Western democracies are desperately searching for Martin Luther. Following the Facebook and Cambridge Analytica data scandal, the social media platform in particular and tech giants generally have found themselves increasingly under scrutiny for their data practices. Such scrutiny takes many forms: is Facebook a threat to democracy?; are tech giants too big?; who has agency over our data?; and so on.
These are questions that have always existed, not just for social media, but for all private interests and indeed all states. There is little objective nuance in these questions. New technologies have always altered the power of the masses, and thus potentially threatened democracy; monopolies have always existed in one form or another; agency has always been a very transient thing. Even when existential questions about social media have been asked, most obviously in the campaign to delete one’s Facebook, the irony has been missed on those who distributed their rallying cries via hashtags, blogs and twitter feeds.
Luther, rather than David, is the character whom many seek as we consider data in a more critical light. We do not wish to slay social media; we want to reform it.
Social media reactionaries, irony aside, do not actually want to delete Facebook. We enjoy the curated information steams decorated in gradients of blue, the egalitarian means by which we might show approval, and the undeniable efficiency of social media as a means of consuming said information and showing said approval. Even the toxic areas of social media – and the Internet as a whole – act as reflections of social problems that exist without social media. At best, social media allows us to peer into fringe groups with relative safety. At worst social media allows these fringe groups to project their ideas outwards, protected by anonymity, which whilst often detestable or uncomfortable, we should recognise is a truth that might not have formally been identified.
Those that believe retreating from social media is the preferable reaction may be accused of suffering a similar belief that the power of invisibility may be gained by shutting one’s eyes. Just because they no longer see the world change around them does not mean change has stopped.
The question, then, is what is being reformed? To take the above scrutiny, the rhetorical questions that precede such outrage may be translated into two statements:
“Why did Facebook collect THAT piece of data about me!?”
“Facebook did WHAT with my data!?”
These statements represent either a shallower, or a deeper – respectively – understanding of the age of data. It is for those who reject the quantification of the world that exclaim the former, and not without good reason. There are legitimate causes for consternation. Privacy is that of most notable concern, and just as we teach young children not to talk to strangers, it seems good practice to not wantonly share data with faceless companies which unravel one’s privacy and leave oneself exposed.
When we realise that we have been doing such a thing by participating in social media – as many surely have since the Cambridge Analytica scandal – it is a natural reaction to withdraw. It is not necessarily the correct reaction.
By deleting one’s Facebook or Twitter, each of us might claim to be reclaiming some privacy and authority over our data. But we forget that companies which collect our data are companies, and insofar as they offer a service we would like to acquire we must pay a price. If the alternative is to pay a membership to these platforms, we would simply find ourselves out of pocket, as well as digitally exposed. If it is to legislate what data may and may not be acquired by these companies, that imbues a whole miasma of regulatory back and forth, arbitration and conflict – issues which, let’s be honest, the vast majority of social media users simply don’t care about.
Concern for one’s personal data security on something as individually innocuous as Twitter or Facebook is simply arrogance – as a data point, we are all very much unimportant. I will, however, return to this conjecture. The point is, for the vast majority, the price of their personal data is a fair cost for the services that the likes of Facebook and Twitter provide. When our data is used to suggest interesting people to follow on Twitter, to organise social gatherings on Facebook, or to recommend delectable entertainment on Amazon or Netflix, we see the great benefit that the tech giants generally provide.
This brings us onto the second statement. It follows quite logically that if the ‘THAT,’ piece of data is willingly given to produce a better service, when that same piece of data is used for an enterprise that is not a better service – the ‘WHAT,’ function – we will criticise the platform specifically, not the process generally.
This is perhaps the cure for the irony identified above: those that were calling for the deletion of Facebook were calling for the culling of a bad actor, whilst Twitter, Blogspot and others had done nothing wrong, and would have been unfairly targeted if this irony were true and discrediting. But this is not the point.
