In Defense of Friction

1903 telephone operator (John McNab on Flickr)

There is no doubt that technology has made my life much easier. I rarely share the romantic view that things were better when human beings used to do the boring tasks that machines now do. For example, I do not think there is much to gain by bringing back the old telephone operators. However, there are reasons to believe social computing systems should not automate social interactions.

In his paper about online trust, Coye Cheshire points out how automated trust systems undermine trust itself by incentivizing cooperation because of the fear of punishment rather than actual trust among people. Cheshire argues that:

strong forms of online security and assurance can supplant, rather than enhance, trust.

Leading to what he calls the trust paradox:

assurance structures designed to make interpersonal trust possible in uncertain environments undermine the need for trust in the first place

My collaborators and I found something similar when trying to automate credit-giving in the context of a creative online community. We found that automatic attribution given by a computer system, does not replace the manual credit given by another human being. Attribution, turns out, is a useful piece of information given by a system, while credit given by a person, is a signal of appreciation, one that is expected and that cannot be automated.

Slippery when icy - problems with frictionless spaces (ntr23 on Flickr)

Similarly, others have noted how Facebook’s birthday reminders have “ruined birthdays” by “commoditizing” social interactions and people’s social skills. Furthermore, some have argued that “Facebook is ruining sharing” by making it frictionless.

In many scenarios, automation is quite useful, but with social interactions, removing friction can have a harmful effect on the social bonds established through friction itself. In other cases, as Shauna points out, “social networking sites are good for relationships so tenuous they couldn’t really bear any friction at all.”

I am not sure if sharing has indeed been ruined by Facebook, but perhaps this opens new opportunities for new online services that allow people to have “friction-full” interactions.

What kind of friction would you add to existing online social systems?

23 thoughts on “In Defense of Friction

  1. Totally with you on the importance of what some call friction. I’d call it ambiguity, since ambiguity is simply what remains unresolved in communication and which propels it ever forward.

    I’d go as far as to say that social systems can by technically dysfunctional and still be socially functional. People find their way, again, through communication, to what they need to know. And are more likely to believe what they see others post than what they read on any About page.

    Not sure about trust systems undermining the need for trust in the same place. I prefer to distinguish trust in people and social situations from confidence in systems. There are systems in which we invest an awful lot of confidence. Do I need to personally evaluate each and every doctor I’m considering — or are accreditation systems good enough?

    There’s a distinction I think worth making between the symbolic system and an action system. We invest confidence in symbolic systems — taking various signs as substitutes for trust (e.g. ratings agencies). So there are ways to capture, represent, and signify trustworthiness.

    Action systems are those patterns of interaction you describe as having been diluted by frictionless mediating systems. Possibly. But the real value to an interaction system is processual — it’s in the actual participation of people in rounds of communication and interaction. Risk and trust are here constantly in the balance — games benefit from raising the stakes; rituals are a means of reducing risks. It’s all very complex, and there’s no replacing it with codified substitute behaviors (eg the efficient birthday greeting).

    The trust that can be sedimented out of a system into symbolic equivalents — ratings, etc — is different from the trust wagered during interaction. I think the two may have been conflated in the study on trust. Anyways, good coverage.

    1. Thanks for the thoughtful reply, Adrian.

      I like your framing around ambiguity and the two terms you introduce.

      I think the study of trust was actually trying to get at the distinction between what you call symbolic and action systems. If I remember correctly, Cheshire was trying to distinguish between trustworthiness and trust, which I think would be the equivalent of what you call symbolic and action systems.

      I think Cheshire would argue that in the example of the accreditation systems for doctors, we rely on the symbolic system itself, not on the doctors. But that after interacting with a particular doctor for some time we reach “action-based” trust that would exist even if the symbolic system disappears.


      1. Bingo. That’s the distinction. This kind of confusion is hard to avoid. It’s a product of our tendency to use predicates — to thingify our language. Trust, influence, capital — they’re relational terms but we tend to think incorrectly of them as fixed attributes.

