A few months ago I got a lot of great responses in relation to my “quest” to connect how people use Twitter, with the concepts Malcolm Gladwell laid out in the Tipping Point (see here and here). While progress has been stalled on this front (even though I’m convinced that there’s enough data out there to do a thorough analysis of it), I’ve continued to be interested in the topic, and have recently been focused on looking at how different brands are using Twitter.
My objective during such explorations is always to come up with a simple framework for demonstrating different behaviors, with a particular eye on develop two by two matrices, which former colleague Alex Lowy literally wrote the book on (seriously – you can look it up). I’ve come up with a variety of them to test so far, but the one below is the one I find most interesting so far (with a very small sample of a few of the brands/ accounts I have looked at so far):
The basic idea is simple. The X-axis tries to seperate out brand-centric Twitter accounts from individual-centric accounts. For example, if you follow Marvel Comics, it’s very brand-centric – no particular individual is featured. In contrast, if you want to keep up with what Ford is up to, you are likely to follow ScottMonty – the head of social media for the company. This is what I call “individual centric”, though I’d like to come up with a better term. The argument is that these are two very different approaches / experiences – but importantly neither is neccesarily right or wrong. Also, as I worked through the data, I’ve noticed that many, many companies / accounts fall somewhere in the middle. For example, SouthwestAir may initially appear to be brand-centric, but the bio makes it clear that it’s Christi you’ll be dealing with.
The Y-Axis is a somewhat crude metric to indicate how the accounts are being used – broadcast centric (i.e. sending a message to your followers) versus conversation centric (i.e. responding to individuals). Now to be upfront, the data sample I used for this initial exploration was fairly small – activity over a select few days in March – so they are subject to change in a more thorough quantitative analysis. But basically what I did to come up with my “crude metric” was take the number of tweets that were @ replies to someone (i.e. answering a question) as a percentage of total tweets to come up with what I’m calling the “conversation ratio”.
So tying back to the graphic above, in the sample of companies I looked at ComcastCares was the most active account, averaging almost 70 tweets per day. Every single one of them was an @reply to someone – either responding to a question, or trying to engage in a conversation. There are a variety of other companies / accounts (like Whole Foods) that seem to follow the same model. I threw Al Gore up there as an example of the opposite extreme (even though he doesn’t really use the account much) – he’s basically all broadcast messages (here’s my thought on this, go look at this, etc.). The Zappos account (driven by the CEO) is actually quite similar – mainly broadcast messages. The bulk of Twitter accounts obviously operate somewhere in the middle – with conversation ratios on average in the 60% range (again, based on a fairly small data set – I’ll state again this is the start of the exploration, not a final report).
So what do you think – could it be a useful framework? I obviously kind of like it, and think it’s a fairly intuitive base to start from. The next obvious step I see is trying to figure out which approach works best for which purpose. I’d bet most people in the Twittersphere would naturally lean towards the “more conversations = better” approach, but I’m not so sure – as it appears that those people that broadcast more than they converse are amongst the most popular in terms of followers. But I might leave that discussion for another day…
{ 4 comments… read them below or add one }
Re an alternate name for “individual centric” — what about “personality centric”?
yes – that is much, much better. I like when the answer is so obvious – once someone says it to you – that you feel a little stupid for having missed it!
The Y-axis could alternately be labeled the “clue” axis, since people using Twitter as a broadcast medium don’t have one…
Great visual! I’m thinking we (SAS) land somehwhere close to Scott Monty. I’ve been wondering lately about oversaturation when you go that route and up with dozens of individuals from the same company using similar behaviors. Even though the individual behaviors are, for the most part, appropriate – does the combined volume of all those people start to become too much?
{ 7 trackbacks }