What We Talk About When We Talk About Data

The diverse metaphors that we use to describe our online data can reveal a lot about our beliefs and assumptions — and even point the way toward better products.

Our online activities leave behind traces of data that we rarely see. Instead, we rely on metaphors to make sense of our data. Like footprints along a forest path, we can follow our data metaphors to understand people’s attitudes about digital information, to uncover new product insights, and to tackle broader debates about the role of data in society.

Metaphors are familiar ideas we use to explain less familiar ones. “The essence of metaphor,” George Lakoff and Mark Johnson write, “is understanding and experiencing one kind of thing in terms of another.” [1] People surf the internet. We store information in the cloud. But metaphors do more than reveal the meaning we lend to our experiences — they help shape that meaning. For Lakoff, metaphors aren’t just figures of speech. They are figures of thought. [2]

During a recent study for Meta, I talked with American participants about what “data” meant to them. Their responses opened a rich reserve of insights. Although we might not normally pay attention to people’s metaphors — letting these turns of phrase wash over us — I didn’t take them for granted. By consciously interrogating people’s data metaphors, I came to see how our words can reveal the attitudes, behaviors, and mental models that structure our relationship with technology.

Tech literacy and literary tech

To build products for everyone, it’s important to gather opinions from people with diverse degrees of tech literacy. Our interviews should include representative sample populations of participants who are familiar and unfamiliar with the workings of digital interfaces.

Now, consider “tech literacy” literally. If metaphors are a means by which we read the world — integral to the narratives we tell ourselves and others — then representative study samples should also include people with diverse literary techniques: that is, diverse metaphors for technology.

For a Meta study, I asked participants “What comes to your mind when I say ‘data’?” My goal was to better understand the significance that people attributed to data in their lives in order to uncover new opportunities to build privacy controls for data sharing.

For some participants, spreadsheets and numbers came to mind. Others mentioned pictures, songs, or documents. The wide range of metaphors they used to make sense of data also seemed to frame their expectations and prime their attitudes.

Their answers fell into five general categories.

1. Organic metaphors

Data was sometimes likened to “dog hair” — shed wherever you go, perhaps annoying, but ultimately meaningless. Participants who used organic metaphors were not very concerned about the state of their own data. They were resigned to leaving strands scattered across the internet, and saw little value in data control products.

Other participants, however, considered data to be more like the dog itself. For them, control over data became a significant concern. As one participant remarked, “Just as you take care of a pet, you’ve got to take care of your data.” This side of the metaphor represented something essential to be cherished. As a result, privacy and control over data were vital concerns.

2. Storage metaphors

Data store deposits of our digital selves. The metaphor conveys two related but opposed views. On the one hand, a participant told me that data are “lists of things for us to sort.“ She emphasized the container. On the other hand, a participant said data are “any and everything that could be stored.“ He emphasized the contents. Both participants invoked the storage metaphor, as I soon realized, because they cared most about the safety of their data. What mattered for the container and the contents alike was that they remain safe from leaks.

3. Knowledge metaphors

Data confer truth. As one participant put it, data are “devoid of opinions; data aren’t just rumors.” In this view — the least metaphorical one, since the Latin datum means “given” — data represent the foundational knowledge that can guide both corporate strategy and personal decision making. I couldn’t infer any privacy concerns from this type of metaphor.

4. Collection metaphors

Whether data were collected by means of cookies, search engines, or pixels came to participants’ minds. After all, digital products actively cull data from our online behavior. In emphasizing the mechanisms by which data are collected, some participants saw privacy and control as shields against surveillance. A few mentioned the myth that social media use phones’ microphones to record conversations. Everyone wanted to be protected from sites and apps that take their data without consent.

5. Visual metaphors

Data are like shadows: projections of ourselves cast by the digital environment. One participant who brought this visual metaphor to my attention noted that we don’t process data by ourselves. Just as the shadow depends on the sun, it’s the digital platform that generates the data we consider to be “ours.” Another participant invoked the metaphor with chagrin, “Anytime I go online, my data go with with me; there’s nothing I can do about it.” For participants who feel like their data follow them, it is imperative to build control products that enable them to co-create their data. And when people use data controls, social media can become not just a data protector, but a data partner.

All companies should work to partner with people in their data management. That means concrete solutions exist to empower people. And the numerous categories of data metaphors offer a resource for designing those solutions. Far from evacuating any objective sense of meaning, people’s metaphors can serve as a guide to integrate options into data control and privacy products — options such as the categories of data to collect, their retention limits, and how data are used to personalize output in recommender systems.

How to mine metaphors

When it comes to user research studies, pay close attention to participants’ metaphors. The words they use are hardly innocent. Like gestures, pauses, or surprises, participants’ metaphors can also offer a treasure trove of insights. Here are some tips to get the most out of them during studies.

