The Fifty-Foot Rule Reconsidered

I have heard cited several times the so-called fifty foot law of sociology, which says that most collaborations happen among people who work less than fifty feet apart (the idea is generally credited to Tom Allen of MIT; the primary reference seems to be his monograph, Managing the Flow of Technology, MIT Press, 1977, which I admit I haven’t actually read). Let’s generalize and assert that most relationship interactions happen among people who live/work 50 feet apart, plus or minus an order of magnitude, say 5-500 feet. This being a probabilistic, phenomenological law, it should be interesting to mull how it is changing in Tom Friedman’s flattening world, and to what extent lives are getting transformed in terms of changes to this law.

(Note: be sure to check out the follow-up article, “The Twitter Zone and Virtual Geography”)

For a good part of 2004, as a postdoc at Cornell, I was among the most egregious violators of this law. My postdoctoral adviser was off on a sabbatical in Europe, and we interacted over the phone and email. My PhD adviser, with whom I was still collaborating, was in Michigan, as was my girlfriend (now wife). My parents were on the other side of the planet in India, and most of my friends were scattered around the globe. I had no social energy and made no friends in Ithaca for months. My phone calls from that period had practically no local calls, and the average distance of my calls was probably thousands of miles.

It didn’t last too long — my girlfriend joined me in Ithaca, I began working with a few students who I met with face to face, and my supervisor returned to Cornell. Despite my utter social laziness and deliberate self-isolation in a remote apartment complex outside Ithaca (one that Thoreau would have liked), I managed (for the first and probably last time in Apartment Land, America) to befriend not one, but two of my neighbors.

But no, that wasn’t to last either — neither a traditional localized pattern of relationships, nor a comically distended one was to be stable. My supervisor went off again, students left for the summer or graduated, and my girlfriend went off to Nepal.

Then she returned, we got married, lived together for a year and now we are back at being 500 miles apart, with jobs in Washington, DC and Rochester, NY, respectively, piling on the frequent flier miles. At my job, in Xerox research, I started out working with a team whose other members were down the hall, but with a couple more projects underway now, I am back to messy global collaborations, one of them involving team members in Europe, California and Canada. 11 AM EST is now the only possible time I can schedule full team teleconferences. Now that blogging is a major part of my life, and since I am reconnecting with old friends (at the expense of potential new ones I should be out drinking with), my active social network is again all over the place. Again, I find myself being lazy about developing new local friendships, outside of lunches with co-workers (“what’s the point” is a pretty valid complaint, since my Ithaca friends too have joined my Jamshedpur, Bombay, Austin and Ann Arbor friends in being all over the damn place).

So a time graph of my key relationship distances, both personal and professional (say the top 20 people I interacted with), over the last few years, looks something like this:


Collaboration Distances

Will it ever settle? I am not sure. The mean may drift down, but the variance is likely to remain high. Now keep in mind that I am a pretty ordinary guy (among educated information workers that is) when it comes to such things. Though I experimented with pen friendships, ham radio and backpacking when I was younger, none of those things ever captured my fancy for long. I realized about 8 years ago that I am a homebody, and so I am by no means an eager adopter of the globe-trotting, international-numbers-on-speed-dial lifestyle. In a previous era, I would not have sought out a career in foreign affairs or sailing. I’d have settled somewhere boring and lived out my life within a 100 mile radius. I don’t have much wanderlust or global-connection-itis left in me. All this happened naturally and largely against my will. I am sure the more enthusiastic globalists among you have much wilder graphs than mine.

Another way to slice such an ordinary-for-2007 lifestyle is to break down every relationship into its interaction units, conversations (including, say, face to face, phone, email and IM), and look at the frequency distribution of the distances of interactions. Here is a sketch of what I think might be happening:

Frequency distribution

I suspect that while industrial age technologies (postal mail, telegraph, telephone, fax, email) progressively moved the mean of the interaction-distance distribution rightwards, the globalized era is actually fattening the long tail. The bulk of the distribution, for most of us, may remain under a hundred miles, but the tail will start to increasingly thicken. More so for work than for community, but increasingly for community life as well. My top 20 friends are not within 20 miles of me, and will likely never be, no matter where I move, even if it is to the most packed metropolitan or super-suburban sprawl.

