Observations on social experience design

Aug 20 2011

In my last post, I wrote about core strengths and weaknesses in the social experience design of Grub With Us, a new Y Combinator-backed social dining service that recently closed a very notable round of angel funding.

I argued that the nuance and judgment involved in human curation of social experiences is hard to “algorithmize,” and that the next big breakthrough for the social web will, I think, be in figuring out how to cultivate and facilitate the creation and development of “strong ties.” A social dining service that helps you meet new people by bringing interesting strangers together is a great start. But what I want as a user is a service that provides greater structure to social interactions in a way that facilitates stronger relationship development. 


The problem of strong ties 

Typically when I meet someone at a mixer or happy hour, it’s unlikely we’ll stay in touch unless there was a special connection, or something unusually interesting about them, or a strong shared affiliation (like having the same professor in school). This doesn’t happen often. Even if it did, it may not often reveal itself in the first encounter. Nevertheless, the reality is the most valuable connections to me are those that actually deepen into more meaningful relationships, rather than those where my main interaction is lurking the other person’s LinkedIn profile. In other words, what I really want is strong ties.

But what drives the shift from weak ties to strong ties? How do you design a truly engaging social experience that incubates strong ties? And what does this tell us about core principles of social design?


Observations on the dynamics of strong ties 

When I was living in New York, particularly around the time I was preparing to move to California, I spent a lot of time organizing dinner parties and social gatherings. I brought together friends — and friends of friends — frequently to meet, mingle, and hang out over thoughtfully curated social events. Typically, I would anchor a social gathering around a well-defined event, say, a theater event or a summer parks concert. I would then manually curate the attendance list through a painstaking process of individual e-mails, Facebook invites, in-person conversations, group status updates, and continuous calibration of the attendee mix so that the optimal mix of friends — balanced by background, interests, gender, fit, conversational compatibiliy, and familiarity — would interact. Doing this iteratively over and over taught me key lessons about social design.

These key lessons were:

1. Bring people together in clusters. There should never be an attendee who knows no one in the group beyond the host. Otherwise, you end up spending disproportionate time easing their comfort level rather than fulfilling your true host responsibilities of facilitating interactions for the entire group. By contrast, if every attendee knows at least one other person besides the host, then they won’t experience awkward conversational lulls, a death curse in social design, because they can always revert to chit-chat with their “default” non-host friend.

2. Balance the gender mix as evenly as possible. Even if it is easier to rally friends who are the same gender as you, don’t. Try hard to persuade a balanced gender mix to attend, even if that means reducing the entire group size to achieve this balance. A balanced group yields far better outcomes for interesting conversations and a good experience than a group skewed in either direction. It’s about who bears the burden of easing others’ comfort levels: in a balanced group, the burden is spread evenly and therefore unfelt by any particular person; in a lopsided group, the minority always feels the burden uncomfortably, a death curse in social design.

3. Mix no more than 2-3 social circles at a time. Limiting each mix to 2-3 of your social circles makes it much easier for attendees to categorize and remember new people they meet by their association to you the host. (It also helps mitigate the problem of bringing people together in clusters.) If, for instance, an event has attendees only from your work, college, and volunteering circles, then attendees will more easily make (and remember) those connections. When they more easily remember connections, they are more likely to follow up afterward, increasing the probability of strong ties. Their thought process will be: “Who did I meet tonight that was cool? I thought Alice was cool. What was her connection to [you, the host]? Oh, right, they went to college together. I’ll Facebook friend her and see if she might want to meet up again for lunch sometime.” By contrast, if you compose an attendance list with 1-2 attendees each from 7-8 different contexts of your life, it can feel random and overwhelming to attendees. Although they may be unable to articulate it as such, they will feel: “Who did I meet tonight that was cool? I dunno, the event seemed kinda haphazard and chaotic. I thought Alice was pretty cool. But I forgot how she knows other people in the group. Maybe I should follow up with her sometime.” Then, some time passes, but no follow up happens. The opportunity for strong ties is lost. I am simplifying here, of course, and there are plenty of exceptions and counter-examples in both directions. But this insight isn’t about certainties. It is about probabilities; it is about propensities. And my experience has taught me that the more social circles (beyond 3) are involved in any event, the harder it is to remember specific individuals because it gets increasingly difficult to associate newly met people with the correct social circle they belong to. The harder it is to make those connections, the less likely it is that any strong ties will develop.

