Sunday 26 May 2013

What is the Half Life of a Social Network? The Problems of Digital Baggage and a ‘Move Fast and Break Things’ Philosophy

It is a little over a year since Facebook went public in a c.$90bn IPO, and I find myself amongst a growing group of people questioning what the future holds for ‘The Social Network’. Social media is a new industry and the rules of competition are being discovered as we go along. And the same things that caused Facebook’s meteoric rise may end up making it irrelevant.

Online technology companies (by which I mean social media as well as services such as search, email, blogging, etc…) face the unenviable situation of being violently pulled in two directions at once. On the one hand is the imperative for constant change: that constant struggle to capture the attention of the general public while hundreds of competitors are trying to do a similar thing. In this light we see firms as locked in an arms race, repeatedly adding features and revising designs in a bid to stay fresh. On the other hand, there is the need to provide consistency and reliability in the service being provided. Users become accustomed to the current set of features, so any changes will tend to cause discomfort and lead to inevitable outcries.

In a way this tension is faced, in some form, by all businesses. But in the online tech industry it is particularly acute, because there are such low barriers to entry, and thus there are so many entrants trying to displace whoever is leading at any point in time. There is also an implicit premium on mere ‘newness’ as trend-setting users like to explore new services in order to be ‘at the forefront.’

In most industries, the most common problem is resistance to change, an inability to adapt to changing user needs and changing social pressures. It’s my strong feeling that in social media, the opposite problem holds. Companies like Google and Facebook are pioneers at experimentation, trialing possible changes with randomized subsets of users, and implementing the ones that appear to work on a wider scale. This is epitomized in Mark Zuckerberg’s famous management philosophy: “Move Fast and Break Things.

The problem is that while this philosophy may be suitable for a small or mid-sized start-up, it suddenly becomes a lot more dangerous for a large corporation. With every change that is pushed through, Facebook risks alienating some fraction of its users, and of those, some fraction is likely to disengage (see my comments on community pages in 2010). Take, for example, Facebook’s algorithm for deciding what gets put on your Newsfeed. This is one of its most important pieces of technology, as it determines both how users interact with each other, and how advertisers interact with users. When this algorithm is changed, it will always have the effect of giving some posts more attention and other posts less attention – and the people or businesses getting less attention will be irritated. I noticed this last year, when Facebook changed the algorithm at the same time as it introduced paid-for “promote this post” options, earning the moniker 'The Biggest ‘Bait N’ Switch’ in History'. I’ve also noticed another change in the last few months, which has led to my newsfeed being filled with less relevant posts (and more advertising than before) and my own posts getting less attention than they used to. I am already changing my Facebook engagement habits as a result, and using Twitter more as a source of interesting links.

This excess of change is far from the only risk. Two others are worth highlighting. An article in the Financial Times last week focused on investor disquiet over FB losing its ‘coolness’ as a result of people’s parents joining. Young people are the key audience for social network, and tend to direct posts to their friends, and the growing presence of older relatives is reducing the perceived freedom of the space. I’ve seen an example of this myself, when one of my friends posted about having a ‘too many drinks’ – and minutes later their father making a rather embarrassing comment on the post, along the lines of ‘I thought you were more sensible than that.’ This is all just rather off-putting, given that the original appeal of Facebook in its early years was based around it being restricted to peers at the same university. Private social networks, such as Microsoft’s Yammer, could begin to capture the attention of users looking for a more exclusive forum.

Then finally there is the issue of digital baggage. This is something I’ve realized recently: much of my activity on Facebook in the early years after I joined in 2005 has left a very large, quite personal digital trail. Furthermore, since I created the content, joined groups etc., the architecture of FB has changed, as has its privacy policies. This makes it rather difficult to see who has access to what. And even if I correctly manage my privacy settings today, there is no guarantee that privacy policies won’t be changed in the future. To give concrete examples, I discovered several photo albums which I thought were ‘friends only’ were actually reachable by the general public with a simple google search. Also there was a set of ‘groups’ that I set up for University colleagues (which are now years out-of-date) but were visible to the general public. These examples struck me as disconcerting, and I realize now that the only way to ‘leave that baggage behind’ would be a make a fresh start with a new social network – for example Google+. I’m not going to do that just yet, but I expect other people will. The Facebook timeline was famously glorified in this video as a way of having a record of your entire life in one place. But that idea is actually quite scary. I, personally, do not want that record in the (potentially) public domain, and I doubt that I’m the only one.