The Implications

While a perennial idealist-optimist like Friedman (is he on drugs?) might see nothing to mourn here, I have enough of a tragic streak in me that I see a definite and irreversible loss of quality of life and community here, most famously noted by Robert Putnam in his classic, Bowling Alone. No, I am not pining for some romanticized small town Norman Rockwell painting life (or Naya Daur village life, to pick an Indian cultural reference). I am talking about the harsh reality of forms and structures of traditional social capital slowly disintegrating, without a clear replacement emerging from Google Labs or Linden Labs. Much as I enjoy spending my evenings in Starbucks reading, arguing with my wife on the phone and keeping in touch with good friends over email, often I wish they were around me rather than, on average, 500 miles away. My local cafe baristas change so often, and the clientèle is so variable, that they are hardly hubs for a renewal of community life.

You might believe that virtualization will soon break this stalemate, that we’ll be able to live where we like and keep at least our families in one place. You might imagine that communities of like-minded teleworkers will form. But that is delusional. Work, as much as family and community life, is a source of enduring and rewarding relationships, and while we’ve gone overboard in sacrificing the latter for the former in recent decades, technology seems set to simply swing the pendulum the other way now. It is still an either-or between two major categories of relationships. The fattened tail will never slim again, it seems to me. We’re never going to have an unfragmented personal and communitarian life again, because the globe remains 24,000 miles in diameter, and now that it is connected, it is highly unlikely that any web of high value creation will remain localized enough to feel comfortable to our ape brains. We’ll all continue to drift, or put down roots somewhere out of necessity, accepting a far-from-optimal median radius of community life.

Not that this is a consequence of new things. It is just the culmination of a dynamic that started when we first created the distinction between ‘work’ and ‘life’ and created the problem of ‘work life balance’ back around the invention of agriculture.

Read the follow-up article, “The Twitter Zone and Virtual Geography”

Get Ribbonfarm in your inbox

Get new post updates by email

New post updates are sent out once a week

About Venkatesh Rao

Venkat is the founder and editor-in-chief of ribbonfarm. Follow him on Twitter


  1. tubelite says:

    Different axes: distance, mode of communication (face to face, telephonic, email, physical transport), type of communication (casual, goal-driven) A few important combinations:

    For spontaneous, casual, face to face communication (different from goal-oriented, planned communication, like work-related or relationship-related) I find that the 50 foot rule works for me. If you’re not on the same floor and within 50 feet of me, you might as well be on the other side of the planet. (Things like are trying to chip away at this)

    For planned, goal-oriented communication, distance matters little: there’s no difference between 1 mile, 100 miles and 1000 miles. It only begins to matter when the earth curves and you’re no longer within the same “synchronous communication” time zones, at which point voice and IM give way to email as the default. i.e. I usually talk to UK folks, but email US folks.

    So…as far as I am concerned, there are 3 levels of distance : 0-50 ft, 50ft – X (both parties can talk when they’re awake), > X. I am completely ignoring communication by physically moving your butt across country, when the difference between 1 and 1000 miles is substantial – too lazy.

    Also: does social group size (Dunbar’s number – ~150) change with all these newfangled modes of communication? Not for me. Inside the Monkeysphere is a funny and insightful article on the same topic.

  2. Hmm. The communication-by-transport is definitely on its way out. The costs are simply too high, esp. with gas prices. People who insist on old-fashioned face-to-face will lose to people who learn virtual workflows. I think twitter is cute, like its name, but it adds no real value that I can see.

    I didn’t think about the types of communication making a difference, but of course you are right there. Though sometimes you’ll need the >X and near-face2face experience, so I expect you are not always able to relegate US-India conversations to email.

