retail company that makes 40% of its sales from digital and physical media. Thus, Amazon doesn’t care much about Kindle margins, because it is primarily a landing pad and consuming device for content providers selling through Amazon. Amazon knows how to make deals with those and is thus growing 50 percent per quarter in revenues, and could reach $50 billion in sales soon. And now comes the Amazon Kindle Fire with the EC2-backended Silk browser. While Apple focuses on using iCloud to make the synchronization of devices a no-brainer, Amazon takes control of the complete Internet experience and has the opportunity to mine the complete user behavior for a better user experience.
While Apple iCloud is – at least not yet – entering the custom application development market that Amazon is offering, the Kindle Fire combined with EC2 and Amazon Silk is a great response that nearly puts Amazon on par in terms of a tablet offering, because it too controls the complete vertial stack. At the same time Amazon is calling ‘check’ in its chess game with Google. because Google wanted Android to be open source. Amazon is not only pulling Android (the old open sourced 2.3 version is the code base for the now custom Fire OS) away from Google, but it is also disconnecting Google from its search engine users through the vertical integration of Kindle Fire -> Silk -> EC2.
But there are some threats too. While the Kindle Fire is not officially an Android device, lacking certification from Google, it is derived from Android 2.3. Google faces an uphill patent battle about Android that could also hit Amazon. Goldman Sachs estimates that Microsoft is already making more money from Android royalties than from its own smartphone revenue: $444M. The Kindle license between Microsoft and Amazon does apparently not cover Fire. Oracle is also suing Google for billions in intellectual property violations related to Java. For Google it is a substantial loss that the most probably most successful Android tablet won’t be officially Android. And to top it all off, the Amazon Android Appstore launched earlier this year is a big success, which means that Amazon does not need Google or its Android Market. Yes, it is a pretty smart move by Jeff Bezos and it will truly interesting to see how it pans out.
Amazon Silk and EC2
Amazon Silk uses a split browser paradigm across the Kindle Fire and EC2 to boost the Web page processing, taking into account network conditions, page complexity and cached content location. EC2 already serves some of the most popular sites and has immense bandwidth to the outside. It even gives the Kindle Fire the opportunity to become a natively connected mobile corporate device that runs cloud applications on EC2! The price of $200 versus the $800 of an iPad will be very convincing to corporate bean counters.
Mahi de Silva, executive VP at Opera Software considers Amazon Silk a copycat move as several Opera products apply compression for a mobile device since some time. He is right as even some ISPs do similar things. But he is missing the point. Amazon is itself a major content provider with its Instant Video streaming movies and TV shows, Amazon Cloud Player streaming music service, and the Amazon’s Kindle bookstore. The successful Android Appstore makes Amazon a complete one-stop shop for all your mobile needs!
There are obviously some privacy concerns as Amazon has now access to all it’s users browsing history and may also be supoened to hand over such information in litigation. But the same is true for Google and Facebook and users seem to care little about it there. Additionally Amazon already knows the owner of the Kindle Fire through its device and therefore there is no need to promise anonymization if they won’t pass on the information. And they won’t as it is a competitive advantage! It already has a large database on every Kindle user from their Amazon buying behavior. Silk in EC2 will also act as SSL proxy and while it might not have access to the data being transmitted, it will still know which sites you access when.
From Vertical to Horizontal
And that is how the vertical integration from device to EC2 backend goes horizontal, meaning beyond the simple connectivity. Amazon can track and learn from user actions much more than any Cookie trails in a browser ever could. The huge advantage lies in the real-time data access. If you think data-mining Facebook Timeline is huge then this opportunity is a magnitude larger. Amazon can improve its knowledge about customers without data mining and can directly use social networking information without needing to worry about Facebook or Twitter blocking its query engines to competitors. Amazon will know which products you researched before you even come to the Amazon store allowing much stronger cross-channel marketing than just past purchases. The opportunities for using the browsing data are endless. I do however expect that some of these capabilities will be scrutinized much more closer once the Kindle Fire will start to be sold in the EU due to its stricter privacy laws.
The Enterprise Opportunity in Vertical Integration
So let me close by pointing to what I think can be learned from all this for the enterprise. I have been advocating (and offering since many years) a solution that provides vertical integration from the user interface via the application model to the database. Adaptive linking of backend services by business users allows for horizontal integration of processes across applications and with it a huge opportunity for emerging innovation for the business. It is clearly what empowers the most innovative companies on the net: Amazon, Apple and Google. I am however told – by experts and analysts – that enterprises rather want to buy the software pieces and integrate them directly i.e. through Java, especially with all the old software they have already ‘invested’ in. While that may be true, I consider that a waste of money and opportunity. Protecting old applications will make the business very vulnerable and much less resilient to change. Take a look at how EC2 and thus all Amazon applications work and it is not your typical COBOL or Java shop. They use horizontal process integration across all apps through the vertical stack. Also Apple can do the same: If you are an iTunes customer, a sales rep in the Apple store recognizes if the card you use to pay is the same. He will just email you an invoice from his iPhone! Done! Very cool and very customer focused!
But jumping on mobile bandwagon in the enterprise as a ‘me too activity’ is questionable. Apple is offering an iOS app development ecosystem for the consumer world that is pretty amazing, but it does not offer practical horizontal integration as is needed for enterprise apps. iOS Apps can’t really talk freely to each other and if at all only trough some pretty limited and hardcoded APIs that Apple strictly controls. There is some opportunity for process integration through SOA interfaces but Apple doesn’t want other development tools except Objective-C. Hardcoding many (client/server based) Apps for the enterprise would be a substantial step backward. I propose that it is the vertical to horizontal integration – such as that Amazon is targeting with its EC2 backend for consumers – that will bring the full innovation potential to the enterprise. You will need a powerful integration backend to sensibly drive mobile apps. So if you like it or not, all the software packages enterprises have right now will either have to go or end up as horizontally linked backends-only to a vertical integrator that will provide user-driven, continuous innovation and the ease-of-use for wide-spread adoption.