Saturday, June 14, 2014

How to beat Amazon at its own game ...

I was asked a question today of 'What can IBM do against AWS?'

About ten years ago I tried to convince IBM to build a utility service for infrastructure (based upon Fotango's Borg system) and about six years ago I spent several hours of my time explaining to an IBM C-level on the economics of change and how IBM needed to build an AWS clone and create a price war in order to increase demand beyond AWS ability to supply (building data centres is a constraint) and naturally fragment the market in IBM's favour.  None of this happened and so my first response to the question was unfavourably ... "die from inertia?" 

However, given it's a friend who asked the question, I thought I'd spend 30 minutes mapping out the environment, looking for weaknesses and write a quick plan.  So, I've written a map but I don't have time to convert this to an image.

Before I tell you the play, you need to make sure you're familiar with the ILC ecosystem model and why information flows are critical to it. Understanding mapping, evolution, situational awareness, the importance of not just harvesting but nurturing an ecosystem is also essential. For some, this post will seem like gobbledygook.

It's enough to say that key to the play is ...

1) Understand your opponent. The value of AWS is not the utility service but the ecosystem of companies building on top of it. This ecosystem acts as a future sensing engine with everyone innovating and AMZN getting early warning through consumption information. If you're going to take them on then you need to neutralise this or at least get yourself into the flow of information.

2) Understand the landscape. The shift from product to utility is inevitable because of competition. Many 'legacy' companies are struggling to find a route through this. Many new companies are building directly on top of this. Beyond simply getting yourself into the flow of information, you also can exploit others to create a counter force to AWS.

So, what should you do? 

1) Provide some form of proxy service to AWS and call it a brokerage. Enable companies to build on your service (in exactly the same way as they do with AWS) but IBM takes care of the pricing / billing of AWS i.e. IBM plays the game of using reserved instances, spot markets etc on behalf of clients.  Use this to provide all sorts of tailored financial terms suitable to enterprise clients. Yes, you're going to embrace and help grow AWS but overtime you want to position this such that all AWS sees is the IBM 'account' but IBM sees which of its clients are rapidly growing. For a fast track route into this space, go acquire Cloud Options as a starting point. 

2) Once you have a large ecosystem effectively under your observation that is building on AWS through your proxy, start mining that ecosystem to identify new services that are rapidly growing. Acquire them - don't copy. Promote an image that using the IBM proxy service is also a fast route to acquisition along with being vastly cheaper (due to the financial engineering done).

3) At the same time as step 1) you should build a platform play on Cloud Foundry. IBM is already doing this - BlueMix - and that's good but it should build the platform not just on Softlayer but AWS, Azure and GCE. Encourage others to do the same because key to the play is a market of public providers all based around Cloud Foundry which are built on a mix of IaaS including AWS. You want Cloud Foundry to be big on AWS to prevent AWS making moves against it. Don't try and differentiate on features, focus on operational excellence. Help others set up as providers in competition

4) Capture the platform assurance market. With many Cloud Foundry providers then companies are going to want to know who is interoperable and who is reliable. First, you want to make sure that a competitive market forms (unlike the mess in OpenStack) so support Pivotal as a Benevolent Dictator. However, build an assurance platform which monitors all the different Cloud Foundry providers and determines interoperability and compatibility to the core. Once you have market of providers which you have assurance information on then offer a brokerage service for platform. Act as the proxy for the market. This again is a capture play for information from the ecosystem.

5) Now you have a proxy role for both PaaS (CF) and IaaS (AWS) and are capturing information from both ecosystems for acquisition plays.  Being a big supporter of AWS (both through proxy on AWS and use of CF on AWS) then Amazon will find it difficult to react against you. At this point, you can start to look to play substitution games at the PaaS layer i.e. providing price benefits for those using Softlayer.

6) Oh, there's a lot more but my 30 minutes are up and I can't type that quickly.

Ok, I'm not saying the above is the only way forward (and certainly most of the above is late in the day) but what I am saying is that if in 30 minutes I can start to identify routes of attack against AWS then with a bit of effort and gameplay there's a lot that IBM can do to change its fortunes in the market. That IBM hasn't been doing this has been a sorry tale though the first signals they are getting better are appearing. Oh, and the above also goes for HP, Cisco, Oracle and all the rest of the old gang.

My advice to my friend on 'what can IBM do against AWS' is simple ... chose one from "learn to play the game fast" or "build a time machine" or "die from inertia"