1 Comment

Summary:

What does it take to move companies toward a data-driven future? EMC chief strategist and Pivotal Initiative leader Paul Maritz spoke at Strucuture:Data in New York on how to move toward the future through human leadership and strategy.

Paul Maritz EMC Pivotal Structure Data 2013
photo: Albert Chau


Session Name: Pivotal Moments in Technology.

Speakers: S1: Om Malik S2: Paul Maritz

OM MALIK 00:00

Welcome back, Paul. It’s been a while, 2 years. A lot has changed, you are the MC now.

PAUL MARITZ 00:51

Yes. As you know, I am in transition, and until the middle of last year I was at VMware, and then on April 1st, my title will formally change again when we introduce the third member of the family.

OM MALIK 01:09

And that member of the family is Pivotal?

PAUL MARITZ 01:09

Correct. We hope to be a new platform for building data-centric applications.

OM MALIK 01:30

And are you focusing on a certain market, or is it all purpose, consumer and corporate?

PAUL MARITZ 01:31

Well, the interesting thing and the motivation behind this move is that across quite a broad spectrum of industries there is a pattern of applications, or the need for a pattern of applications emerging that can’t easily be done with existing technologies. These applications all combine to some degree the need to analyze and reason over much bigger data sets very cost-effectively, and to combine data that is increasingly going to be coming from the world of telemetry. This means data that arrives in large quantities, in real time, and that needs to be reacted to in real time.

The key point is reactions. It’s not enough to be able to do analytics, but also drive some kind of experience or value that combines what you’ve historically seen and what you’re seeing in the instant. Lastly, we believe that these applications should be built in a cloud-independent manner. Infrastructure level clouds like Amazon and OpenStack are the new hardware, and applications shouldn’t of necessity be restricted to one of those clouds. We’re bringing together assets from the big data space, the fast data space, and cloud abstraction/virtualization.

OM MALIK 03:16

Let’s talk about that a little bit. What you’re suggesting is that the first 50 years of technology, which was about paper automation, making business processes more efficient, is over, and going forward in this new data-driven era we need new kinds of platforms.

PAUL MARITZ 03:41

As you and I have discussed, most of the existing IT world is in the mature stages of a very successful 40 to 50 year journey to automate everything that we do with paper, whether that be bookkeeping, keeping track of customers, or communicating. You could argue that there are certain industries like healthcare that haven’t even really begun this process, but in general businesses today need those capabilities, they can’t function without them, but they are no longer differentiating. They aren’t going to offer new competitive capability.

If you talk to many companies today they will tell you that the competition is changing and they’re not sure how to address it. If you’re a retail company, your new competitor isn’t the big bricks-and-mortar retail store down the road, it’s Amazon. If you’re in the payment space, your competition isn’t necessary one of the existing payment players, it’s PayPal. Those existing companies are looking out seeing that their competitors are using IT in a very different way, and this is a big problem for them. So we think it’s interesting to look at who has driven new business value in the last 10 years, and it is the big consumer internet giants, such as Google, Facebook and Amazon. They’ve learned how to do three things. Firstly they are reasoning over much bigger datasets much more cost-effectively than could be done with traditional data-warehousing techniques. Secondly, they have found ways of automating much of their operations. They have aligned their underlying IT capabilities in an automation-centric fashion rather than a people-centric fashion. Lastly, they have a culture of using those capabilities to rapidly develop and introduce new products. What we’ve been thinking about is how to take those three capabilities, add in the need to process fast, high-volume data from the world of telemetry, and how to combine those in a way that existing enterprises could use.

OM MALIK 06:56

The question is, will it really be possible for the established data processing vendors to adapt to this new world? Can you teach the old dog new tricks?

PAUL MARITZ 07:13

That’s a question I ask myself a lot as an old dog! If you go by history, the answer is probably no. If you take the thesis that there have been three IT eras, the IBM mainframe era, the client-server platform, and now this emerging data-centric cloud-based platform, with a few exceptions, people who did very well in one era haven’t done so well in successive eras. Another way of looking at this is that each of those eras had their canonical set of applications. In the mainframe era it was about automating bookkeeping. In the client-server and web era we tried to things that required a more sophisticated data model, and that drove the relational database. Now in this cloud era, the data model is going to have to change again. You can’t drive these scenarios on top of a traditional relational database centralized kind of model. Every time the data model changes, you find that everything above and below it changes too, because it really is the glue that ties applications to the underlying infrastructure.

OM MALIK 08:49

This injection of web-scale ideas and practices into various companies involves a whole different kind of thinking, not just at the IT infrastructure level. It affects thinking at the executive level, at the middle-management level, and all throughout the company. People insist on seeing data as this special thing. I feel that it’s more about the culture of the company. If the company doesn’t have a culture of innovative use of data, there doesn’t seem to be much they can do to adapt.

PAUL MARITZ 09:23

I think it’s tied to business models. These underlying structures allow you to rethink business models, and unfortunately I’ve learnt that it’s very hard for an organization to change its business model. All of us become very comfortable with a business model, it allows you to look into the future, it’s your compass” that allows you to make predictions, and it’s very difficult to let go of that. It’s not so much a matter of technology, it’s the fact of a company coming at you with a very different business model. Amazon for example doesn’t have the same focus on managing brands in the way that a traditional retailer does. They manage a customer relationship. Their whole business model is predicated on getting value out of the customer rather than squeezing value out of brands or products.

