Computers and artificial intelligence have come along at an exponential rate over the past few decades, from being regarded as oversized adding machines to the point where they have played integral roles in some legitimately creative endeavours.
Data is a raw material for creating entirely new digital goods and services and should be treated as a form of capital by organisations looking to stay competitive, Oracle’s big data strategist claims.
“There is a larger story behind all the talk about big data and digital transformation, and that story is an economic one: It’s the rise of data capital,” Oracle’s Paul Sonderegger told CMO.
“Data is kind of capital, on par with financial capital, for creating new digital products and services. Unlike the metaphors we use about data - that it’s the new oil, gold or new electricity – what we’re saying with data capital is that it fulfils the literal textbook definition of capital. It is a produced good, as opposed to a natural resource, and it’s a necessary input into other goods or services.
“It’s the same as a retailer who wants to go into a new region but lacks the financial capital to do it and can’t go. By the same token, if that retailer wants to create a new pricing algorithm, basket analysis algorithm or recommendations engine and lacks the data to feed it, they can’t do it.”
Sonderegger was in town to keynote the Big Data and Customer Experience World Show in Sydney, and caught up with CMO to discuss the idea of big data as capital, who he thinks should be the chief data officer, and what it takes to win customers in a world where digital platforms are the competitive battlefield.
Hasn’t data always been a form of capital, by your definition?
There has always been this idea that data is a useful record of what happened, and that you can use that record to make the process that produced it more efficient. We have been doing that for over 100 years and it goes back to Frederick Taylor and shovelling coal at Bethlehem Steel [the Principles of Scientific Management]. That whole push towards standardisation and productivity in the beginning of the 20th century was tied up with capturing information. That data was then used to make that specific process more efficient.
What has changed is that data is no longer merely a record of what happened, it’s a raw material for creating entirely new digital goods and services.
Is this ‘data disruption’ phenomenon similar to the ‘digital disruption’ businesses are experiencing?
Disruption has come in for a lot of criticism; it’s become an HBO sitcom, and even had its own exposé in The New Yorker. But we can all agree there’s something going on there, when Alibaba is valued at US$10bn and owns no property, or when Uber is valued at $50bn and owns no taxis.
What’s harder to see is that the digital disruption we’re familiar with because of these new consumer-oriented, online services is really platform competition. These platform companies are taking advantage of the fact that by capturing and using more data, they can turn an existing industry’s value proposition into a new set of services, which in turn, connect to markets. These new companies are positioning themselves as the platform that connects those two sides.
The challenge is that this phenomenon is going to come to real-economic industries that haven’t seen it before. One example is farming. There are drones now that will take these spectrographic images across fields, analyse those, and by looking at how much green there is, figure out the chlorophyll content of the leaves. That information can be shared with a fertiliser around the back of the tractor, which will change the fertiliser mix, as well as how much is put on different parts of the field. That tractor in the middle is now in competition to be a platform for agricultural services. Yet most makers of heavy equipment in agriculture don’t think that way.
How about the automotive industry? Is Ford’s FordPass in-car mobility app, for example, the car manufacturer’s attempt at becoming a platform provider?
That’s the same. Whenever you think about platform competition, you have to think about what the two markets are being connected. One way to look at it with automotive is you have information providers as one part of the market, then your drivers on the other side.
But the best way to look at it is to have information providers on one side, and developers on the other. If you’re Ford, what you want to do is attract more automotive app developers to your platform than BMW can to its. But it’s not as simple as that.
Meanwhile, technology companies like Apple, who are very familiar with platform competition, already know that the car is going to come down to a platform battle. That’s why you see things like CarPlay from Apple.
What are the key principles for understanding data as capital?
The first is that data comes from activity. Data is not a natural resource; you have to spend some money and put in some kind of computer to capture it, such as a sensor or application on a mobile device. The important implication for competitive strategy is that these activities are like new lands to colonise. If your rival gets there before you, they get the data riches from that activity. That’s why there is a real race among manufacturers and financial services firm to digitise and ‘datafy’ activities before their rivals do.
We worked on an initiative with San Francisco Park to digitise on-street parking in the city. We worked with the city, some academic folks and partners to put out smart meters for parking, as well as a method to pay, then placed sensors in parking spaces in order to find out what happens to parking when you modulate the price based on demand.
The average price of parking fell 4 per cent in the neighbourhoods where it was used, and the target occupancy rate of 80 per cent was hit 31 per cent more frequently. All of this was made possible by the insight that there was information being lost from the activities around parking. This is the talent companies have to develop – looking out at the world and seeing the data that’s not there.
The second principle of data as capital is that data tends to make more data. This is where you’re feeding the data you capture into analytics and algorithms in order to use it. When that data gets used, you create more data and feed it back into the systems to improve things further.
