Most of us are aware that social media sites, apps or platforms collect a lot of personal data through the things we do with them, but the processes are understandably foggy.
The simple assumption is that our personal data are sold to marketers and advertisers to serve us accurate ads. For the most part that’s right. The kind of fine grained targeting that Facebook achieves is staggering. But there’s a lot more going on. And in response, there is an equally urgent need for data literacy, a better understanding of social media data drawn from content and activity, and how they connect with social life.
Disconnection no longer seems a workable option. We should know what social media data is, how to access and read it, and be able to use it to understand the world around us. Maybe this is too lofty a goal? At the very least, becoming digital citizens will increasingly require breaking down some of the barriers of trust that shadow the idea of our data trace.
Whether we like it or not, metadata are vital for organizing and accessing the petabytes of digital stuff circulating online or resting in server farms. At the turn of the century, Bonnie Nardi and Vicki O’Day explored the idea that a library is an information ecology, not just an organized archive of books and journals. In their schema, an information ecology “is a system of people, practices, values and technologies in a particular local environment”. Libraries are not reducible to the collection or to the architecture, but are a function of the pairing of people with information, by whatever means. And right from the start, metadata and catalogs have been central to that system.
In the rush to define “big data” over the last five years or so, less attention has been paid to data itself, in all its varieties.
Data is actually a nebulous concept. It’s hard to pin down. There is often confusion between metadata — or the usually unseen digital statements about a potentially informative object, as Jeffrey Pomerantz puts it — and use-data, or data exhaust as it’s sometimes known. Data exhaust flows from all the things we do online and because of its volume, velocity, variety, is often linked to big data. The “variety” part is what obfuscates. Almost anything can be datafied.
What concerns most is the idea of a data footprint or map that might be extracted from our everyday use of social media platforms and apps.
Social media blurs the boundaries between the categories and standards established for metadata that help to organise and structure digital material, and data exhaust or use-data. Content itself — the media we produce, the posts, tweets, conversations, exchanges or clicks and shares — are like metadata, bits of operational and analyzable information, but messier.
In the move to new “cultures of connectivity” as José van Dijck puts it, data transparency may be the most significant social battle of our time. But the next step will be literacy. Data literacy requires controlled, but also better access to social media use-data (particularly our own). How do we read and make sense of our own data and the data publics that encompass us?
To edge toward this goal, I’ll detail some fundamentals underpinning social media metrics and analytics. My comparison of social media content and activity data across platforms aims to place that data within its material context, as it cuts across, defines and reflects aspects of social media ecologies, constituted by a dynamic array of platforms, practices and social media publics.
Social media use-data not only underpins our new media ecologies, but data literacy is crucial for intervening and finding and maintaining the public good within the commercial imperatives of the dominant, all consuming social media platforms.
Social Media Analytics: Platformed and Segmented
Social media analytics is popularly understood as the knowledge or insights generated by processing data produced within networked contexts. The current name of Facebook’s native analytics tool, “Insights”, denotes this process of “revelation”, extracted out of continuous and large data flows. Definitions of social media analytics take in these various elements. Analytics are the systematic and continuous computational analysis of data. It’s the breaking apart of a set of data, its calculation, and re-assemblage in a new, more meaningful form.
Metrics, on the other hand, are the individual units of measure, the instances and varieties of data produced and stored ready for calculation and query. They are standardized measures that allow descriptive and comparative insights to be drawn out of messy and complex social activity and interactivity. Where analytics are open concepts and comparators, metrics are more precisely defined, and are tied to site features and affordances.
Social media analysis does not happen in an objective computer bubble. Just as algorithms are human designed, analytics designate and prioritize concepts for making sense of social media content and use-data.
These are also the mechanisms by which platforms segment people, experiences, content and digital objects of all kinds as a factor of specific, storable, comparable metrics. In other words, analytics are knowledge targets, or the kinds of things we might want to know about the social activity producing the flow of data.
Even at the highest level we understand how platforms segment us: Reddit is for mischievous college students or news boffins, Gen-Y are migrating to Snapchat or away from it, Beliebers and political pundits are all over Twitter, designers and fashionistas find a home on Pinterest, Tumblr offers marginalized teens a set of subcultural hideouts, or amateur porn. If you’re trading you’re on Ebay, if you make you’re on Etsy, Facebook has been colonized by moms and dads, and invaded by corporate Pages.
A micro analysis of metrics across platforms reveals both the standardization of social media activity, and some of the platform differentiators. The table below brings together a comparison of some of the key types of metrics and corresponding analytics targets across five social media platforms.
Facebook and Twitter and other platforms, along with developers of third party social media analytics tools, have a particular interest in providing insights on user identity and profiling, visibility or “brand awareness”, engagement or influence. This is because social media analytics are first and foremost techniques of market segmentation, so traditional market insights are adapted to social media use, but also guide the ways platforms like Facebook will organize and measure user activity and hence channel content. (There’s just not enough room to list all the points of profile data Facebook requests of its users — it is a powerful profile engine).
Profile data, for instance, loves a binary or a quantity — male or female, age, single or “in a relationship”. Binary segmentarity encompasses us even if we’d prefer an alternative middle ground. Segmentarity occurs in other ways too: through the places and locales of our everyday activity, our work space, school or home, neighborhood, city, state, nation, as they encircle us and characterize the circular segments of so much of our everyday social media activity. Similarly, our life stages, events, occurrences, and all the temporal factors that make up everyday experience indicate our linear segmentarity.