        So yes, to develop your line of thinking further, not only is friction-free a bad idea (and certainly doesn’t lead to trusting relationships), but friction may in fact be a means by which we shake out levels of trust and confidence. If crises involve high degrees of friction, isn’t it precisely through crisis that new levels of trust, or breakdowns in trust, occur? (By analogy, the value of adding noise to improve signal in low noise audio recordings.)

        There’s another problem with friction free mediated interactions: as they commodify interaction, and expression in particular, they lower the value of interactions to their lowest common denominator. I sometimes wonder if this is just an unavoidable consequence of social tool design and use. Efficiency needn’t be the over-arching aim of design.

  2. Guys, I’ve stumbled across this blog during my Web Science research. I am thinking about trust as a sub-component of credibility and how semantic web researchers are endeavouring to enhance trust with machine-based solutions.

    I am intrigued by your ideas regarding friction/ambiguity. In my research, I am struggling with the distinction between what we do online and offline. For example, I may go to a close friend’s birthday party to signify my commitment to our relationship but still write ‘happy birthday’ on her FB wall to broadcast to our shared community that I’m a good friend who’s remembered. I may also write on the wall of someone with whom I have a tenuous relationship that I can only be bothered to maintain with lazy gestures. It’s hard to tell what our relationship is just by analysing our activities online. Similarly, people trust brands online that have a legacy of positive offline interactions. I found during my research people trust whatever they read on our National Health Services’ website if they have had affirmative treatment experiences offline. Given that trust is already established, users want a friction-free experience with the NHS website.

  3. I think it’s good to have the option for frictionless sharing and “friction-full” sharing and being able to communicate to your audience how much friction you had to overcome to share that piece of information.

    Frictionless sharing is helping to recreate the quotidian experiences friends experience when hanging out or working in the same room as each other. There is importance being able to understand the less substantial, not share worthy, occurrences in a friends life. I believe this information helps strengthen relationships and believe anyone can attest to that if you have ever taken a road trip with someone. Fritctionless sharing has a lot to go to achieve this level but i believe it will try to repicate that intimacy.

    Friction-full sharing is as important and should be able to be clearly distinguished from the frictionless. The two do not need to be exclusive of each other, just given diffent priority. Friction-full sharing will tend to be more about “read this now” vs “this is what I’m browsing”. Hopefully this allows users to explicitly share higher quality, more meaningful content since they are aware that that frictionless sharing is already picking up the rest.

    1. Noah, I like your idea that frictionless sharing is like being in the same room with friends. I wonder how that experience could be replicated on the social web. Perhaps the equivalent is more like a chat room where a) all the people sharing the experience know who is online (i.e., who of your friends is reading) b) The information is ephemeral by default.

  4. Pingback: On Friction and Sharing | iterating toward openness

  5. Pingback: The Facebook Illusion « The Red Eye Portal

  6. Pingback: The Frictionless Life and Its Discontents: You Heard It Here First « The Frailest Thing

    1. It’s… the SYSTEM. 😛

      Good article by the way. Lately I’ve been finding Facebook so unsatisfying. I’m an American living in the Middle East, and it doesn’t stop me from missing my friends. It almost makes it worse. Facebook just doesn’t replace being in the breathing presence of another person.

  7. Pingback: Lažna društvenost i automatizovano šerovanje urušavaju online poverenje | Dragan Varagic

  8. Pingback: Bookmarks for December 13th from 10:46 to 10:46 | LVDV

  9. davidwlocke

    Friction is an allocatable good. Design it. Be friction-free here. Have some friction there. Learning is an exit barrier, that requires some friction.

  10. Pingback: Types of Social Sharing - Mostly Baked

  11. Pingback: Social Friction & Social Media | Spinstah

  12. Pingback: » Why Friction Should Be a Part of Your Strategy PayPerTrends by Sam Harrelson

  13. Pingback: Checking Under the Hood | ServeSurfer Foundation Blog

  14. Pingback: Higher Friction = Higher Reward

  15. Pingback: Compartilhando em redes sociais: com atrito ou sem atrito? | Modelo Mental

  16. Pingback: Compartilhando nas redes sociais: com atrito ou sem atrito? | Modelo Mental

  17. Pingback: The end of social - O'Reilly Radar

Comments are closed.