Over the past few weeks, I’ve spoken about Facebook and Twitter to a few friends who work in tech in New York and Northern California. We spoke about self-promotion — how much promotion my friends felt they did, how much others seemed to do, and how it all felt. We spoke specifically about how Twitter is used in our professional circles.
Criticism of social networks has centerd on a lack of “meaningful” conversations. One critique suggests people use Twitter, Google Buzz, Facebook, and similar tools to talk about themselves, rather than talk to others. By some (dubious) estimates, 6% of tweets contain explicit self promotion (e.g. “Read my new blog post” or “Check out the new scarves in my Etsy store”), and 80% of tweets are about their author (e.g. “I won my kickball game Sunday afternoon.”) “I was shouting into a vast echo chamber where no one could hear me because they were too busy shouting themselves,” Leo Laporte concluded in a scathing blog entry. I’d put Malcom Gladwell’s criticism of Twitter activists in this category as well.
Earlier criticism of Twitter, circa 2007, centered on over-sharing mundane details. What did you have for lunch? What does your infant daughter say? How lovely was that sunset over the Marina? The excellent @forlunchihad Twitter account turned this criticism into commentary, but it dogged Twitter for its first few years — and still does in some circles.
In the past three years, the naysaying around Twitter in the tech community has broadly moved to the first person from the second person. The shift to “No one cares about me” from “No one cares about you” may seem subtle, but there’s a stark difference in expectations: we seem to expect the people who follow us to care about what we have to say, and we’re upset when they don’t. Yet we’re not using Twitter as a diary; we’re using it to push information in a mostly un-targeted way, and when we do that, we’re emphasizing tweets (content) over other users.
Three aspects of Twitter’s design encourage this:
1. Targeting is unnecessary – One argument goes like this: Twitter streams surface interesting tidbits, but they can lack consistent quality because “there’s too much noise.” Yet “too much” noise implies there’s less-than-optimal targeting, and that’s not likely what is happening. The cost of reading a tweet is low, and the benefit of coming across an “interesting” or “valuable” tweet can be high. So tweet-reading is a high-variance, potentially high-average activity. [1] Your followers, who best understand their preferences, are better positioned to sort your tweets between “irrelevant” and “interesting.” Contrast this with email, when receiving an irrelevant email has relatively higher costs: you have to open the email, determine it’s irrelevant, potentially reply (it was, after all, sent directly to you), and then dispose of the email. Phone calls have even higher costs. On Twitter, the costs of targeting are often higher than the benefits of having done so.
2. Following is optional – If you dislike someone’s tweets, it takes two clicks to receive no more of them. Twitter does not force you to consume content you find irrelevant or uninteresting. (Although social pressure can.) Contrast this with mailing lists. The “quasi-Darwinian” aspect of Twitter’s network means two things: users are encouraged to think about the reception of their tweets before tweeting, and users receive immediate feedback. (Hi, declining follower counts!)
3. Re-following is an acceptable practice – Decisions to follow or unfollow are neither permanent nor high cost. Edge cases notwithstanding, there is (almost) no such thing as an unwelcome follower. Accumulating another follower — another person interested in your content — is flattering. If you consider followers who have followed, unfollowed, and re-followed you “less welcome” to your stream, you’ve implicitly downplayed your future content. That’s tough to do.
There’s not a “right way” or a “wrong way” to use Twitter in absolute; instead, each community on Twitter (startup tech, K-12 teachers, Buckeye football fans … ) sets its norms. Twitter’s network provides social cues (e.g. @replies, RTs, and link analytics) to re-enforce those communities.
Some may complain about Twitter’s structure and the resulting behavior, but I don’t think it has changed human behavior. It’s simply made some actions easier than they had previously been.
Notes:
[1]: What is the cost of scrolling past an irrelevant tweet? What benefit do you get from interesting information? Imagine the cost of reading an irrelevant tweet is -2 units, and the benefit of a valuable tweet is 20 units. Under these assumptions, you’d be willing to read 10 “bad” tweets for every one “good” one. Expected value (0) = (-2) * 10 + (20) * 1. Any more bad tweets without good ones, and you’ll give up reading your stream as it currently exists; any more good tweets without bad ones, and you’ll keep reading. These numbers are baseless, and the example simplified, but I think it’s directionally correct and hopefully helps illustrate a broader point.