CIO UK: How did the book come about?
Vivek Ranadivé; In some ways the book is a combination of my life's work – where if you get the right information to the right place, at the right time and put it in the right context then you can make the world a better place.
The basic premise of the two-second advantage is it's more valuable to have just a little bit of the right information just a little bit beforehand – a couple of seconds, a couple of minutes, a couple of hours – than it is to have all the information six months later. After all, what's the point of knowing that you've lost a customer after the customer has gone?
Where did you discover this concept?
It turns out that super-successful people have internally developed that two-second advantage. So if you look at the great ice hockey player Wayne Gretzky, he said that his success was because he went to not wear the puck was, but where the puck was going to be and if you look at Wayne Rooney he does that on the soccer field as well. They would combine patterns with events and close the loop to be able to anticipate at the moment of truth.
Is this something that can be applied to business?
For example, a telecom company adds 5 million customers every month but they lose a couple of million every month. By using 20th century database technology and CRM technologies and doing some data mining, they try to find out what went wrong six months after and they don't make any headway.
So, we've been pioneering the concept of a real-time event driven world. A world in which, when something happens, it is put in the right context over several patterns and we can see that if the telco's customer gets six calls dropped in a 24-hour period then they will switch supplier. Every dropped call is an event and after the fifth dropped call the telco offers their customer free SMS messages if they top up their prepay card in that 24 hour period and the problem of churn is gone.
Isn't this something that conventional database-driven platforms can do?
Let's say, if you get on a plane in New York and six hours later you land in San Francisco and you go over to the baggage carousel and there's no bag. Then you wait in line and you have to tell them that you didn't get your bag. Now, why are you having to tell them? The reason is because the old database technology is like a phone that doesn't ring. Say something happens and it goes into the database and nobody else knows that something happened. It gets lost in the application noise.
In our world everything is an event – you get on the plane, it's an event. Your bag doesn't get on – a non-event is still an event and we connect the dots and now when you land there will be an SMS or there will be an agent waiting for you to say sorry, your bag didn't make it, give us your home address we will send it to you and we had deposited 5000 Air Miles to your account. So really, the two-second advantage is about giving you that advantage to do the exact right thing at the exact right time.
How does this work in detail?
The world is moving from a transaction architecture to an event-driven architecture. You go on a website and you click on a mortgage calculator. That should be screaming that you are shopping for a mortgage, but nothing happens. So in our event-driven world, we check you out, we check your payments and check out your credit and before you leave the website we can make you an offer on the spot – that is the two-second advantage being able to tie patterns with events.
If you can put this advantage in a box and make it a corporate skill it will allow companies to sell to their customers better. Let's take an example like cyber security. What's happening there is people are trying to build bigger and better locks and every time you do that somebody builds a way to pick that lock.
In the world of the two-second advantage we look for suspicious events in the neighbourhood and we look for suspicious activity so it's like a neighbourhood watch. We pick up suspicious events and we get the bad guy before he ever gets to your door
What you're talking about is instinct. How can you bottle that?
If you get on a plane and your bag doesn't, there will be a problem. That's not voodoo. If there is a delay due to the weather then there will be a forward chain of events that you can adapt your value chain to, so that planes don't show up, the crew doesn't come, the food doesn't show up because there's been a delay.
What I'm talking about is factual, it's not voodoo. Anyway, I believe that math ends up trumping science. You don't really have to know how something works, you just have to find the pattern. You know that if you drop six calls, the customer will switch. The mathematics shows that there is a correlation.
How would this event-driven platform look?
By now every corporation and the government has a transactional platform, usually Oracle-based and that's great for storing transactions, but in order to be successful in the 21st century you have to have a platform that is made up of another system, a bus to pick up and tie events together. It's made up of the brain that is able to find patterns, it's made up of a memory to be able to look at sets of data and it's made up of muscles with a business process management engine to be able to take action.
How much interest do you think there is in this new kind of platform?
The rate, in which interest in this type of computing is growing, is accelerating. Normally we have five or six people interested in a beta test. Last time we had 500 people wanting to do the beta. We are seeing industries that we never saw before. I had a grocery chain come to my offices. We were wondering why they were talking to us and is not IBM? Why not just stick with the transactional systems?
They said: No we need to become event driven. We need to tie our product-set all the way to through the value chain to the customer and we want to make an offer to the customer before they leave the aisle, not six months after they have left the store
So this demand for event-driven computing is going to grow quickly?
Take a bank with 10 million customers in the 1950s. You might have had just a handful of touch points with each customer – through a letter or at the branch. Now you have millions of touch points and information is continuously online and available. So now take the same 10 million customers in the 21st century and that's multiplied to 100 million touch points as these customers can be on your premises, on a website, on the phone, making a credit card transaction, Tweeting about you.
Then, sitting in the back of the bank is hundreds of terabytes of data and what we have to be able to do is take each of those events look at it uniquely, tie it back to that storage and create that two-second advantage. To have a transactional platform was the static data. Event driven platform is the dynamic data. The killer app then is not an app at all – it's the business process itself.
Investment in new IT isn't popular at the moment, how are CIOs going to sell this to the board?
The beauty of this is that it slides into the existing architecture and so you are not ripping anything out. You are actually making the phone that doesn't ring, ring. So, you are patching an adaptor to the database and unlocking that information.
And, the way that you are selling it is you are not calling it a new architecture, you are saying let's do a customer upsell and cross-sell application where the customer gets the exact right offer or you might sell it as we need fraud detection or we need cyber security all, we need to optimise the supply chain. You are selling it as agility, you are selling it as risk management. These are all 21st-century problems and opportunities, so you are not going in there and making it a religious sale about putting in a new architecture.
So event-driven computing is here to stay?
If you look at the companies of the future, if you look at Amazon at Google or Facebook, the little secret that people don't want you to know is that they don't actually use databases – they use the database to store transactions but much of their business is run in memory.
If I told you 10 years ago that most of what you do wasn't going to be on the landline but on the cell phone, you might have been sceptical. Similarly, what I will tell you today, is that 10 years from now, much of what business does will happen outside of the database, it will happen at the moment of truth. It will happen at the event stream.