Blog: Evite

For most of 2002, I lived in Cupertino, California, a sector of sprawl in Silicon Valley near San Jose, which is to say pretty far from the more exciting centers of the Bay Area like San Francisco or Berkeley. One of my biggest complaints about living there (and in the Valley in general) was that the driving necessary just to go have dinner with friends in Berkeley (about an hour's worth) was prohibitive enough to make it uncommon and require a lot of advance planning.

Despite the distance, we (the four people sharing that house) hosted afternoons of board games, cooking, and music playing every one to two months that were quite well attended. We had the advantage of a large house (the benefit of living in suburbia) that could comfortably fit fifty. Some months the number of people from Berkeley outnumbered the number of people from the Valley itself. One time, six separate cars of people made the pilgrimage down from there.

Now, while we are all nice folks, I'm under no illusion that 20 people would regularly drive an hour each way just to see the four of us. Rather, the gatherings themselves possessed a measure of social capital. An individual deciding whether to go would remember that last time there were 40-50 people there, many of whom friends from college who he or she doesn't see frequently. The presence of so many people worth seeing makes it worthwhile to attend. But each person is making the exact same calculation independently. One gathering is valuable because it's well attended, and it's well attended because the last one was valuable. The gatherings have a social inertia all unto themselves.

This inertia takes time to build, but sustains itself almost automatically. I recall acquaintances of mine who never came to a gathering until one time, then enjoying themselves and attening regularly afterward. It's possible these people just were always busy until the first attendance, but I believe that they happened to be underestimating the value of the gathering to themselves and therefore not prioritizing attendance as highly as they did afterward.

If many people underestimate the value of a gathering, then it becomes self-fulfilling since attendance ends up being low. There are a few ways to combat this perception problem. One is to be a very charismatic person who many people want to attend. Then, there isn't an estimation error because going just to see the host is worthwhile in itself, and it doesn't matter as much what kind of prediction people make to themselves about the other guests. Attractive and/or personable women especially benefit from this technique, because everyone wants to be around them. Similarly, having a few hosts (as our gatherings in Cupertino did) helps to bootstrap the attendee count early on.

Evite provides another method. With Evite, potential guests can see who else is attending and what they write. Therefore, if I'm considering going to a party hosted by someone I don't know that well and therefore wouldn't go just for him or her, I can see if the rest of the attendee list looks promising. There's not a lot one can see on that list, mainly the first names of the respondents and their comments. But that's enough to see if it looks like a lot of women will be there, for example, which has definitely swayed my decisionmaking from time to time. The "+1 guest" label often means the person is attached, but sometimes there is more information given. Alternately, as with the Cupertino gatherings, guests look for whom they know on the list and are more likely to show up if there is a critical mass of friends and acquaintances.

With Evites, we end up with a "tipping point." Once there are enough friends or cute sounding girls, more people will decide to go, further growing the list of attendees and creating a snowball effect. If few people respond initially, many others may bide their time and remain undecided until later. Those people may not be aware that their actions affect others, but they do. Just like trying to be "fashionably late" and arrive around the average point, many people want to RSVP on the Evite around the middle of the pack.

It would be fascinating to get a look at some Evite usage data to actually establish these tipping points. Is there a certain number or percentage of attendees that leads to a "tip"? Does the presence of clearly female names make a difference? Numbers of additional people being brought by the respondents (like the "+1 guest" labels mentioned earlier)? Does it matter what people write in their comments? What about the people who aren't going - does it make a difference if they just mark themselves as not coming, versus adding a note like, "Wow, that sounds so fun! Too bad I'm out of town at a wedding that weekend"? How does this vary over the different types of events Evite distinguishes?

As for me, I am trying to resurrect the board game gatherings now that I know enough people to make it possible. But I ended up making a tactical error by using Evite; I had a good sized core group of people who really wanted to come, and the time I picked worked for those I talked to, but it turns out most of the rest of the core group are out of town. There are plenty of people in town, but most are from the class whose decision also depends on the other attendees - meaning they are waiting to see who responds first. Ultimately, those who do attend will enjoy themselves and the social capital of these gatherings will grow over time. Alternatively, there's always the phantom woman stratagem.