The launch of google+ has occasioned a plethora of posts discussing the circles. This discussion has raised interesting and important questions about the design of present and future social networks. Having read people’s theories and projections about the future of circles and any such social grouping as a basic capability of SNs I have come to the conclusion that the arguments presented are premature and the truth is we simply don’t know how people will respond. What I have not seen discussed however, which we do know to be fact is that there are many counties in the world with very different ideas about privacy, and hence perhaps the necessity for groups, than we do and a failure to recognize this can easily result in conclusions that will look ridiculous outside the culture that spawned them.
To set the scene it is helpful to mention (again) the influential work of Paul Adams. Adams discuss the relationship between offline group behavior and online models of such behavior in talks such as ‘The Real Life Social Network‘ - http://www.slideshare.net/padday/the-real-life-social-network-v2 His key point is that social networks such as Facebook that lump everyone together as ‘friends’ treat everyone in a homogenous way that is quite unlike the way people behave in the real world. Offline relationships, he argues, are much more clustered and nuanced. Adams does not make the point but it is worth noting in passing that such complexity and nuance is universal across human cultures.
So the key argument for circles goes like this: if Facebook friends doesn’t reflect offline reality and Google circles does then circles is a step forward, right? It may or may not by itself result in market share shift to google, but it is a better model of offline behavior, providing a better user experience and hence is a better product.
Early reviews have been largely positive, notably with respect to its slick UI, but there are skeptics. And some of their criticisms are profound - they may or may not be correct, but they go to the heart of the value of sophisticated modeling of offline behavior and hence the value of circles.
The criticisms I have encountered fall into three schools (2 & 3 are related):
- sharing isn’t necessary because people don’t actually want it
- sharing is just too much trouble to be worth the marginal benefit
- sharing can’t work because offline relationships are too subtle and dynamic to model usefully and maintain easily
All of these serve to potentially undercut online social groups and must be taken seriously.
- Not necessary
A good example of this was Yishan Wong’s interesting and provocative post on Quora titled ‘How Google+ Shows That Google Still Doesn’t Understand Social” in which he states that Facebook has observed that unlike power user “nerds” who are prepared to sort friends “into buckets” Facebook’s “casual users” are motivated by “histrionics, narcissism, exhibitionism and voyeurism” and such users indicate “increasing desire for more public information sharing.” For Wang the reason Facebook does not put privacy oriented group capability “front and center’ is because of Facebook’s experience with real users. Their testing convinced them that most people don’t actually want groups. Regular people are quite happy to publish publicly and use offline social behaviors such as cliques to filter the resultant noise. Wong further develops his position with what constitutes a theory of individual behavior that he believes explains his position on sharing and groups. http://www.quora.com/Yishan-Wong/If-we-accept-as-I-feel-we-must-your-statement-that-the-clear-lesson-from-facebooks-experience-is-that-most-people-prefer-most-SN-data-to-be-public-and-that-social-behavior-such-as-discretion-can-filter-such-data-do-you-believe-groups-to-be-fundamentally-doomed-Do-you-believe-there-is-practically-no
Needless to say, if Wong is right, circles is indeed unnecessary to anyone other than a power user nerd. From this perspective, any positive early reactions to circles can easily be interpreted as early adopter such nerds finding a SN that meets their needs but don’t expect it to help G+ cross the chasm.
- Too much trouble
Kevin Cheng’s post ‘Can we ever Organize our Digital Friends’ is an excellent example of this position. http://kevnull.com/2011/07/can-we-ever-digitally-organize-our-friends.html Scrupulously itemizing the alleged advantages of circles ‘Why we need Groups’ Cheng nonetheless has major reservations ‘Why Grouping Sucks.” He quickly came to the conclusion that “..it wasn’t worth the effort to rigorously group everyone.” This idea that it’s just too much trouble seems in turn to have two variants (i) people are just too idle to learn anything new, and (ii) a form of (c) it’s just too hard to categorize people and maintain such a schema. The only way we find out if people are too idle to learn anything would seem to be to offer them something hopefully great and discover if it is possible to find ‘the spot’ where it is worth their while or whether all such attempts founder on user indifference. It seems to me far too soon to consider this process exhausted. With respect to the argument that it’s just too hard to do properly, this is properly considered under c) below.
