Influence is a complex process in the development sector. We have known this for some time – the work of the RAPID programme at ODI on understanding how evidence influences policy is very clear on this. But the wider socio-economic system within which development cooperation is embedded is no less difficult to influence. Many corporations, especially in new media, are turning to complexity and evolutionary sciences as a means of measuring influence. But there is considerable potential for misuse and abuse, as illustrated by a new report on Facebook’s contribution to the European economy, and a recent critique of the latest wave of social media analysis firms.
I: Facebook’s Impact on Europe
In a report published in January 2012, Facebook asked the global business advisory firm, Deloitte, to estimate the benefits it generated for the European economy. The report findings make intriguing reading for anyone with an interest in performance, accountability and transparency.
Deloitte’s analysis looked at the direct economic impact of Facebook – such as paying tax, profits and wages. These are the so-called ‘narrow economic effects’ of on-site activities. These impacts are seen as Facebook’s value added, and are described as ‘analogous to contribution to GDP’. Facebook also has a series of what Deloitte calls ‘ecosystem effects’ – namely, how Facebook enables other businesses to ‘reach customers, make sales, create and monetise apps and even boost demand for products such as broadband and smartphones.’ These ‘broad economic effects’ which result from the Facebook ecosystem give us a measure of Facebook’s influence. The diagram below sets the model out in more detail.
Facebook’s ‘narrow effects’ suggest that it has a narrow impact of €214m and supports 3,200 jobs. But its broad effects are considerably more. For Europe as a whole, the economic impact of Facebook was estimated as just over €15bn in revenues, supporting 229,000 jobs.
By far the largest part of this broader impact was seen as ‘the impact on business participation, where Facebook enables other businesses to advertise, promote their brand, raise awareness and therefore generate new sales… much of this effect is associated with the brand value created for organisations through the social links prevalent on Facebook and the new ways of engendering loyalty and interest that Facebook provides.’ [emphasis added]
Facebook Chief Operating Officer Sheryl Sandberg had this to say when launching the report:
Today’s report shows that Facebook is about a lot more than sharing pictures or keeping up with friends. Increasingly, social media means growth and jobs. Social media is proving particularly valuable for small- and medium-sized businesses, which form the backbone of the European economy.”
These are strong statements, and certainly in keeping with the ‘Facebook boom’ narrative that is so prominent in the media at the moment. In a time of economic gloom, Facebook seems to be one of the few glimmers of hope.
II: Spurious reasoning, dodgy numbers
However, Deloitte seems rather more measured than Sandberg. While clearly happy to put their name to and launch the report, the preamble to the report qualifies this in the following passage:
As set out in the contract, the scope of our work has been limited by the time, information and explanations made available to us. The information contained in this report has been obtained from Facebook Inc and third party sources… Deloitte has neither sought to corroborate this information nor to review its overall reasonableness… no responsibility or liability is or will be accepted by or on behalf of Deloitte… or any other person as to the accuracy, completeness or correctness of the information in this document or any oral information made available… [emphases added]
How about that for an object lesson in distancing oneself from ones work? But if we look at the detail of the analysis, we can start to see that the ‘ecosystem valuation’ is based on some very sketchy assumptions.
For example, Deloitte attributes €6.6bn of the €15.3bn European-wide economic impact of Facebook to “brand value”. This is based on the attribution of a fixed cash value of a Facebook fan of a particular product (€4.69), taken together with the total number of fans (4.2 billion), with some downward adjustments. While all of the numbers used are based on other studies, the overall calculation and final figures still seem fantastically overblown.
The report also suggests that Facebook contributes €5.5 billion to the European economy by generating technology sales. €0.4 billion of this is down to additional device sales, and the rest is seen as broadband. In effect, the report is saying that large numbers of Europeans are buying devices and signing up for broadband just to – or mostly to – use Facebook. Again, this is a claim that would be very hard to substantiate. These two figures alone make up €12.1 billion of the stated impact €15.3bn of Facebook.