The point is this reaction to the ‘WHAT,’ function is a more rational response as it accepts the transactional nature of data and social media – that data is given only insofar as it creates and improves desired services. When Facebook as the trusted keeper of this data oversteps their mandate, or is lax in their protective duties, or both, situations like the Cambridge Analytica scandal arise. For policy makers, at least initially, this should be the area of legislative interest.
However, this simple model of social media and agency is incomplete, and not wholly due to simplicity. Whilst for some the collection of data (THAT) is the point of incredulity, for many it is the outcome of having given that data (WHAT). But how does THAT become WHAT?
As stated above, individually our data is not substantial. Most people understand this, hence the emergence of Big Data. But Big Data is often an overvalued asset; social media sites will gleam any and all data they can from us, and they will surely have fantastical ideas of the services they might provide in the future (the THAT and the WHAT, respectively). A specific example is Cambridge Analytica, who siphoned the data of Facebook users with the intention of supporting a particular election outcome.
Perhaps it is unique to the Cambridge Analytica scandal, with its politicisation, or perhaps it’s a condition of our data obsessed lives, but the scandal was not about the THAT and the WHAT. Users had already given Facebook that data, and citizens were already exposed to campaign advertising via the Internet and traditional platforms. The scandal, I believe, revolves around the HOW.
It would be abjectly unfair to place the blame for the scandal at the door of behavioural economics. For the most part, Big Data analysis doesn’t really care why a person with a particular set of data points is more likely to support one candidate/product/idea as opposed to another. Big Data simply identifies the pattern and targets the resources associated so as to match the pattern. Here, behavioural economics becomes the far too fastidious an advisor, the consultant who didn’t get the memo about going fast and breaking stuff.
As such, in lieu of the Cambridge Analytica scandal, pillars of behavioural economics such as nudge theory probably find more of a role as villainous underlings in targeted news campaigns (the presumed resources of a political advertising machine) than they do in deciding who will be targeted in the first place. Indeed, would the Cambridge Analytica scandal be a scandal if people did not believe – rightly or wrongly – that the efforts of the company were relevant to whatever outcome they were trying to facilitate?
Possibly, but it is much less clear than it might be otherwise. If the HOW of the matter is to receive some blame, and be subject to some reformation, behavioural economics may be as worthy of criticism as Big Data is. I find myself coming to this conclusion on the Cambridge Analytica story: Big Data is the new corporate sexy; cognitive deficiencies make us feel dumb. The latter should not be forgotten because the former is a more palatable creature.
This point, however, should not be conflated with the Cambridge Analytica story. That circumstance is more unique; the tandem utilisation of Big Data and behavioural economics need not be.
In an election, everyone is a potential consumer. Whilst Big Data might be used to target those more susceptible to a particular side’s advertising, for those who are accidently targeted the message is not wholly lost – the message is just less efficiently received. But for a commercial advertising campaign, where the promises of bang for your buck may make or break a would-be Cambridge Analytica’s business model, behavioural economics becomes more relevant.
Here, there is no need to go fast, and certainly no desire to break stuff. On the contrary; where the demands of Big Data are the maximum return from the number of advertisements placed, behavioural economics comes to the aid. The manipulation or exploitation (both controversial words) of ubiquitous cognitive shortcomings may reduce the cost of accidental mis-targeting, whilst the framing of products as defaults, offers as loss averting and the use of multiple ads to invoke herding effects may make advertisements too effective for those already considered susceptible.
In other words, Big Data may tell us who to talk to, but behavioural economics may tell us what to say. In this sense I return to the title of this piece; behavioural economics may be the force that pulls the lever of Big Data. This is a logical realisation I believe many in the field of behavioural economics will come to. A nudge such as a default option effect is far from perfect; for benefits of this nudge to be seen, a great many observations are often needed. As such, the domains in which Big Data and behavioural economics rely are the same: across populations.
Whilst the discussion of this piece has been framed around a scandal, it is not my intention to suggest either Big Data or behavioural economics are malignant. As with the discussion of the benefits of social media, such an accusation would be far too simple, and far too easy. But a great deal of the commentary on the nature of data and its place in society misses the point: it is HOW we use data that matters. If this discussion is to be had, I feel behavioural economics must be included.