I feel I don't need to add any sick burns about how the phrase 'data is the new oil' is a facile, capitalistic examination of digital power dynamics. But one thing I used to think about is the idea that data is a commodity. Even the people who reject that data is like oil, will talk about data as having inherent trade-able value. They might talk about having data as a proxy for having power.
So, the first time a friend said with frustrated fervour 'data is not a commodity', I looked up the word commodity to remind myself of what it means. Both definitions from the top of google raise... questions.
The first: a commodity is something that has a set value and can be traded on an open market — think gold or copper.
The second: a useful or valuable thing.
The more I thought about it, the deeper I shared the sentiment that data is not a commodity.
Let's start with the first definition: data is not oil or gold. Why? Because it doesn't hold universal value. If I give you a spreadsheet, and you wanted to sell it, how would you put a price on it? Would you count the rows? Count the columns? Of course not. You might look at the data, but really you would probably think: what could someone do with this? How might this data be useful?
If I gave you a barrel of crude oil, you'd look up the price, and then you'd find a refinery and see how much they would charge to turn it into oil that could be sold.
So in dissecting that first definition of 'commodity', we can see how data does not really fit.
Let's take the second definition: something that has 'value' or is useful. Data only has value if you know what to do with it and have the resources to do something with it. If I had a bunch of spreadsheets which detail every retail price of one product in every store nationwide, and sent all of those over to my mom, we would probably have a long conversation about how to open Finder on her laptop. But, if I handed the same spreadsheets over to Amazon — a leading corporation in retail and logistics — they would turn that data into an inviolable market position to fix prices by area for that product, and to always undercut the brick and mortar competition but without operating at a loss anywhere. Amazon could turn that data into millions of dollars.
But the platform economy does treat data like a commodity — which is what can make us believe it is one. It does so in at least three interesting ways.
- It leans on a complex division of labour and robust set of infrastructure that can turn one thing into another, with a system that is mostly predictable in terms of cost. So, if I run a small business, I don't have to have an oil refinery in my backyard to turn crude oil into oil that can power my delivery vans. I also don't have to run an ISP, or even a server to host my website.
- I can do incredibly complex things, without actually engaging in the complexity to leverage that thing. In other words, readers, I have all of your email addresses, and that data point enables me to get in touch with you for free. It's a complex system which is made simple for my (and your) convenience.
- It allows for an accrual of value in some places, and an accrual of harm in other places. And in most cases, never the two shall meet. No one that trades 'gold' on the open market cares about where it was mined, and there is very little connection between the sale price of gold and the devastation of communities by those who mine and sell it. Those capitalising off of data and data systems, are in a very different place (physically, economically, politically) than those who are harmed by them.
These are interesting characteristics of data, and if we explore the way data can be commodified it might be helpful for understanding some macro systems and patterns. But when we get literal, and think about data as a commodity, we undermine our ability to understand power and value in data systems. It can be so tempting to reduce the complexity of our world in order to increase understanding, but sometimes reducing complexity means you are no longer connected with the very reality you are trying to understand. Even though data can be commodified in some circumstances, let's not dumb things down to thinking it's a commodity.
And if data is not a commodity — what is it? No, really. Hit me.