NA IMPRENSA INTERNACIONAL > AS MÉTRICAS DO TWITTER
The metrics we really need from Twitter are not the metrics we have
Por Jeff Sonderman [Poynter, 25/9/12].
Twitter co-founder Ev Williams acknowledged Monday that available metrics like follower counts are not great measures of success.
At a panel discussion in New York hosted by BuzzFeed's Jonah Peretti, Williams said the follower count "doesn't capture your distribution. … The dream metric is how many people saw your tweet."
And coming from Williams, that's not just wishful thinking. He currently sits on the Twitter board of directors and oversees the company's product strategy. So is Twitter working on a better measurement tool for users, and what would it look like?
We know Twitter has an analytics service that so far is offered only to advertisers. It measures things like the number of retweets, mentions and clicks for each tweet, the number of followers gained and lost over time.
That would be a nice start. Unfortunately, Twitter CEO Dick Costolo just told the Online News Association convention there's still no timeline for when its analytics might be available to non-advertisers.
Twitter analytics exists already, it's just not out yet.
There are several layers of potential metrics that would be helpful to Twitter users. Currently we have good measurement tools for some, workarounds for a few and almost no idea about the others.
* Follower count: We know this number now, it's just not very useful, as Williams said. In addition, it would be useful to know more about the geography and demographics of followers.
* Active followers: We should know how many of our followers are real people who use Twitter regularly. A third-party tool called Fake Follower Check attempts to analyze how many of your followers are either fake or inactive users. It's not perfect, though – it only takes a sample of your followers and it relies on educated guesses about their status.
* Reached followers: How many of your active followers were exposed to a particular tweet? Only a fraction of Twitter users are active at any given time, so it would be extremely helpful to know the number that actually read your last 140 characters. This is deeply internal data that only Twitter itself could provide. The upside of Twitter's recent crusade to push users toward its own website and apps is that tracking users' activity may be more feasible.
* Engaged followers: How many users not only read your tweet but did something with it (retweeted, replied, favorited, clicked, etc.)? And which of those engagement actions did they take? These are the metrics Twitter's current analytics system tracks best. You can also get imperfect measurements through Bitly click data, the "t.co" referrals in your website analytics, or engagement trackers like Twentyfeet or Crowdbooster.
* Non-followers reached and engaged: Finally, this metric would tell you how productive your engaged followers were in extending the reach of your tweets. You got five retweets? Great, but how many more impressions and engagement actions did those people produce from their own followers? This also is data Twitter itself would have to provide, although services like Tweetreach can add up the total number of followers a tweet potentially could have reached.
Analytics is one area where Facebook is miles ahead of Twitter. Facebook Page administrators get access to the Insights tool, with metrics such as total likes, total people engaged and total people reached. You can drill down through those metrics for individual posts or time periods, and you can get demographic and geographic breakdowns of the audience.
It's long past time for Twitter to catch up. Here's hoping Williams is pushing the company in that direction.