4. Mix social circles where interest gaps aren’t too large. This is about creating enough interest proximity between circles to kick-start conversations in the first place, but also architecting enough interest diversity to make those conversations engaging and refreshing. If you simply cluster people in the same profession together, guess what they’re going to do? They’re going to talk about work. Talking about work isn’t fun. If conversations aren’t fun, people walk away blaming the event itself for the disappointment. Guess who bears that blame? Yup, it’s you, the host. The other extreme yields even sorrier results. If you mix circles with yawning interest gaps between them — like artsy people and hardcore sports fanatics — it is unlikely in the first place that the circles will have enough to discover about each other to pique interesting conversations and, in turn, to generate strong ties. People will walk away feeling they made no connections at all, and blame the event for it (as opposed to feeling like there was at least some connection, even if the conversations weren’t fun because they were about work). The right balance involves bringing people together who have proximate enough interests to create relatively quick conversational ease and comfort, but diverse enough interests that conversations actually flower beyond work topics and into personal passions. When people start talking about personal passions, those conversations tend to be remembered as engaging and refreshing, and those memories are associated with the event itself — and the skills of the host who organized it.

5. Generally cap attendance at ~10 people per event. While there are always sensible exceptions to this guideline, generally if an event has more than 10 people attending, any particular attendee is unlikely to meet everyone; the people she does meet, she won’t get to know well. That defeats the purpose of curating an event in the service of strong ties in the first place. With 10 or fewer people, however, there is greater social pressure not to “size anyone up” too quickly before looking for a way to “duck out” of a conversation, since such behavior is more easily noticed (and shunned). Each attendee feels implicit social pressure to spend more time (on average) talking to other attendees, which given longer average duration increases the probability of discovering commonalities that can spark strong ties.

6. Be thoughtful about seating arrangements for sit-down events like dinners (and try to shuffle seating midway). Striking the right balance between seating friends together to preserve social comfort vs. mixing them up to encourage intermingling is an art that involves “facts-and-circumstances” judgment on a case-by-case basis. Some people pair better with strangers than others; sometimes the venue itself influences who pairs well together; single vs. couples attendees create their own challenges; and continuously updated attendance lists (and cancellations) can send seating plans back to square one. To complicate matters further, an ideal seating structure would involve at least one shuffling between meal courses to encourage people at opposite ends of the table to get a chance to meet. Of course, that is more work on the host because it multiplies the effort required to create a second seating arrangements for a single event (as well as disheartening restaurant staff trying to coordinate food orders). But thoughtful seating can generate huge social value fostering strong ties.

7. Be thoughtful about how conducive the event itself is to creating interactions that yield strong ties. My own view is that events at clubs, bars, and huge mixers are terrible for facilitating interactions because they are dark, noisy, uncomfortable, crowded, or all of these. That basically guarantees that interactions will be short, superficial, and meaningless. Events like organized movie outings also are not conducive to strong ties conversations because participants are unable to see, face, or even talk with each other. But what about a movie event paired with an after-show discussion over drinks / dessert? That can provide both a forum to mingle plus a topic starter (the movie itself). Or better yet, what about an outdoor parks concert or play, a food festival, or a museum outing? Events like these are intellectually accretive, which I strongly prefer because they attract a higher-quality group in the first place. But they also provide enough atmospheric flexibility for attendees to “half-engage” or fully engage as they wish, or mix and mingle if desired instead. The best events, I think, allow people to mix up, shuffle around, but hold longer conversations comfortably as well, stimulate conversation topics based on the event itself, and provide a warm comfortable ambience conducive to engaging discussions. These ingredients, when mixed together, create fertile dynamics for strong ties.

