Data Integration Blog -

I first read about “the democratization of data” a few years ago in a Google blog by Hal Varian, Google’s Chief Economist. The use of the word “democratization” is a little misleading, here. It’s not about data getting to cast votes and elect representatives. It’s about the consumer being able to consume data in a way that previously only organizations with big IT budgets could.

If you look at a set of technology price graphs, the cost of everything in IT has fallen dramatically over decades – the cost has collapsed. It fell with the proliferation of PCs and fell further as the capabilities of PCs increased. The cost of PC software held steady for a while, but then Open Source software emerged and put a downward pressure on the price. And then recently, cloud computing exploded and suddenly it was possible to rent servers or software at a very low cost per day – or even per hour.

In effect, the financial barriers to using just about any kind of software and computer resources, including those used by large organizations, have fallen to the point where Joe Public can, if he wants, do many of the things that large organizations can do. Certainly, he can collect data and analyze data. But does he want to? Probably not.

Towards the Monetization of Data

The cost of information has fallen dramatically, too, of course. But it’s a patchy picture. There are exabytes of data out there on the web and most of it available for free. Wikipedia has killed the encyclopedia business. News web sites and aggregators have been gradually putting newspapers out of business for years and the electronic book, courtesy of Amazon, is starting to eat up the book publishing business. But, with some exceptions, there isn’t a mass of data out there for free. Admittedly you can get megabytes of historic stock prices from Yahoo and you can also get a fair amount of free government data if you want it.

But while data may be cheap to store and process, it isn’t cheap to collect, clean and maintain. And it’s only cheap to analyze if you know how to analyze it – and you’re not likely to do that unless there’s value in doing so. So right now, Joe Public doesn’t do that, both because he sees no point in doing it and because he probably has no idea how to do it.

So instead organizations do the work for him. Consider, as examples, LinkedIn, Flikr and Facebook. All of these web sites are backed by serious amounts of hardware resource and their primary purpose is to collect data for the sake of sharing it and turning a profit. They have different business models, charging for some of their services or funding them with advertising or a mixture of both. It works fine and it relieves Joe Public from having to do much. They monetize data by aggregating data, and reward Joe Public with some free benefits.

And if Joe Public really would like to analyze his own data there are many web sites that can help him out, most of them retail sites that are eager to sell to him. Netflix, for example, does a pretty good job of determining which movies Joe likes. Pandora and iTunes are both good for analyzing his music tastes and Amazon is good at recommending him books to read.

All of this leads me to conclude that Joe Public’s data isn’t being democratized in any meaningful way. It’s being monetized.

The Monetization of Data as an Opportunity

And that leads me to what is an emerging opportunity for a least some businesses. On the one hand we have organizations gathering data on their own behalf for their own benefit. That data is being exploited by them, hopefully in a useful way. But it is not being used of itself to generate profit. For example a retail chain may gather terabytes of data as to what Joe Public chooses to buy and what special offers work for him and what product placing on shelves is most effective and so on. And armed with such aggregated data, it can improve its operation.

But such data may be immensely useful to other businesses, particularly those that make the products it sells. They could make use of that data to improve their own sales through that retail outlet and through others. Similarly their competitors might be interested in that data. So the data can be rented out – and in some cases, but not many, it is.

Now consider the fact that the organizations that permeate the economy, including of course the various arms of government, also have data that could in some circumstances be rented out for analysis. And of course the social network web sites. They too gather vast amounts of personal data. And that too could be rented out for analysis.

Now consider the idea of mashing up that data. For example, correlating the music Joe Public likes with his tendency to buy certain types of insurance or to take a particular type of vacation. What I’m talking about here is cross correlation between sets of data that have never been data mined in that way. Without a shadow of a doubt there is gold in them there hills of data. And there is lots of data that no-one has yet tried to monetize.

But there’s a problem with this. The question is “who owns Joe Public’s data?” Monetizing data is fine until it gets personal, at which point, maybe Joe Public doesn’t want his data analyzed. Maybe he doesn’t want any business to share his data. Or maybe Joe feels that he should be rewarded if his data is used in any such analysis.

This is a political issue. At the moment there’s little to stop anyone trading in Joe Public’s data, but there is a potential for such data to be used to Joe’s disadvantage – even to be used criminally. So I suspect that ultimately the monetization of data will have to be regulated. Right now there’s very little regulation that protects Joe’s data from being monetized and traded on the open market.