During its initial rise, the trend of Facebook’s subscriber figures was exponential growth. To look at the graph, if it referred to a share price it would definitely look like a bubble. Of course it is not a share price, and so I don't expect it to ‘pop.’ But I think user engagement with FB will decay. However much effort it puts in to try and stay relevant, it is nevertheless likely to be displaced by newer start-ups offering cool new services. The proportion of cognitive real-estate it can command will decline as other services manage to win more. My message to Facebook investors is therefore, get out while you can.

Agree? Disagree? Please leave comments below

Monday 6 May 2013

Schelling, Focal Points and the Problem with Partial Vegetarianism

Following my previous post on Thomas Schelling’s contributions to Game Theory, I have been reflecting on another key idea from his 1960 book ‘The Strategy of Conflict’. Schelling introduced the concept of ‘focal points’ in situations where two parties have to coordinate. The idea is that when faced with several options, people often have a sense of which is the ‘most obvious,’ and when coordination is required they are likely to pick this obvious solution, in the knowledge that the other party is also likely to pick it.

 The classic (abstract) example of such a coordination problem is when and where you would go to meet somebody in New York City, if you had not agreed this ahead of time and communication was, for some reason, cut off. What do you think? When and where would you show up? (see below for the solution*)

The idea of the focal point is important for several reasons, of which I will highlight two.

Firstly, focal points do not fit well with standard ideas in economics. Most modern microeconomic theory deals with curvilinear relationships between variables: equilibrium is defined by the utility-maximizing point on a curve. We find this point using calculus and marginal analysis, and so we assume the curve is continuous. The focal point concept throws continuity out of the window. Even when we introduce game theory – a critical implement in the modern economist’s toolkit – this only tells us that Nash Equilibrium outcomes are stable, not which Nash Equilibrium is most likely. Somehow focal points allow people to efficiently choose between different equilibrium positions, even without communicating.

This leads onto the second, deeper, point. Focal points challenge the economist’s mechanistic idea of ‘rationality.’ It is very hard to specify what constitutes an ‘obvious’ choice. Often the cues that lead to particular choices are idiosyncratic to that particular context. Given a number of different equilibrium positions, standard models of rational behavior might suggest picking randomly between them, or undertaking exploratory search to see what the other players pick. It is difficult, if not impossible, to identify focal points using a mechanistic or analytical solution, the best we can do is identify heuristics for finding them. Yet gravitating towards focal points tends to lead to better outcomes for individuals than a strictly ‘rational’ behavior, and so behavior is in some sense ‘hyper-’ or ‘super-’ rational.

The idea of focal points has been broadened from Schelling’s original coordination-game context, to the general tendency for people to fixate on certain ‘obvious’ equilibrium points of behavior. The tendency to do this seems to be deeply rooted in human psychology. This has major consequences for self-control and our inability to ‘do things by halves.’ It is much easier to give up a certain behavior outright than to stop doing it half the time – for example, I am fully aware that meat production is environmentally very damaging. Vegetarianism would therefore be an ethical behavior for me to adopt – but I love eating meat so much that I can’t bear the thought of giving it up entirely. Now, I could halve the environmental damage my meat consumption does by halving my consumption. However, as I have found (upon trying to limit my meat consumption), it simply does not work like that. A plan to only eat meat on Fridays falls apart when there is a delicious steak on the cafeteria menu on Wednesday!


*This example, conceived by Schelling, has been used many times by researchers as a quiz question. Overwhelmingly people opt for 12 noon as the most ‘obvious’ time to meet. The place chosen varies between two obvious locations: New York residents tend to pick Grand Central Station, whereas out-of-towners choose the top of the Empire State Building.