Visa credit cardsMasterCard credit cardsAmerican Express credit cardsDiscover credit cards

Wednesday, April 23rd 2014


Scared of Big Brother? Too late, says 'Big Data' co-author

Large-scale predictive analysis is here to stay, with many implications


Forget about trying to keep your personal details private in today's "big data" age, says Viktor Mayer-Schönberger, co-author of the book "Big Data: A Revolution That Will Transform How We Live, Work and Think."

Organizations ranging from credit card companies to nonprofit hospitals already have access to an unprecedented number of facts about the way you live, thanks to a trail of information you leave behind when you go about your everyday activities.

As more companies tap the power of the data available to them, consumers will have less control over the personal details that are amassed by third parties and used to make decisions that affect them, says Mayer-Schönberger, a professor of Internet Governance and Regulation at the University of Oxford, who co-wrote the book, released in March, with Kenneth Cukier, data editor of The Economist. 

Collecting vast quantities of information about you -- and the people that you interact with -- allows companies to better predict who you are, what products you prefer and what action you're likely to take next, says Mayer-Schönberger.

That could lead to a wide range of consumer-friendly innovations if companies use that information responsibly, he says. But it also means that we are entering a world in which privacy is a thing of the past -- and everyone from your neighborhood lender to the federal government knows far more about you than they ever have before.  

Viktor Mayer-Schönberger,
co-author, 'Big Data: A Revolution That Will Transform How We Live, Work and Think'
Viktor Mayer-Schonberger, 'Big Data'
Big Data: A Revolution That Will Transform How We Live,  Work and Think

As consumer data gathering techniques evolve, people are becoming increasingly concerned. However, "Big Data" co-author Viktor Mayer-Schönberger explains that while many fear their privacy is being invaded, the end result may be products and services that can save us money and make our lives easier. caught up with Mayer-Schönberger and talked with him about the massive changes afoot. The term "big data" gets tossed around frequently. What exactly does it mean?

Viktor Mayer-Schönberger: One way of looking at it is that "big data" means that we can do analysis -- we can gain insight -- with a lot of data that we can't with a little data. What are some examples of that data?

Mayer-Schönberger: That could be anything. It could be as mundane a thing as search terms sent to Google. It could be something quite sophisticated like vibrations given off an engine, either a jet engine or a car engine, that is being captured and then analyzed ... The truth is that more and more of the reality that surrounds us, we are able today to translate into data. That's what we call "datafication." ... [For example,] I now have an application on my iPhone that records when I walk, when I bike and when I'm in the car every day. And I don't have to tell it that I'm now biking versus walking. It just knows it by my movement, by the acceleration and so forth ... That is only possible because with location and with acceleration being datafied, inferences can be made about what I am doing at a particular point in time. And that's the kind of information that other companies would love to have and are trying to gather, isn't it?

Mayer-Schönberger: Indeed. The thing is, 20 or 30 years ago, PR agencies would say, you cannot not communicate. Today, it is you cannot not leave a data trail. Basically, whatever we do today, we leave a data trail behind and that data, in itself, has some value -- some hidden value.

And if we try to not leave a data trail behind, it just becomes quite costly. You know, just think about being unable to order anything online because you would have to pay in cash in order to not leave a data trail, or to pick things up, because the moment you have it delivered to your home, you leave behind your address and then your telephone number and other information. So we generate a lot of data knowingly and a lot of data also is generated unknowingly. You write that more data is really better data for companies. Why is gathering so much information about the world and the people in it so powerful to them?

Mayer-Schönberger: Companies, when you look at it, want to understand reality better. They want to understand what their customers want, and they want to understand better what the competition does ... So what they can do is take the data and make really good predictions. And to have really good predictions is just extremely powerful ... [With] more data, we just get better clarity and better detail. And that's just enormously more valuable, compared to what we had before. It lets us predict infections in preemies 24 hours before they happen. And it lets us predict whether people are going to take their medication on time or not, just by whether or not they pay back their mortgage on time. All these kinds of things are predictions that make commercial transactions more efficient, but also improve services and improve products. And this is something that companies, such as credit card companies and credit bureaus, which are sitting on huge amounts of data, have been experimenting with for years.  