- Offline too subtle
Cheng alleges that whilst we are incredibly good at navigating the nuances of our social groupings in the real world, the real world is just too complex and constantly shifting for any grouping to either capture it usefully and in any event, attempts to maintain it will be unduly arduous. Such complexity may be such that we can never usefully model it and “what I really wonder is whether we should be trying to mirror real life interactions at all. Instead of mapping, wouldn’t it be more interesting to change or create new behavior.’ What that new behavior may be is left as an exercise for the reader.
So what are we to make of these arguments? Do they spell the death knell not just for circles but for any attempt to group our offline relationships? Below we examine each argument in turn.
It may be that we will discover that regular people just don’t want privacy, that they would much rather publish what may be intimate details of their lives indiscriminately and that traditional notions of privacy are eroding and fit only for nerds and activists, but even if Wong is right his theory of individual behavior that supposedly underlies this readiness to expose all doesn’t seem to hold water. He states that ‘Social network usage is driven by two forces - people’s own desire to share and people’s desire for information about others.” This suggests the default behavior with respect to one’s own behavior is that people want to share; that this is the fundamental propensity; and that it is this drive that may be moderated by various considerations. But the rest of the post doesn’t proceed from this assumption. it doesn’t discuss the ‘desire to share’ at all. In fact the ‘product formulation’ discusses its opposite - a desire for “strong privacy for themselves.” which is not sharing, which in turn is in stark contrast to his statement that Facebook found people love sharing.
I see two further areas worth questioning about the argument as presented. Firstly, this is an area where the user experience matters. If it is cumbersome to manage groups it may well be that people can’t be bothered, but what if circles is frictionless enough? Or - as we shall discuss below - algorithmic solutions automate the problem away? Will users still not care? IMHO we just don’t know yet. Secondly, in a world in which Facebook may be ubiquitous but cultures vary enormous, does it not seem a remarkable result that ‘users’ don’t care about privacy? that ‘users’ are narcissistic and happy to publish their lives to the world. IMHO this has an implicit assumption about ‘users’ that is probably quite mistaken. Perhaps cultures are converging and future generations wordlwide won’t care what they publish but that seems a long way away. Facebook is worldwide so they may well have data on such cultural differences but Wong does not discuss them. And Facebook’s group UI is clunky so is it surprising people didn’t like it. In the light of these concerns I feel it is far too early to take Wong’s conclusions as definitive.
With respect to grouping being too much trouble it is worth considering the work of the startup Katango that addresses exactly this problem. (Katango’s blog is essential reading http://blog.katango.com/page/2/) Cheng mentions them in his piece and pronounces “I was impressed with how accurately they created logical groups that hadn’t occurred to me.” Katango ’reached out to’ 9,000 people to arrive at their own assessment of the real situation with respect to sharing and came to two conclusions:
- Social sharing is fun and useful, but we don’t want to share *everything* with *everybody* all the time.
- Sharing with just the right people usually means putting together email lists or messaging groups to talk with; and that just takes too much time and effort.
In other words they don’t buy the ‘let’s all just publish to the world’ argument but they do buy the ‘it’s just too much trouble right now’ position. Arguments about the subtlety of our offline worlds are important because if those who claim that such subtlety is forever beyond computer assisted grouping then Katango will fail and even a killer UI may to too much trouble for a subset of users. How big is that subset? We don’t know because we have not yet run that experiment. But Katango has an impressive team so let’s assume they make huge progress. The question now arises - stipulating a killer UI and effective clustering will users still find groups too much trouble? The truth is, we don’t yet know, and this points to the broader point that we don’t yet understand what tradeoffs most users will be happy to make.
It seems to me entirely healthy that Adam’s assumption that the online should map the offline be questioned. But if we try to sum up the above arguments it is clear to me that conclusions are being drawn extremely prematurely. We have never seen an SN which makes and will doubtless continue to make a serious attempt to support offline groups. We know that google is as good as anyone at algorithmic solutions. If Katango can derive impressive results it is surely fair to assume google will. We are potentially changing social behaviors and we don’t know what may prove to be a new stable equilibrium.There is a huge amount to learn here. Furthermore, it is luminously clear to me that arguments that assume that attitudes and behaviors of what clearly represents a subset of the world’s potential SN users is seriously broken.