If you look at the industrial control world, modern manufacturers of jet engines are increasingly selling power by the hour”, instead of just selling engines. This means they subcontract to provide an airline with 10, 000 hours of power for example. What makes you successful in doing that is how successfully you can service the engine during those 10, 000 hours. Today, servicing is done incredibly inefficiently. In the worst case, every 2, 000 hours you unbolt the engine, break it down into its constituent parts. You have to assume that everything is broken and prove that it isn’t so you can put it back together again. If you can get the telemetry off that you can start making very different decisions, about how to service those engines. I’ve been told by people in the electricity generations space that today they don’t trust their customers, so they essentially de-rate everything! Whereas if they could get the telemetry data off those generators, they could drive them much closer to their theoretical maximum. This would allow them to make decisions about how close to get to the maximum, which means they could be selling excess power, but they would be trading that off against shortening the life of the piece of equipment. This is a very different business model, and that’s what I think is disruptive.

OM MALIK 11:51

So what will it take for companies to become data-informed and data-intelligent?

PAUL MARITZ 11:57

I think there’s two ways that could happen. I think unfortunately the first is Darwinian evolution! I think what history teaches us will happen is certain companies will fade and others will rise because they see the opportunities of using this type of business model. Others will change and change can only come through leadership. Somebody in the organisation has to really step up and drive change. I spent ten years running VMware and it was interesting to see why customers had such widely varying rates of virtualization. It got the point, and this is slightly tongue in cheek, that I could meet the CIO and tell you with a high degree of certainty what percentage virtualized that company was. Change requires leadership. It requires people to understand what is happening and really get behind it and drive organizations to transform, because none of us really like to change.

OM MALIK 13:18

I’m going to press you a little bit on your comment about Darwinian aspect of data. Do you feel that we are entering an era where a lot of the classic businesses are going to fall by the wayside because we have a data tsunami” coming which will really change how the business is supposed to work?

PAUL MARITZ 13:45

I think that’s a possibility. If you just look at the disruptions that have been caused in industries by people who start from a data-centric perspective and Amazon is really exhibit one, in disrupting a whole industry based on data. The interesting thing is that as the costs of telemetry come down, other industries that we would not have thought were susceptible to that are going to be just as susceptible. Anything from control, electricity generation, jet engines, all the way through to the communications world.

I think you can see this emerging pattern of applications that apply across significant numbers of sections of industry and having been in this business for 33 years now, when you see a pattern of applications emerging across a wide variety of industries and when there is no easy way of doing those applications, that is generally a strong signal that a new platform or platforms will emerge.

OM MALIK 15:03

There’s another interesting aspect to what you’re saying. I’m an American Express customer, and the only thing I look forward to from them, because everything else is a bill, is the year-end report they send you as to where you spend your money. It’s interesting to look back on, but it doesn’t really solve any specific problem; I already spent the money, I already went over budget. This is a company that has massive IT spending, and even they can’t come up with any more data intelligence for customers than that.

PAUL MARITZ 15:46

This goes back to the culture issue, and this is not just an American Express issue, this is an industry wide issue. A lot of enterprises have just come off 15 years of best practices”, which is to outsource everything, including their IQ. All of a sudden they find themselves having to deal with competitors who are using intellectual property and IT as an offensive weapon. They’re either going to suffer a Darwinian fate, or they’re very rapidly going to have to rebuild a culture of being able to build intellectually based products, which means thinking very deeply about how to use data, and not just having insights into data but how you translate that into the experience, the real time experience that the customer has.

One of my favorite examples of this is, if you look at telecommunications today the average company is divided into two almost separate universes. There’s the universe of the network that runs the actual communications and then there’s the customer care side, that generates the bills, and these two worlds hardly talk to each other. At best there’s a batch file that goes over once a night. So the experience that the customers are having in real time, on their mobile device, that’s impossible for the Telco to understand today. They can tell you they dropped a call, but they have no idea whose call they dropped. Today in the average Telco, if you wanted to ask when was the last time you dropped one of my calls? it typically takes them days, even weeks to find that out. To really transform their businesses not only from a customer satisfaction point of view but from a business point of view they’re going to have to bring those worlds together. Every time they drop your call the least you’d expect is a text message apologizing. They can’t even do that today let alone do more ambitious things on top of that.

OM MALIK 17:55

On the flip side of that, it’s actually a good thing for smaller companies. If you’re a small startup and you want to take on Macey’s, it’s much easier to do that if you have best data practices in place.

PAUL MARITZ 18:09

I agree, although that’s where I think you need platforms and services to do that, because small companies can’t build this infrastructure. Amazon web services has demonstrated the value of that, but in our opinion this has to go to another level of abstraction.

OM MALIK 18:35

Hopefully Pivotal will do that, and thank you so much for taking the time and talking to us, and hopefully next time you will give us the state of Pivotal.

PAUL MARITZ 18:40

Exactly! Now it’s just the small matter of execution. Thank you.

[applause]

You’re subscribed! If you like, you can update your settings

firstpage of 2
  1. Paul is the right leader to not only move Pivotal to new platform. He has the unique capability to move the industry to this new platform. He is also very correct that this transformation can only happen when the new data driven platform is not tied to one cloud.

    There would be two challenge that Pivotal / VMware would have moving forward. One would be engineering & technology and I think they would be able to execute on this front.
    The bigger challenge Pivotal / VMware would be in running large cloud. Pivotal / VMware lack this DNA and it would be some time before they would learn this. It is more than just Dev/Ops automation. A service company requires a very different culture, how you think and interact with customer is totally different. Simple things like selling and customer’s service is very different. Service company product definition and execution is 180% opposite to product companies. Companies like Microsoft which have been running service for more than decade do not execute well on services because of their legacy product background.

Comments have been disabled for this post