Social Does Not Work That Way! Good Night! The INFORMS Marketing Science 2010 Preview

SOCIAL MEDIA DOES NOT WORK THAT WAY!!!

Do you remember Morbo from Futurama? Yes. Well, I certainly do channel my inner Morbo periodically, especially around the INFORMS Marketing Science Conference time of year.

This the fourth time I’ve tried writing this post – and not make it sound like I really am being Morbo. Here goes.

The conference is next week and I’ll be in room 14 – soaking up all the papers at this curious nexus of marketing, social media, data mining, statistics, and modeling.

Here’s a flavor:

On Thursday, Sabris and Grewal are presenting: “A Blog-eat-Blog World: A Multivariate Poison Process Model of Competitive Performance Implications of User Generated Content”. And of course, I’ll suffer at 4:35 when poor Ghosh and the Park’s present “Modeling Member Behaviors in User Generated Content Sites: A Semiparametric Bayesian Approach”.

Not exactly easy material to absorb or present at 4:35pm.

Certain titles are deceptively accessible. “The Elites vs. the Crowds: How Third Party Experts Influence Consumer Opinions in the Marketplace” by Yubo Chen et al is one such case. Heavy Markov Chaining isn’t evident in the title, however – Yubo Chen – the lead author – is well regarded in the community. He has journals in the top tier such as JAM and MS. You bet he’ll bring the math.

Buried in Saturday morning are some potential diamonds. “Endogenous Homophily in Social Networks” by Katona might prove to have something original to contribute – though, homophily is a topic that goes back to the 1960’s (at least) in the MS literature. It might be old wine in new bottles. At the same time that morning, Hu and Hui are presenting on “Monitoring Brand Perceptions using Jump Diffusion: a Bayesian State-space modeling Approach”. I reach for my wallet whenever I hear the words ‘JUMP’ and ‘BUMP’. Those algorithms don’t exactly scale. They both present at 8:30am on a Saturday…and are equally interesting and relevant from their own perspectives.

You just never know what is going to be really great and what will be less than applicable. So at a conference with 15 tracks, 12 sessions each, with 180 individual sessions, and around 600 presentations – it’s a role of the dice. Just by showing up at any one of the 15, there’s an opportunity cost of 14. In the end, if I’m lucky, I’ll walk away with 50 unique presentations. And I’ll have seen a fraction of what’s on offer.

I’ve gone on well too long about this – but that’s the flavor of the conference.

I’m bringing this back to where I started with this post – back to Morbo.

Every community has their own language and jargon. Underpinned in language is literally how that community interprets the world and stresses what is important and what isn’t. I head into that world – where even the smallest nuance of how something was sampled can draw blood – and emerge on the other side a lot smarter and inspired. But also a lot more stabby. There’s the world where you talk openly about whether or not a viral-branching model of WOM is really the best way – and then there is the world of reasonable assumptions and tipping-point logic.

Then there’s the world where it’s perfectly reasonable to make a Futurama reference in the context of advanced social analytics.

They are what they are.

It’s just a matter of chilling out and patience.

Windmills really don’t work to cool anybody down.

GOOD NIGHT!