Mayer-Schönberger: That's absolutely right. If you look at Visa or MasterCard, they have billions and billions of data points that they sit on, and analyzing them was almost impossible just 10 years, 15 years ago, because of the enormous storage costs and then the enormous amount of processing power needed to analyze it ...[It used to take weeks to analyze a credit card transaction.] They are now down to just minutes. And that is, of course, much more powerful because then the credit card companies, or any other companies that have asked for the data, can now ask hypotheses -- can ask questions -- to the data and don't have to wait three weeks for the answer. You also write about accepting the "messiness" of big data. It's often chaotic and incomplete and can even be inaccurate, but the more of it there is, the more likely it is to be right. Can you explain that?

Mayer-Schönberger: In the small data age, we only had limited amounts of data available and collecting data was extremely costly and time-consuming, and so we limited the number of data points that we'd connect because it was just so cumbersome.

If we try to not leave a data trail behind, it just becomes quite costly. You know, just think about being unable to order anything online because you would have to pay in cash in order to not leave a data trail, or to pick things up, because the moment you have it delivered to your home, you leave behind your address and then your telephone number and other information.

But because we only had a few data points, a dozen or two dozen or three dozen, we needed to be extremely careful in connecting every one of them because if you have 50 data points and one is really off the charts, your result will be wrong and your decision that is derived from the result will be terribly wrong.

In the big data age, when we have enormous amounts of data points, and it's just a couple thousand that are wrong, so be it ... The value of the rest of the enormous amounts will just drown out the small mistakes. It's interesting, because that idea that you have to accept messiness in order to accept the power of big data is a tough concept for people to wrap their minds around. I spoke with one banking representative who said that his bank was considering using alternative data beyond what's in a traditional credit report, but he was skeptical of the reliability of that information because it's not consistently reported. Similarly, I've also spoken with people within the financial services industry who are skeptical of using social media data for credit scores because one source told me computers don't get context well. Are they missing something?

Mayer-Schönberger: I think they are missing something ... Using the example of FICO's medical adherence score when we look at data, we say, "Oh, why should financial data be able to predict whether or not somebody is taking a medication on time? That's ridiculous." But the truth is that behaviorally speaking, we can see that people behave consistently and that means that you can use correlation to establish whether or not somebody is taking their medication by looking at whether somebody is paying back his mortgage or her loan on time.

And it might sound strange, but if it serves the purpose, if it solves the problem at hand, then we should absolutely use it because that will give us better insight ... If we don't use these insights, often times we have hunches and these hunches are often times wrong ... There's a lot of value for the financial services industry to be had from social media data and from many other data sources that are not directly related to our financial behavior. Because it turns out that humans do behave relatively consistently ... At the same time, we need to be careful about these correlations. We need to do our analysis right. We must not forget that at the end of the day, we must remain the masters of our interpretation of the data and not succumb to the lure of data and not be dictated and controlled by the data we use. One thing that people might find a little disturbing is that they might be judged by what people similar to them or connected to them are doing, rather than what they actually do.

Mayer-Schönberger: Yes. And it's kind of saying that our lives are more interconnected than we realize.

Mayer-Schönberger: Absolutely. But that is true, not just with social media data. Think about it from the perspective of genetic information as well ... when your sibling or your kids or your parent has his or her DNA analyzed, he or she is not just revealing data about themselves, but about you ... So, in that sense, the people that we have as friends on social media are revealing a lot about us ... The truth is that human beings are just, almost by necessity, social animals. And that means that it comes with a lot of upsides, but it also means that those people that we hang out with reflect, to an extent, the values that we have. What happens when the data's wrong? For example, even though you may share genetic traits with your parents, it doesn't necessarily mean you'll inherit the same diseases.

