Earlier this month, to paraphrase Mr. Kenobi, a million @replies cried out and were suddenly silenced. By now, much has already been written about the #fixreplies fiasco, but what I haven’t seen is any analysis on how the mobs were gathered, and when and where the pitchforks and torches of the masses wound their way through the conversation. Indeed, upon closer inspection, it seems a large proportion of the mob was unarmed.

First, the numbers:

In the first 26 hours of the uprising, there were 20,463 tweets with the #fixreplies hashtag from 15,423 people. Not a huge number, considering twitter’s userbase is estimated somewhere in the tens of millions. However, another recent twitter blowup, #amazonfail, reached only 14,162 tweeters in the first 26 hours, so perhaps that mass is more critical than the overall percentage would suggest.

But what do we define as conversation? For starters, consider the retweet ratio: 11,027 retweets vs. 9,445 tweets. Over half of the overall conversation is an echo. In fact, at 5am, when #fixreplies was 6 hours old, retweets peaked at roughly 65% of the hourly conversation, and sustained that level for another 5 hours. For those five hours, the echo was twice as loud as the original content. And like any echo, it lost strength over time. After that 5 hour stretch, retweets fell to 40% of the hourly total, jumped to 60% the following hour, but fell towards 30% over the next several hours.

Perhaps more telling, 8,682 of the individuals involved were ONLY retweeting the content of others. To return to the earlier metaphor, over half of our agrarian uprising left their pitchforks at home and were sharing with the other half. Still an impressive sight, but somewhat less deadly.

The norm?

So is twitter really the echo chamber it appears to be? Looking again at #amazonfail, we see an echo in full effect. For each of the first 6 hours of #amazonfail, retweets averaged 40% of the hourly content. After that, RTs slid to 30%, bounced back to 40% and then declined to around 20% of the hourly total. But unlike #fixreplies, this mob is more diverse in its methods. Rather than the domination of aggregators over creators, the opposite is true with #amazonfail, with roughly half of the tweeters involved tweeting only new content, and another 37% tweeting mixed (RT & original) content.

So what?

The mobs still arose, the respective brands were still torched in the town square, and they’ll still be fodder for hundreds of “what not to do” conference sessions. Can we sift anything new from the ashes? Sure.

  • Take a good hard look at how your audience is speaking before wading into an issue.
    After twitter posted a compromise position on their blog, the retweets began to die down, taking the overall volume down several notches in the process. If you can appease the aggregators, the voice of the content creators is blunted. Without the echo, the voice doesn’t carry as far.
  • Personal and complex issues breed personal and complex responses.
    With #amazonfail, the issues involved were both highly personal and highly complex. To properly explain a position, a tailored response was necessary, and agreement was shown through the @reply more often than the retweet. With a focused campaign such as #fixreplies with one clear goal, the retweet was a simple tool for a simple job. In a campaign such as this, the response is much, much easier to tailor – provided the communicators on the receiving end are listening.
  • Users at the higher end of the follower spectrum play key roles toward the outset of an issue.
    The initial conversation spikes in each issue showed a dominance toward these super-users, with a corresponding ripple focused among the middle of the road users a few hours later. More study is warranted, but it stands to reason that if the super-users can be addressed early enough to avoid the initial spike with even a “we’re working on it” message, a situation may be defused before a tweetstorm level event kicks in.

Of course, early warning is key to staying ahead of a storm, but in the event you are caught out, analysis of the situation is key before acting. It’s best to at least look out the window before sticking your head out in the rain.