I was watching tonight’s Letterman interview with Joaquin Phoenix, bizarre at the very least – check it out. Three or four minutes into the interview, we’re trying to work out if this is just weird or if it was W-I-E-R-D. First place we go was Search Twitter, where pretty much up to the minute postings confirmed it…Yep, this is indeed W-I-E-R-D (check out yesterday’s Twitter postings on this subject here). Once we saw all of this, we went of to check out Juaquin’s interview history on IMDB and it seems that he’s done that type of thing before.
A couple of lessons for you SocialMediaMarketerati here, this is how things work today, consumer wants up to date information, goes to Twitter and gets it on the button, then can run over to YouTube to confirm, because after all, seeing it on YouTube is believing (lonelygirl15 excepted).
When I went off to Advanced Search Twitter, I did notice that there was a Sentiment Check-Box where you could filter all positive or negative messages or those that asked a question. I first checked the positive filter (for Feb 12th) and got one response. Then the negative and got a couple more. I finally checked the question filter and there were the pages. I suspect that Twitter are now using some simple text processing in a further effort to try and add some value to the analysis, afterall they are about to change their business model to try and add value for paying business users. Facebook are also sitting on a goldmine of text that demands some pretty sophisticated text analytics – I wonder how accurate their Engagement Model actually is.
There are also a number of other Text Analytics products on the market, however, I know that Overtone’s OpenMic product has a dashboard that includes a very accurate Feedback Sentiment Index (FSI) that is worth checking out.
Just getting back to Joaquin and Letterman, it didn’t take long for the Gawker’s blog post on the interview make it to #2 position with Google search terms <joaquin>, <phoenix>, <letterman>.
Time for bed (because the interview only finished 30 minutes ago).