Mayer-Schönberger: Absolutely. But again, that is just understanding the limitations of the implications of big data. Big data is predictive, but it's also probabilistic.

So it cannot predict you. It can only predict that out of 100 people like you, 70 act that way. You might be among the 30 percent that don't act that way. And therefore we must always be careful whenever we make a decision based on prediction and probabilistic prediction that we let people show that they belong to the 30 percent rather than the 70 percent ... We must not forget that predictions can and will be wrong in any number of cases.

As we move into the big data age, we need to evolve our thinking from thinking in terms of privacy and control of personal information over to thinking about what is a good use of our personal information and what is a bad use of our personal information. So we're entering a world in which data-driven predictions are, to some extent, taking away our power as individuals to exert control over our circumstances? For example, one analogy I was thinking about when I was reading the book would be credit reports. People know what's in their credit reports right now and they know what actions they can take to improve their credit scores based on what's in the reports. But if lenders and credit scoring agencies are using all kinds of data -- from who's living next to them to how often they move to the value of their home -- to make decisions, how are people supposed to control all of this? Is this something that we just have to accept about the world we are living in now?

Mayer-Schönberger: I think that it's not about control, but it is to an extent about transparency and about the ability to do something about it ... Even today, when we think about credit reports and credit scores, the credit bureaus only give very general information about the factors used for credit scores. So they don't tell you exactly what you have to do in order to improve your score precisely. So even today, we really don't have 'control.' A lot of companies seem to be keeping their experiments with big data quiet for fear of consumer backlash. However, privacy doesn't really exist anymore anyway, does it? Our data is already out there.

Mayer-Schönberger: Well, the desire to keep information private continues to exist. There's no question about that. But you're right, there's so much data out there, it would actually be better to be open about it.

It would actually be better to be transparent about all these experiments because society might find value in it and might begin to like it and might be willing to accept that sometimes there is a trade-off to make between additional insights that lead to better services or lower costs versus informational privacy. Do you have any advice for the everyday consumer who's concerned about all this? How are we to face and get used to this kind of brave new world in which our details are getting collected and sold to perfect strangers?

Mayer-Schönberger: In a way, the idea of privacy, the idea of controlling information and controlling every single piece of personal information -- who I have given it to, for what purpose it can used and so forth -- that is thinking in the small data age. If you only have 15 data points about you, like your Social Security number, your date of birth and so forth, then you can kind of protect that, or at least hope to protect that.

But if there are a billion data points about you out there -- like how you walk, where you were, what was your location at any given point in time, who did you call and so forth -- it doesn't make much sense to try and control every individual data point yourself.

So what I would suggest is, as we move into the big data age, we need to evolve our thinking from thinking in terms of privacy and control of personal information over to thinking about what is a good use of our personal information and what is a bad use of our personal information. And try to be open for experimentation there because that will show that there are a lot of very potent and very powerful uses of personal information. That's not only a power for the credit card industry or the financial services industry, but a power for consumers as well.

See related: Congress asks if data brokers invade consumers' privacy, FTC privacy report backs more credit data access for consumers, Debt collectors' use of social media raises concerns

Published: May 2, 2013

If you are commenting using a Facebook account, your profile information may be displayed with your comment depending on your privacy settings. By leaving the 'Post to Facebook' box selected, your comment will be published to your Facebook profile in addition to the space below.

Three most recent Spotlight: People and their plastic stories:

Share This Story

Follow Us!


Credit Card Rate Report

Updated: 04-23-2014

National Average 15.00%
Low Interest 10.37%
Balance Transfer 12.64%
Business 12.80%
Student 13.27%
Cash Back 14.84%
Reward 14.96%
Airline 15.30%
Bad Credit 22.73%
Instant Approval 28.00%

USA   |   UK   |   Australia   |   Canada