• Free associate. Explicitly ask what comes to mind when participants hear key words. For example, I posed the question, “What does the word ‘data’ make you think of?” Make sure to include similar questions up front in your discussion guide.
• Probe. Ask participants why they chose their words. Following a woman’s offhand remark, “my data swap hands; everyone leaves their fingerprints,” I asked where she heard the metaphor. It turns out that she watched Law & Order. Like crime scene evidence, data can can make their appearance anywhere.
• Trace inferences. What sentiments follow participants’ metaphors? Do they feel in control of their data? Do they care about data at all? Such inferences frame people’s broader attitudes toward technology.
• Scan for recurrence. Do participants repeat rhetorical devices? Do their metaphors shift? One of my participants likened health reports and social security numbers to the trunk of an information tree. His Google searches, however, were akin to branches. Metaphors often form a network — in this case, of arboreal figures — which structure people’s thought.
• Code. After studies wrap up, classify participants’ metaphors according to their underlying themes. Do patterns emerge? And what feelings ensue? Is it empowerment or resignation, understanding or confusion, agency or fear?

The social frontiers of data metaphors

Data metaphors have profound implications. Should we treat Uber as a taxi company or a software developer? Should Meta be considered an algorithmically-driven platform or a publishing house with liabilities for its content? Such metaphors are a battleground on which laws are drafted and regulations are enforced. They also have tangible effects on peoples’ lives. When data leaks were characterized as a “privacy Chernobyl,“ the metaphor conveyed the severity of people’s concern not only for their data security, but also for the corresponding accountability demanded of business and government “meltdowns.” Indeed, our relationships with data-dependent organizations is at stake.

Pivotal to debates over the role of data in society is whether we think of data collection as an act of extraction or sharing. Evgeny Morozov offers the concept of “data extractivism” to describe the covert means by which social media allegedly steal data from users. [3] The figure of thought treats data not as meaningless breadcrumbs or exhaust left behind online but as essential traits of ourselves. For critics of data extraction, the term conjures allusions to the HeLa cell line, which was robbed from Henrietta Lacks while she was a cancer patient at Johns Hopkins Hospital in 1951. [4] Her cells enabled discoveries in genomic and vaccine research. Yet, they were extracted without consent or compensation.

Whereas extraction connotes violation, sharing connotes volition. Users share data with digital products as part of an exchange, something key to data partnership. Aided by privacy and data controls, people can selectively give their data to responsible tech companies and receive valuable tools in the form of personalized services, effective search engines, and social messaging.

At the heart of this metaphorical duel are deeper questions about the nature of our digital identity. When sharing data, we offer off a superficial byproduct of our social interactions. Much like the dog sheds fur, our shared data represent digital dander. And organizations provide value when they act as skillful artisans, weaving our remains into a tapestry of useful products.

Data could, however, be construed to constitute much deeper attributes of our digital identity. For Shoshana Zuboff, data lay bare the very essence of our selfhood. In her view, tech companies extract data from people by translating their online activities into behaviors whose signal is detected. As she puts it, our “personal experiences are scraped and packaged as the means to others’ ends.” [5] The signal is used not just to improve companies’ products. Companies also generate revenue by selling personalization and prediction products to advertisers. Data beget more data as a marketplace of personalization and prediction products expands. For Zuboff, data are an inalienable aspect of our lives, so any form of collection amounts to extraction.

Metaphors for the future

Do data represent inherent or extrinsic aspects of our digital selves? It’s not my aim here to resolve the debate but instead to make clear that metaphors needn’t reduce consequential questions about data’s role in society to subjective contests of opinion. Concrete product solutions exist to afford people more agency over their data. Metaphors serve as our guide to design those solutions, to define use cases, and to evaluate outcomes.

That starts with new metaphors. Social media can be a steward of people’s data. Companies should take care of their passengers, as it were, by investing in transparency and privacy controls. These impart the technological literacy to understand data categories and empower people with the means to determine precisely which data are gathered. Therein lies the key to data partnership. It is a future in which social media companies design intuitive and useful controls for us to mold the footprints that our data leave along our digital pathways.

Author: Larry S. McGrath, UX Researcher at Meta

Contributor: Blaze Owens, UX Researcher at Meta

Illustrator: Drew Bardana

Citations

[1] George Lakoff and Johnson, M. (1980). Metaphors we live by. Chicago: University of Chicago Press. p. 5.
[2] George Lakoff (1986). A Figure of Thought, Metaphor and Symbolic Activity, 1:3, 215–225
[3] Evgeny Morozov (2022). Freedom as a Service: The New Digital Feudalism and the Future of the City. Farrar, Straus and Giroux.
[4] See Rebecca Skloot (2010). The Immortal Life of Henrietta Lacks. New York: Crown.
[5] Shoshana Zuboff (2019). The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. New York: Public Affairs. p. 10.