III: Social media hyperbole
This is a particular form of social media spin, and is part of a wider movement described by Philip Sheldrake in the Guardian last week. Sheldrake argues that a whole spate of social media organisations are using and abusing the tools and ideas of complexity science in order to demonstrate their influence, all with an air of scientific credibility.
The rest of this post draws extensively from Sheldrake’s critique. He begins by describing what influence is:
You have been influenced when you think something you wouldn’t otherwise have thought, or do something you wouldn’t otherwise have done. …ultimately no one wants to communicate without influence; that wouldn’t be a good use of the communicator’s time and energy, or indeed that of those on the receiving end. The focus on making sure you’re influenced back is vital too… Individuals (and organisations) that best absorb the zeitgeist are heuristically more able to respond in ways their audiences (stakeholders) might well appreciate…
But things aren’t all that straightforward, and he turns to complexity science to show why:
Complexity is the phenomena that emerge from a collection of interacting objects. The interacting objects could be molecules of air and the phenomenon the weather. It could be vehicles and the phenomenon the traffic. Human objects could be the population of Cairo, the 99%, sports fans in a sports stadium, people who like photos of cats, your customers, or your employees; in fact, any collection of people interacting with each other, influencing each other. A characteristic of complexity is that studying the individual rarely betrays anything about the phenomena. You can’t learn much about the termite mound by studying the individual termite or the traffic jam by studying the car.
Sheldrake then relates the ideas of complexity science to the phenomenon of influence:
Take almost any of your recent thoughts or actions and try and decipher how in fact that thought or action came to be; what did you take into account, consciously and unconsciously, over what timescale? You soon begin to appreciate that your thoughts and actions are outputs of a complex system. You are reconciling multiple inputs, multiple influences.
The article points out that companies such as Klout, PeerIndex and PeopleBrowsr all claim to provide systematic insights into individual influence, using ideas of complex systems (specifically social network analysis). This is problematic, however:
In my opinion, complexity and network science will continue to unearth insights of important commercial and societal value, but I’m considerably less enamoured about seeming to translate today’s analytical capabilities into some kind of a score of an individual’s influence. Right now, we have no scalable facility to ascertain or infer who or what caused someone to change their mind or behaviour, without falling into some kind of last-click attribution trap, so how then can we pretend to score an individual’s likelihood to exert that influence, and as if they did so with apparent Newtonian simplicity? We’ve barely even attempted to correlate proxies for influence, assuming that universal correlates even exist. Today, these scores are apportioned in such naive fashion that your so-called influence changes following a fortnight offline.
IV: Navigate complexity, don’t ignore it
This seems to be precisely the kind of thinking that can be seen as underpinning the ‘Facebook ecosystem’. On this basis, we might say that the Deloitte analysis was weak not merely because they did not seek ‘to corroborate this information nor to review its overall reasonableness’. It also falls into the trap of attributing benefits to Facebook in far too simplistic and straightforward a manner, through over-use of the metaphor of ‘ecosystem’. To cite Sheldrake again:
Perhaps these companies attempt a measure at online popularity, or perhaps online authority, or more exactly the likelihood to have one’s online output shared/forwarded, but not one’s influence. Nor indeed one’s trustworthiness.
Sheldrake also cites Duncan Watts, noted network expert, who has argued against such applications of network and complexity science:
Influentials don’t govern person-to-person communication. We all do. If society is ready to embrace a trend, almost anyone can start one – and if it isn’t, then almost no one can.
This is not great news for Facebook and other social marketeers: ‘many [of whom] have claimed to be able to identify the influentials, get to know them, and influence them. They are effectively claiming to be the influencer of influencers, a sort of influencer-in-chief if you like.’
Sheldrake closes with a message for marketing and PR consultants that is equally pertinent for development and humanitarian agencies seeking to demonstrate their influence:
However, successful [organisations] of the 21st-century will avoid such simplistic thinking, such hyperbole, and recognise complexity and navigate it appropriately.’ (emphasis added)
Facebook and other firms who are well advanced in their use of complexity science ideas should be paying careful attention to Sheldrakes’ assessment. Those of us in the development sector – despite being at much earlier stages in both our efforts to use complexity science and analyse influence – would perhaps also do well to take note.