This is a followup from last post on Social Media Monitoring Case Study for Columbia Business School Alumni Club – Part 2 where I was able to cull data from Radian6 and determine how much twitter traffic was generated over the last year for Havana Central and by whom.

Today I want to focus on how many of those “twitter handles” actually showed up at some point and went to one of the three restaurants.

Again, I sorted by those who visited the most and I determined it by focusing on phases in the tweets that told me they were likely going to, at, or had been at one of the Havana Central Restaurant locations and referring to this in their tweet.

Here’s a link to this aspect of the  case study -3 – those who visited a location and talked about it.  I won’t swear that I got rid of all the noise, but did manage to get most of what I wanted in this file.

One thing I would think makes sense to look at is velocity -  the same study I did here could be broken down to weeks instead of months – or daily – if desired – it would just be more work.  The point being  – this is information you (we) can actually take action on.

The question then becomes, what action do we take.   If we can engage some one as they are in a location, eating (because it’s contextually relevant and immediate – which works well with the internet, as we know) , so much the better – but even if we can’t – we can reach out, build our lists, span our loyalties and start counting what ROI is.

And I think, at the end of the day, the ROI here is in cultivating a house list – building it, refining it, promoting off of it – and tracking it.

Sure, we can do A/B tests on the list too – if we give a drink or a meal to someone who comes often and track that – will there be more lift than if we just randomly gave away free drinks or meals to some lucky persons inside the restaurant.

I don’t think we know the answer to those questions yet – but I see the need for a “lab” type approach – and as I was talking to some people at Altimeter Group recently (they had a webinar today, unfortunately, I could not attend it – though my name was to be mentioned in it) – if I had the chance – I’d love to lead a lab – where we try out stuff like this – and find what works – build our framework out of what we are actually doing and we know how to measure (or creatively figure out how to measure) and then build some case studies around each measure of a framework.

Cause right now – I see a lot of framework type stuff – but when you try to apply it – you can’t – it doesn’t work – too many pieces missing – not enough thought given to how to apply ideas that are written about.

Anyway, enough for one post – maybe if anyone has any feedback on this case study – comment please – I’d like to hear your thoughts and promise to respond to ever comment.