Is data a commodity?
This is my dataset. There are many like it, but this one is mine.
Apparently, the first thing that happens when one announces a theme of the month is that other topics rush to the front and demand attention. So while attempting to focus on the alt data market, I realized that the roles of the intermediaries—the consultants, scouts, advisers, marketplaces—are fundamental to understanding the data market. And besides, the short answer on alt data is that they’re interested in anything and everything if it might deliver useful trading signals.
I also realized that I’m going to keep coming back to the financial sector as a customer as I go through all of the specialist data categories. That’s just data monetization strategy meeting Sutton’s Law.
Is data a commodity?
Collecting sources for a view of the alternative data market (which is mostly hedge funds and investment banks looking for data that will help them outperform the market), I have a lot of tabs open. There are some great newsletters, and it doesn’t take many to generate an overwhelming set of rabbit holes to explore.
One of those tabs was open to Todd Harbour’s proposal on evolving the trade in data from informal marketplaces to a formal commodity exchange. I’m glad he’s done the exercise, even if I don’t quite accept the idea.
Right off the bat, I wonder: is data a commodity? We can standardize the technical layer (formats and connectors), but can we standardize the information content in a way that meets the expectations of a commodity? Most products aren’t sold as commodities, so why data?
For some categories, I can sort of see it. Many competing providers offer similar datasets that might answer the same questions—consumer data seems like a candidate, for one. But should commoditized data be standardized around content (what fields are in the source) or purpose (what questions are answered)? If filtering a dataset can turn it into a different product before it goes to market, is that commodity behavior?
Other categories of data come from scarce resources (such as satellites) or exclusive sources (pick a major social media platform). Does being a tradable commodity depend on having many suppliers in the market? Does the model fail if there are too few sellers?
I think the missing analysis is cui bono—who benefits in the shift to a commodities-style market, and who’s better off in the current enviroment? What are the knock-on effects of commoditizing data, assuming it’s possible? Everything’s a tradeoff; what’s the cost of the proposed benefits?
Mostly, my immediate reaction to the idea of trading data as a commodity is that companies generally try to resist commoditization of their products and services (it’s not every day that my initial reaction references an article from 1998, but it’s a good one; an update today could add the data wrapping strategy for more ways to avoid commoditization).
I do recommend the article, Data as an Unacknowledged Commodity: The Case for a Formal Data Exchange. It’s a thoughtful and thorough analysis of many of the elements worth considering, and it’s a good starting point for a longer discussion on how market structure and regulation could help address some of the well-known issues with data sharing. Agree or disagree, it’s nice to see the logic that supports his thinking.
Refocus on the supply network
Life keeps reminding me that words often mean different things to different people. Markets and marketplaces are no exception. I tend to use market the way I remember it from undergrad economics courses, as a broader construct involving buyers and sellers, the theoretical space in which homo economicus makes rational decisions based on perfect information, or in the real world, the abstraction of aggregate supply and demand. Say market to others, and some expect a certain… structure.
Marketplace is looking similarly challenging, since every cloud provider seems to provide one, specialist intermediaries operate others, and there’s a broader abstraction there, too.
I knew terminology and the ontology of market participants would be important. More than ever, it’s looking like the first task. While battling constant distractions from all these fascinating data stories, of course.
Inspiration
Horizontal ecosystems at the intersection of data and… everything?
Added to the must-read list: Data Harbour and It's Pronounced Data
Discovery
A Suspicious Pattern Alarming the Ukrainian Military – Is the Russian military using civilian satellite platforms for pre- and post-attack intelligence?
Digital Exhaust Opt Out Guide – When the resource is in the last paragraph of the article.
Technology, venture capital, meet industrial policy.
TBR
Co-Intelligence: Living and Working with AI – Hoping for a well-balanced take—optimistic but not foolishly so. From Wharton prof Ethan Mollick.
How to Win an Information War: The Propagandist Who Outwitted Hitler – If those who cannot remember the past are doomed to repeat it, can those who remember the past do better?
Squirrel!
Why we’re not going to get ornithopters.