None of this is to say that we have become pure marketing drones in the great capitalist expansion into Web 2.0 (although many make that argument quite convincingly). It’s rather to ask that we consider the greater ecology of forces, experiences and contexts that implicate social media data. Zizi Papacharissi refers to big data as offering not a new kind of knowledge, but forms of “situated knowledge” — “they tell a rich story, but they are also part of a greater story” (2015).
Well before the data metrics and analytics of social media, we were always, as Gilles Deleuze and Felix Guattari put it, “segmented from all around and in every direction.” They go so far as to say that “the human being is a segmentary animal”. Their understanding of segmentarity that I’m adapting here was taken out of the work of political anthropologists trying to understand the parallel operation of macro and micro social forces on how people act socially.
Even if so many of us resist the granulated fields of Facebook’s personal information requests, we’re a generation schooled on the art of self-presentation, and self-curation. Social media platforms specialize in curatorial segmentation. But it’s a two-way process.
Social codes, categories into which we fit (or don’t, but are placed), places, times and events have always segmented us as social beings. Social media data, metrics and analytics simply work to standardize and extract further value from these segmentary forces. Likewise, social, cultural, commercial, economic, political interests always underpin decisions about what counts as significant social knowledge.
One of Deleuze and Guattari’s contributions to social theory was to understand the micropolitics at play in these processes. And they saw the slippages, breaks, ambiguities, plurality, in the multiple lines of segmentarity: the “lines of flight”. Just think about the play, politics and cultural practices surrounding hashtags these days as an example. Or look at the subtlety and irony at play in the coding of Instagram images with hashtags like #depression or #happiness. Data literacy entails an awareness and account of the messiness and contradictions of lived experience as they become entangled with datafication.
Social Media Ecologies: Platforms, Practices and Publics
Despite the business insights and ad targeting evident in the kinds of analytics that Facebook, Twitter, Instagram offer commercially, we can make sense of social media data in other ways.
Thus, in the academic literature, platform studies and politics (including infrastructures, data, software, and algorithm studies) is well underway (see the Social Media Collective, or Share Lab, or Data Active). Likewise, social media’s user practices have been studied, from the activities and tactics of influencers, the production and circulation of GIFs, emoji, political memes, trolling, or as captured in the ethnographic “why we post” project. And research into the characteristics of networks, the emergence of activist publics, or hashtag publics, a Twittersphere or intimate publics has already yielded significant new knowledge about the “real time” workings and “pulse” of societies (have a look at the Australian Twitter News Index, or CityBeat, or CSIRO We Feel).
These often separate spheres of social media research together constitute an attempt to understand our new media ecologies. But on a more micro scale, getting started with building data literacies involves placing data and the processes of segmentarity within this new media ecology, as an assemblage of each of the elements (not one or another in isolation).
Government agencies seeking to define “digital literacy” or “media literacy” as do the Ofcom in the UK or the ACMA in Australia, need to keep pace with the imperatives of “big data” within the new media ecology. Data literacy now means paying attention to the specificity of the platform or app, considering our and other’s practices carefully, and recognizing the “publicness” of those actions, and the publics they bring into being.
Data literacy is lost if it’s not thinking media ecologically. Social media experience, as “captured” by standardized elements of profile, content, activity and interactivity underpin a dynamic, overlapping ecology consisting of platforms, practices and publics:
The life-world, personal experiences, behaviors, feelings, events, places, profiles (selves) always exceed the bits used to distinguish and measure them. The broad goal of data literacies is first and foremost to make this fact understood. There are any number of ways to slice up a life or social event. The social media content and use-data collected for this purpose, standardized for comparison or time series measurement, must be read with this in mind.
Equally, a data literate person sees the benefit and potential in the data-slicing that social technologies enables. I think this is the tacit acknowledgement in the proliferation of self-tracking for health or “self-optimization”, or the endless flow of wellbeing and #fitspo content that frames the daily struggle to obtain or maintain social capital online.
To know and understand, and intervene in these processes of segmentation — or segmentarity, to use the more expansive concept — we can read social media metrics through the material ecology within which it sits.
Building on, but also working parallel to critical data studies and data activism, new data literacies and critically informed research might begin to map the standard objects of social media metrics onto the socio-technical domains that produce and are produced by them.
In the same way, social media data constitutes the glue in an assemblage of platforms, people, behaviours, social activities and community or network formations. Profiles, activity and interactivity have a material basis in the world. Visibility may be extended as a factor of profile, activity and engagement, but it remains a very material measure tied to the publics we inhabit and imagine.
To understand and intervene in this ecology, data politics will become increasingly important. The first step involves broadening our reflexive knowledge, or literacy, and analysis that runs against the grain of marketing-oriented, segmentary insights. This includes the search for new, different types of knowledge that can be generated from social media data. The ultimate goal would be to foster rights claims and ethical guidelines relevant to life as data citizens and as digital citizens.
Deleuze, Gilles and Guattari, Felix (1987) A Thousand Plateaus, Minneapolis: University of Minnesota Press.
Nardi, Bonnie and O’Day, Vicki (1999) Information Ecologies: Using Technology with Heart, Cambridge, MA: MIT Press.
Papacharissi, Zizi (2015) ‘The Unbearable Lightness of Information and the Impossible Gravitas of Knowledge: Big Data and the Makings of Digital Orality’, Media, Culture & Society, 37(7), 1095–1100.
Pomerantz, Jeffrey (2015) Metadata, Cambridge, MA: MIT Press.