By B.L. Ochman
Do you really know what Twitter statistics measure? If you think it’s influence, think again. And if you think you’re using the best tools to measure Twitter, you’d better take another look.
I’ve tested more than 45 of the current crop of 350+ paid social media monitoring platforms – which can cost $10K a month and more – and found their measurement of Twitter stats were all lacking in some way. But there are some free (or very cheap) Twitter search tools that come close.
FollowerWonk – which just updated its platform on Friday – and Twiangulate – which has been great and getting better for a year or so – are the best of the many dozens of Twitter tools I’ve tried. Add ReTweet Rank to the mix, and you have some pretty impressive intelligence about Twitter
Twiangulate was developed by Henry Copeland and the team at blogads.com because they wanted a better way to measure Twitter stats. Twiangulate’s FAQs say “Twiangulate discovers hidden tweeters, friends of friends (or friends of enemies), micro-influentials who only insiders follow… or sometimes just friends you haven’t yet seen tweeting…After all, its not just the size of your following, its what you do with it.”
Boy is that right! Number of followers certainly aren’t what matter – you can buy Twitter followers on ebay. I guarantee you that most of those are bogus, but would still be impressive in a report to a CMO. And you can bet a lot of agencies buy them! So what does matter? Read on…
Twiangulate (left) and FollowerWonk features both include:
• Comparison of three Tweeps’ followers to reveal unique and shared relationships among their followers.
• Comparison of followers and/or people followed by three Tweeps
• Search Twitter bios by keywords and location
Followerwonk includes Wonk Scores – without revealing the formula by which the scores are computed. These, they explain are “proprietary rating of a Twitter user’s influence and engagement on Twitter. It ranges from 0 to 100, where higher scores generally indicate a person with greater influential activity on Twitter.” Ah, but what is influence? There’s a whole industry rising around that measurement and nobody’s got it right yet!
Twiangulate is more granular. In addition to the many features it shares with FollowerWonk, Twiangulate lets you:
• drill down to a Tweep’s inner circle;
• map the relationships between groups of people (like attendees at a conference or a CRM list).
• create a Twitter list from the results
• follow any of the Tweeps in the results whom you do not already follow
• Map the list to show what percentage of the rest of the list the Tweep is connected to, and display a list of the Tweep’s recent Tweets
• Create searches with advanced sets of criteria like hiding people followed or being followed by more than a certain number of people
• Display results in table or interactive form
• Create and save a watch list for your search that will be updated as the data changes
• export the data to excel,
Perhaps most importantly, Twiangulate says “to eliminate the noise of the millions of spam tweeps and focus on intentional follows, we limit the results to tweeps who follow fewer than 11k people who have 1.5 times more followers than friends.”
FollowerWonk’s (left) useful new features include:
• account ages of followers and friends – assuming that “…users who have been on Twitter for a while are early adopters, and generally more valuable than newbies. As well, there is a higher percentage of spam accounts in newer user segments (as Twitter has not yet cleared them out.”
• Tweet recencies – how recently followers have Tweeted. This information is useful because “If a person hasn’t tweeted recently, then their account may be dormant”
• Friends’ Tweet counts “if a user has a low number of total tweets, they may not be a “real” user. Perhaps they tried Twitter and gave it up? On the other hand, those with a very large number of tweets might be avoided (they could be spammers, overwrought, or just too verbose!)”
• Languages of your friends Tweets.
• Word cloud from friends’ bios – very useful IMO because the words one uses in a 140-character bio are very likely to be very good descriptors.
• Friends’ location word clouds – where their bio says they’re tweeting from.
Add ReTweetRank to all of this, and you get a comprehensive picture of how often a Tweep is re-tweeted, and by whom. The re-tweeters are ranked by number and then by the “influence” which is determined by the number of followers a re-tweeting Tweep has. And that’s useful if half of that person’s followers aren’t spammers or bots. ReTweetRank presents an interesting set of stats, but brings us back to square one on what influence really means.
Results are still far from perfect
So, you see, we have a long way to go before we understand Twitter’s impact, let alone Facebook or blogs.
I still say that Google and Twitter soon will put all of these platforms out of business – or buy them. Until then, you can’t really know for sure whether your marketing dollars are being spent wisely, but you definitely can make lots of pretty charts and graphs.