Why the Term 'Metadata' Is not Helpful in Discussions Around Privacy

17 Apr, 2025

Facebook has changed its corporate name to Meta as part of a major rebrand. The company said it would better "encompass" what it does, as it broadens its reach beyond social media into areas like virtual reality (VR). BBC Headline 28 October 2021

When I published my glossary post What Is Behavioural Surplus? this month, I received the comment: is that not just what we call metadata ?

After giving it some thought, I want to propose that the term metadata is not only an inaccurate description of behavioural surplus, it also downplays the severity of the incursion on our privacy that is happening on such a large scale today.

In the article, I explained scholar Shoshana Zuboff's definition of behavioural surplus as presented in her book The Age of Surveillance Capitalism. I recommend reading that post first, if you have not done so already.

There are two problems with using the term metadata to describe behavioural surplus:

Metadata is created and used purposefully in order to categorise and understand the primary data, while behavioural surplus is used for commercial purposes unrelated to the primary data
Metadata sounds abstract and innocuous, while behavioural surplus is currently being used as a privacy invasion tool for profit on a massive scale

But first we need to look at the meaning of the prefix meta.

What is meta?

The word meta means about, after or transcending. A meta research paper looks at sometimes 100s of published research papers and attempts to describe commonalities and conclusions drawn about that field of research. In the field of education, there is a famous study that looks at all the research that has been done on effective in learning in schools, and then lists all factors, tools and methods, from most effective down to downright harmful. Interestingly, investing in improving the school environment has very little impact on learning, according to this study, but teacher expectations score high. This is an example of a piece of meta research, or, research about research.

Another example of meta can be found in storytelling. One of my favourite video games, Nier Automata, has you play through the same story a number of times using different video game genres (first person action, platforming, arcade, etc.). As you repeat the same story from different perspectives and game-modes, you are encouraged to consider essential questions about video games and storytelling, such as, who are the heroes when there is violent killing? Nier Automatais a meta video game, a game that comments on what video games are by having you play a game.

These two examples may feel abstract and philosophical, but there is a type of metadata most of us know well: a library catalogue. I want to use this example to illustrate my first point about why metadata is not an accurate term to describe the recording and sale of our behavioural data.

Librarians use systems to categorise and describe products that contain lots of data. The clearest example of this is the library catalogue (which used to be written on cards ordered in long sliding drawers), where each entry tells you the author, the title, the date and the publisher of a book, film or piece of music.

Beyond that, librarians and other catalogue creators can add tags to describe how these texts can be grouped, for example, by genre, theme, topic or age appropriateness.

These metadata are deliberately put together, and sometimes require a type of intelligent analysis in order to place things in the right category. What are the Star Wars films for example? Science fiction or fantasy? (These films are not really about science at all).

Metadata like those found in a library catalogue have a clear function that relates to the source data (the books), which, in this case is to help readers find the book they want, or to learn more about a book without having to read it all.

The behavioural surplus that comes from our interactions with an application (time, place, typing speed, spelling errors, repetitions of searches, etc.) could be described as a type of metadata but that wouldn't tell the full story.

The type of data about us that are scraped and sold for behaviour prediction are a byproduct of the digital interaction with a tool, like a search engine; they do not serve a clear primary purpose of helping to access the primary data (our searches). To come back to the example of the library, it would be a bit like selling the library catalogue to external companies, a key difference being that books can't give consent.

Therefore, metadata is inaccurate and cannot be seen as synonymous with behavioural surplus in the context of surveillance capitalism.

Metadata sounds abstract and harmless

Reading up on the meaning of meta, my eyes sometimes glazed over as I skimmed various paragraphs and definitions on meta culture, meta humour, and so on. It's an abstract concept, difficult to grasp, similar to a concept like irony.

Secondly, the word metadata has a light, invisible quality to it. You can easily skip it. You don't really need to worry about it.

As Zuboff explains in her book, the scraping of behavioural surplus for profit is something serious that we do need to worry about. Companies are designing new physical products in order to obtain even more personal data on us. They are also discovering these predictions can be used to influence behaviour in the real world, or remotely shut down your car if your driving is not up to the standards of your insurance company.

The importance of language

Language is key in the fight for privacy. Companies use language to downplay and lie about the severity of their incursions into our private spaces when caught. Our current use of metadata to describe the sale of behavioural surplus, both in privacy communities and in general discourse, plays right into the hands of these companies and bolsters the obfuscation they need in order to continue making huge profits.

Perhaps the renaming of Facebook as Meta, when things got too whistleblowy for the company, is a nasty joke on us—I imagine Zuckerberg saying 'I told you exactly what we were doing. It was right there in the company name.'

I don't know if behavioural surplus is the most effective term to describe the wake of extra data that can be captured whenever we write an email, do a search, or even exercise and sleep, but it is better than metadata, because behavioural surplus comes closer to describing what it is: data about how we behave that can be recorded and sold.

Maybe a better term or phrase or acronym will emerge that will make it immediately clear that we are talking about unintentionally produced data about our behaviour that is being taken for profit without our consent, but for now, I would suggest we got with behavioural surplus as a more accurate term.

Update

Someone posted a link to my post on the Privacy Guides forum. An interesting discussion followed there. I've just joined the forum myself and it looks like an active and knowledgeable community!