    Thanks for the reference to Dunbar’s number. Didn’t know such a thing existed. Reminds me of the stuff by Desmond Morris (The Naked Ape, and the Human Zoo), which defines a tribe as the largest social unit where you know everybody. His sense was that the limit was 20 if I recall correctly, but that came out of guesswork, not a regression equation like Dunbar. Our social instincts are only good enough for handling tribal dynamics. We need technological crutches (like family trees, phone books, LinkedIn) to handle larger numbers. One the face of it, 150 seems huge, especially if, as the article on Dunbar’s number suggests, you are talking about knowing the whole web of relationships, not just each individual. Assuming a small world type or scale free type social network, you’d need to keep track of about log(150) hierarchical or flat clusterings I think. At least its not as bad as 150*(150-1)/2, since most people won’t know each other.

  3. tubelite says:

    Let me pick that up in reverse order. Dunbar’s number also comes up in Gladwell’s The Tipping Point, which is yet another book I’ve not read. The number appears large, but it’s certainly possible to imagine a village or a tribe of 150 which is a fully connected relationship graph, and the entire graph is known to each person.

    The Monkeysphere article makes an even stronger assertion: that we are actually incapable of recognizing, empathizing and treating more than this number of people as fully human. All others are caricatures, movie extras, the redshirt guys on the Star Trek away teams. Another important point made is that simply knowing random human facts about someone immediately tugs at our empathy, and makes the person more “real”.

    This is hardly a new idea; every hack writer knows about “character development” and tries to embed quirks and idiosyncrasies in their characters in the hope that it makes them plausible. But it’s still amazing to watch this play out in real life. I’ve seen six months of email cold wars disappear over a single lunch, as you realize that the other chap is not, like you’d secretly suspected, a horned devil whose sole purpose in life was to harass you with review comments, but a rather mild-mannered guy with 3 kids, whose Deathly Hallows predictions happen to match your own.

    There is a clear and obvious ordering of email and physical visits apart from email to deal with > X distances. Not just to overcome issues with email between the Antipodes which stem from very long round trip times, but also to subtly remind everyone that they are dealing with humans on the other side, rather than email addresses. A single physical visit can have long lasting consequences on the nature and tone of further email conversations.

    Which brings me to my last point: one of the defining characteristics of the 50 foot radius is that you are exposed to random, useless facts about the people contained in there. This is both the cause and the effect of striking casual, spontaneous conversations. You know that A has started tennis lessons, that B’s kid has been paining her for a bicycle. They know that you’re going to Delhi this weekend. You might ask A the next day how his lesson went. As per the Monkeysphere principle, it’s these trivia (otherwise informationally useless), which significantly contribute towards humanizing these people to us and vice versa, bringing them closer to membership in our own “tribe”.

    Twitter is similar, but a one-way street. It lets you broadcast, casually, spontaneously, with a low barrier to entry, (like IM and unlike the deliberate purpose of email and blogging, which need more time and structure, thus have a high barrier to entry) something about yourself. I might know, that my friend across the world is visiting Home Depot, that another is slacking off from work and watching Life of Brian. These are not things which they would have emailed or called me about, but this is just the kind of trivia I know about my local friends simply by being around them. For people who haven’t met face to face at all, twitter would be even more useful as a “humanizer”. So it goes beyond cute; I think it has the potential to open up the 50 foot barrier to some extent. Imagine a voice twitter where you record and upload sound bites with your mobile/GPRS. Imagine playing sound-bites from your friends across the world, while driving, for instance. Compare to podcasts.

  4. tubelite says:

    Part of my previous comment got munged, probably due to my not escaping the “less than” sign. Why the heck can’t WordPress provide a preview before posting?! Let me repeat part of the 4th para:

    There is a clear and obvious ordering of “email less than voice less than face2face” when it comes to humanization of the participants. So, like you say, we do need conference calls (audio and video) and physical visits apart from email to deal with greater than X distances.

  5. Hmm, okay, I didn’t quite think of twitter in that way, but that makes sense. Though it is yet another high frequency interrupt-annoyance idea that requires careful gating, like instant messaging, it does make the antipodal person more human.

    You are right about the effect: the few times I’ve met online buddies for the first time, it has had a dramatic effect in changing my opinion of them and toning down the violence of future interactions. Online dating has the same problem in reverse, thinking people to be more amazing than they are in person based on limited information and optimistic assumptions.

