Data is the gold of the 21st-century – Establishing data markets might help to make it work for everyone

Max Tegmark, professor of physics at the MIT, points out in his latest book (Life 3.0 – Being human in the age of artificial intelligence) that “in the future, the only traded resource will be knowledge” because knowledge-driven technology will be able to create all sorts of matter just by reassembling atoms. That sounds far-fetched but even today companies that harvest and utilize data, the precursor of knowledge, are among the highest-valued firms globally. However, you don’t see price tags on all the data that is floating on the internet, because markets have not yet developed. I will inform you from an economic perspective why that hasn’t and why it actually should. 

Of course, already today you trade your data in exchange for services. If that hits you by surprise, I recommend a short reading of the terms of service on your Facebook account (if you still have one). Yet, given the margins of big tech companies, the deal could be more profitable for producers and also more transparent. Markets could help to distribute data’s benefits more equally and make data trade less opaque.  

Researchers at the University of Amsterdam approached the issue of lacking data markets from a scientific perspective. They defined data commoditization and market mechanisms by proposing six data properties that could pave the way for building solid data markets.

1. Data sovereignty addresses data ownership. Compared to oil, data, as an economic good, is non-rivalry, meaning that data can be copied and shared infinitely.
2. Trustworthiness (or trusted data) refers to data being verifiable and auditable. Data is becoming a decisive element of automated decision-making and therefore it has to be trusted. A consumer at a gas station does not have to check the gas for its quality because all parties involved in the gas value chain have agreed on industry standards.
3. Data reusability ensures that data is stored and can be gathered for future projects and applications. Well, oil doesn’t make such a great job here either. 
4. Actionability demands that data purchased by a company is directly applicable to its value chain. Meaning companies should be able to assess the economic gains or savings through data acquisition before purchasing the data. In the oil industry, companies have estimates for the returns they can expect from acquiring a specific amount of oil through a trade market and established value chains. 
5. Finally, measurability refers to the valuation of data. There are different approaches to conduct the pricing of data.
   1. First, there is the cost-based method, which is based on the idea that data creation, sharing, storage, and analysis are costly and therefore should determine the price.
   1. Second, there is problem-based pricing where the consumer sets a price and the providers react upon. Current examples for the latter are tournaments on data science platforms such as Kaggle.
   1. Also, the price depends on the data quality which hinges on multiple factors. For instance, the number of variables and cases, the precision, the accuracy, the actuality, and the temporal resolution of a data set. In an existing market for data, pricing would be much easier through processes of comparison with similar data assets.

What is the current status across significant industries? A look into the (European) perspective.

Currently, data commoditization is mostly pursued by huge tech companies. As they do both the data harvesting and monetization, there is little incentive for them to engage in the creation of transparent data markets. In Europe, the European Union has set the agenda for the creation of open data markets to facilitate digital transformation. To not fall further behind the curve, industries sitting on huge piles of data, e.g. incumbent banks, should embrace the establishment of open data markets as an opportunity. 

by Jonas Röttger, FINDER ESR

Sources:

Demchenko, Y., Los, W., & de Laat, C. (2018). Data as Economic Goods: Definitions, Properties, Challenges, Enabling Technologies for Future Data Markets. ITU Journal: ICT Discoveries, Special Issue “Data for Goods”.