nav search
Data Center Software Security Transformation DevOps Business Personal Tech Science Emergent Tech Bootnotes BOFH

What is the REAL value of your precious, precious data?

Why your personal information isn't as important as you think

By Tim Worstall, 22 Apr 2015

Data Pair – Part 2 Multitudes are getting very excited about what all of this data flowing around the system is worth. If we can know lots and lots about lots and lots of things then obviously that's really valuable, yes?

And to some extent this is even true. We're not just in a bubble here – the more we can do with data these days, the more that data is worth. However, it's worth possibly unpacking how and why data is valuable.

The first thing we need to get away from is some idea of what data “ought” to be worth. Economics is a positive subject, not a normative one. We therefore look at what something is actually worth, not what it ought to be in some parallel universe where the chakras are aligned. This is, to some extent, the point that some of our continental colleagues are missing.

The thinking seems to be that we must protect European information from being sucked up by American corporations because it is valuable and we Europeans should exploit that value. Or even that consumers should not have that value taken from them. This is rather ignoring the point that this information has value only because it is being sucked up by those companies that have the wit and ability to process it. The same data that Amazon sucks up about your buying habits is entirely valueless if sucked up by, say, Waterstones, who wouldn't know how to process it or even what to do with the results if they did.

We cannot therefore say, prima facie, that data or information has “a value”. And we most certainly cannot say that data ought to have such and such a value (OK, we can, but that's metaphysics, not economics or finance).

Value is “value to whom” and that also depends upon the time and place. Knowing the winner of the 3.30 race at Kempton is rather more valuable at 3.15 that day than it is at 3.45 that day. Being able to predict something by processing data is thus perhaps more valuable than knowing as a fact something that has already happened. The value to whom is further important if we consider the gambler and the puritan who insists that betting is for the very devil.

Finally, our value-to-whom depends upon who is the supplier of that information and who is the user of it. Quite obviously there must be a difference of valuation here or there never would be the transfer of that information. This is true of any market: I and John only exchange apples for pears because, at the margin, we think that the last apple/pear we have is worth less than the next pear/apple we will have.

So it is with information and data. A vast number of people think the data about their last book purchase, the most recent online ad they looked at or the web pages they have been visiting is of zero value. A very much smaller number of aggregators of such information think that there's great value in it. So much so that they're willing to spend a few billion dollars here and there to provide a service (like, say, a social media network or a search engine) just to collect that information. It's this difference in valuation that drives the transaction.

Take it. It's useless to me

Finally, we've got the idea of the value of each piece of information and the value of the aggregate pile of all that information. One single piece of information, one data point, is worth – perhaps – nothing at all. It's the agglomeration of millions or billions of such data points into an information flow that enables us to try and assign a value to what we can find out from it. And it's this last that really has the value and is what we can rightly call Big Data.

At which point, of course, we find ourselves in something of a bind. For if it's the process and the accumulation that is adding the value, then what value can we assign to each individual part of that flow? It's a bit like assigning value to a rivet on a bridge. Sure, the bridge is valuable, the accumulated effect of all the rivets is what gives the bridge its value. But what would we pay to ensure that one specific (or any single) rivet were there and working? Probably nothing, given that there's redundancy in such things.

This is, I think, where author Jaron Lanier's ideas from two years back about people being paid for their information falls down. So, too, where certain Eurocrats fail in their insistence that European data is valuable.

We have a number of different possible estimations of value here. What do you have to pay someone to give the information? Not a lot, as it turns out: provide a webpage and people will tell you all sorts of things. So the information, data, in the hands of the originator seems to have a value not far above zero, if it is above zero.

Yet at the other end of the system there's people making billions by manipulating the accumulated and aggregated data. So this obviously does all have some value. But it seems to be the process of collecting, accumulating, aggregating and then parsing this data flow – that is the part of the process that is adding the value.

Thus, logically, the value of doing all of that should be going to the people who are doing it. And the more Marxist we are, the more we should think this, perhaps. The value is being created by the data engineers, not by the originators. The value should, as with all other work by hand or brain, accrue to the data engineers.

Another way to reach the same point is that in a market, things are worth whatever anyone is willing to pay for them, whatever someone is willing to try and charge for them. As no-one buys the data itself, no-one successfully sells it on an individual basis, so the data itself has no individual value at all.

It is the data stream that has value. While we can express that as a value assigned to each contributor to the stream, that's really not the way to look at it. It would be like a surfer trying to assign a value to each water molecule involved in the crestion of a wave. It's the system itself that is the value, not the components of it, so the value must be assigned to the system, not the components.

There’s absolutely no doubt at all that there is value in Big Data. It's just that the value seems to be in the “Big” part and thus there's little to no value that can be assigned to the individual data points. ®

The Register - Independent news and views for the tech community. Part of Situation Publishing