    I totally hadn’t thought of the “random crap” you learn about people in the 50 foot zone that humanizes them. This is I think a pretty important point that relates to the frame problem in AI. AI reasoning/planning systems often make the closed-world assumption to deal with the problem of exploding numbers of irrelevant predicates to keep track of. You simply assume that any proposition not asserted is false. This solves the problem of unnecessarily worrying about block A being on top of block B while you are reasoning about how to restack blocs C, D, E and F. But it happens at the cost of not knowing what else is true about the environment that you don’t need to know for your immediate problem. In the long run, developing a coarse model of the whole world is smarter than the closed world assumption, and I think you’d compute more efficiently if you did that. Real human thinking is somewhat smarter than closed world AI methods, but not by much. Couple of recent books (“Stumbling on Happiness” is one) really look at how we are wired to make assumptions about things we don’t know.

    So the key question is, how do you humanize the people at the other end of the world without making everybody live in a fishbowl?

  6. People at the door, telephone calls and IMs are interrupts which oblige us to service them synchronously. That’s why their nuisance value is high (e.g. door to door salesmen, telemarketers). Email, twitter feeds etc. are not sync. You can deal with them at your leisure, or ignore them if you so choose.

    Especially when it comes to dealing with “random crap”, it’s OK if you miss a few of them. So something running down the bottom of your screen, which scrolls or transitions discreetly to the next twit so as not to disturb your foreground task or your concentration, but which is there to see if you take your eyes off your work and relax.

    Everybody lives in a fishbowl to a certain set of people – the 50 foot horizon. We just need to remove the geographic restriction but keep the number of people in your “inner circle” low (which happens naturally if it’s a 50 foot horizon). You could twit at different radii. e.g. update on relatively sensitive things like medical or financial events will go to only those who you deem part of your intimate clique. But one-line movie reviews like “Saw Blue Umbrella . Beautiful.” can go to everyone.

  7. Hmm… I think we got somewhere here. Combine the sociological 50 foot rule with some form of a Dunbar number, and you can define a virtual, sociological distance metric that maybe you can even apply to the visualization problem I posted about earlier.

    The former is about interactions, while the latter is about overall interpersonal information load-bearing capacity. You can probably interact intensely with a few people (like when a team of people is trying to launch a startup on pizza and late nights out of one person’s apartment), or twitter to a lot of people. On the spectrum between intimacy and caricature of your models of other people, there can be a varying mix.

    The one twist is that as you grow older, you’ve had episodes of close interaction with increasing numbers of people who cannot go back to being caricatures. So gradually, maybe a form of entropy builds up and you run out of personal social computing ability.

  8. I found the reference to the 50-foot (50 meter rather) rule: a couple of colleagues at work pointed it out, as did a LinkedIn user when I asked the question there. Article updated appropriately.

  9. billswift says:

    I didn’t look up your references for Dunbar’s Number, but wanted to point out an independent support for it. Mennonites, communal farmers, rather than let their communities keep growing, split their communities into two when their numbers get much above 150 (max of about 220). Can’t remember the source, I read it over ten years ago.

  10. Terrific post. I came upon it while freshening up a presentation I’ve been doing for many years that includes reference to Tom Allen’s original research done for Digital Equipment Corporation (RIP). Engineers were about to be dispersed around the world; the CEO Ken Olsen was terrified that all innovation would stop; hired Tom Allen, who found that if the engineers were more than about 15 meters or 50 feet apart, they didn’t talk much anyway. Jeff Stamps and I have written about this in our books and it always gets a rise from the crowd when presented. I’m interested in how this has changed given technology, and, googling, came upon your post. Thanks. AND I never knew about Dunbar’s Number but have also observed this same “rule” in effect in organizations, originally at WL Gore (GoreTex) where the founder split units apart when they got to 150 in size. So…I’ve noticed a virtual 50-foot rule in effect, with an ever-changing group of names filling my inbox. Never hundreds (I’m an indie consultant not an exec so perhaps that explains it) but always a handful, whose names come up again and again and again.