In the case of SF Park, in the initial phase of the program, the average price of parking only fell 1 per cent. But by capturing the data about sensitivity to the price changes, we could further improve the system and achieve a 4 per cent reduction.
The idea that goes along with this is experimentation. It sounds very risky to some managers. As a bank, for instance, what if you try a new fraud detection algorithm and it starts flagging legitimate transactions, causing customer satisfaction to tank? That is a real risk and concern. But what if instead, you try the new algorithm with just a small subset of transactions and watch it like a hawk? You set a threshold on the increase in flags, and if you hit it, you turn it off and go back to what we had before. That’s the approach companies need to embrace.
The third principle of data capital is that platforms tend to win. That’s where it concludes, and this has big implications for competitive strategy for established firms.
Many organisations are still struggling to do something with the data they’ve already got. Why is this still an issue?
There is this real issue around a skills shortage for analysing this data and making new use of it. The way we approach that at Oracle is to make better products. One of these is Big Data Discovery. Its purpose is to help people who have questions but not necessarily data expertise shop for data sets inside the company.
More and more of our customers are concerned and don’t feel like they have their data house in order, but recognise they must do something with big data that’s new. Increasingly, a number are taking the approach of setting up a big Hadoop cluster, and then pouring in diverse data sets. They don’t know what it’s going to be worth or what they’re going to do with it, but they’re putting it into an environment that requires little prep for holding diverse data, then putting a discovery and exploration environment on top.
With Big Data Discovery, we provide a visual environment for shopping for data sets in Hadoop. Just like Target online has guided navigation and allows you to browse and search, this allows you to browse and search data and once you see something you like, you can look up the attributes, how many rows and values, and so on. This is one way to bridge the gap.
How do organisations bring data together when there are so many data owners inside an organisation?
It becomes a human relationship exercise to go and find your colleagues who have a question of some kind, but know they need data sources from different parts of the organisation to answer it. The key is to make it as easy as possible for owners to contribute what they have before they change their minds. Then you need to tell a great story off that.
One example we’ve seen is a small US college called Valdosda State University. They had a real mission to increase graduation rates of students from vulnerable populations. The guy who ran the program went around the college, and begged and borrowed data. There was data on badging people in and out of buildings, data from mealcards used in the cafeteria, course enrolment and matriculation, people who had data about actual grades received. He had a story to tell. That human relationship is a big part of these data exercises.
Are there specific data sets or types people aren’t utilising anywhere near enough yet?
It depends on the domain you’re in. Within manufacturing, for example, there is an enormous set of data about the complicated supply networks these companies require. A lot of data captured about shipping products from point A to point B gets used by the shipper. Then there’s data about stuff sitting on the dock, or how long it takes to get to the manufacturing floor, and it gets used by the manager of that facility. But that data could be incredibly rich in root cause analysis, in changing the way these manufacturers handle warranty claims, if only it were brought together.
How about the whole social strata – is that data as useful as we all think it is?
I think with all data, there is a trade-off. It can be used for good or for ill. And there are enormous potential benefits to society from social data. But there’s also a lot of hard thinking to be done around how we secure personally identifiable information, how we prevent unacceptable use, and even what informed consent means in a world where there is potentially enormous value in the secondary uses of data.
We’re seeing increasing number of organisations buying external data sets – according to IDC, 70 per cent of larger organisations are already doing so. There’s also a rise in data sharing between two companies to pool their data. Do you see this trend continuing?
One of the implications of data capital is increased trade of it. We see this now with data for targeting advertising and for media buyers. But we also expect to see an increase in the buying and selling of data like from supply chains, for example, and even from within the healthcare system.
One of the things that needs to be created is public data exchanges that handle contracting around data. A lot of companies may be interested in selling their data, as long as they are assured it only gets used by those trading partners they’re comfortable using it. That’s where we’ll see new data marketplaces spring up.
In light of these processes and demand, do you expect more dedicated chief data officers to be appointed?
Without question. The chief data officer’s role is much like the CFO’s role: Make sure the data assets are used to their full potential, while still staying compliant with policies and procedures about privacy, security and appropriate use. Data chiefs who think of themselves as building new products and services with the data will be incredibly powerful over the next 10 years.
This role really comes down to skills. Those can be resident in one of those other c-level roles – in the CMO, CIO or digital officer. In consumer-facing businesses, the chief digital officer is probably going to have the lion’s share of the responsibility and authority about how to use the company’s data assets more effectively. But in B2B, it may a very innovative CIO playing that chief data officer role. It’s equally probable someone comes out of product design, who really is thinking about innovating with data, creating new capabilities in the marketplace. They could be become a partner to the CIO.
There are different kinds of chief data officers, too. Some will be out of risk and compliance, such as in financial services firms. The chief digital officer tends to be at the head of marketing and customer insights, and you see that in consumer packaging, retailers and media companies. Then some will come out of operations. Where we’re seeing these types is in manufacturing.