8. Be sensitive to how price and venue can create entry barriers for some attendees. For group events, I generally never pick non-food venues that cost more than $10 or restaurants that cost more than $20-25. I also pay special attention to free events. The reality is, people in your social circles likely come from different professional walks of life. Some people are better able to afford a fancy dinner or an expensive show than others. Some may be vegetarians; others may not be. Some may dislike loud noisy clubs; others may prefer them. Some may prefer outdoors activities; others want to dwell indoors. If your goal is to mingle different social groups regardless of income or other restrictions, then sensitivity to the affordability and appeal of the venue itself is important to avoiding attendance barriers that might otherwise arise because of venue. You don’t want to box yourself in unnecessarily by organizing an event that lawyers can afford but college grads cannot. In the end, it is difficult to solve for every attendee’s preferences and restrictions, but being thoughtful about them can definitely increase the accessibility for attendees. With greater accessibility, strong ties become more likely.

9. Organize events once per week; accept all invitations you receive. If the other insights are about quality, this one is about quantity. The more events you organize, and the more you attend, the more you are likely to forge new relationships yourself that yield strong ties. This not only directly deepens your own understanding about strong ties dynamics; it also expands and increases the social circles you can draw upon for future events. In some sense, the larger and deeper your network, the larger and deeper it is likely to become.


Communities with intrinsic strong ties 

Recently, I have also been thinking about how important principles of social design often flourish in one particular institution: the church. Church communities get many pieces of social design right and are powerful “social APIs” because their group dynamics are inherently well-structured to incubate strong ties. Insight about these dynamics can also be applied to social products. What are these dynamics?

1. Attendees who aren’t total strangers. Church attendees may not be best friends, but they aren’t total strangers, either, and are typically friendly and familiar with each other. Most churchgoers know at least a few friends in their congregations, and those initial connections can create natural links to meet other new people. This helps solve for observation #1 above: bring people together in clusters.

2. Strong common affiliation that creates community and fellowship. The fact that at least one central interest axis — religion — brings people together into churches helps establish baseline interest alignment among attendees. This helps solve for observation #4: Mix social circles where interest gaps aren’t too large.

3. Attendees who are already “vetted.” While churches tend to have strong acceptance norms to anyone wishing to attend, other norms — such as a baseline dress code, alignment of religious beliefs, a focus on values, etc — help filter out large interest gaps arising from randomness in the attending population. Population randomness is lower because similarly minded people, at least to a minimum degree, are more likely to be the ones who attend in the first place. This also helps solve for reducing potential interest gaps (observation #4).

4. Repeated facilitated structured interactions specifically designed to incubate strong ties. So many interactions in church settings — from weekly services and post-service lunches, to small groups, to church picnic days or community service days — are designed to facilitate the development of strong ties among attendees. Many religions, in fact, see group fellowship as central to the religious experience, and this motivates both the continuous supply and demand for organized events structured to generate those ties. This helps solve for observation #7: be thoughtful about how conducive the event itself is to creating interactions that yield strong ties

5. Repeat attendees week after week. Highly frequent well-structured interactions involving repeat attendees week after week helps solve for observation #9: increasing the probability of forging new strong ties through increased volume of interactions among a consistent stable population set.


Implications for the social web

I think the implication of all these observations for social web products are straightforward. But the devil is in the details. And execution is anything but easy given that social web products often provide only the roughest or partial simulation of how real-life social networks form, operate, shift, evolve, and interact. But deeply thoughtful consideration of — and personal experience with — the drivers of social design can help product designers architect more subtle, nuanced products that facilitate real-life outcomes that move us closer to strong ties. Alternatively, it can also help clarify those aspects of social design simply unreplicable by technology: an equally useful insight allowing designers to avoid those use cases entirely and scope their products more narrowly for replicable use cases instead.

What are your thoughts? Any key drivers or observations I missed? Any you disagree with? Any I only got partially right?

Andrew Chen is a tech entrepreneur and startup founder in the San Francisco Bay Area. You can follow him on LinkedIn, Twitter, Quora, and Google+.

No responses yet

Leave a Reply