Are you being served

Image credit: Jay Cuthrell

Are you being served ( 2010 )

#include <disclaimer.h>

[and was appreciative of the invite.

The photos I took don’t justice to the demo presented. One demo was a screen named “Sentiment by Media Source” and it was a standout for me. Here’s why… [Just like there is a heat map for most solid analytics packages – the ability to gauge sentiment overall is a critically important tool to have in larger organizations. More importantly, the SAS demo focused on the well worn paths and pairing the tool with a SAS professional service is spot on.

Granted, you have to tune SAS SMA as you would any tool. For those expecting “AI”, that’s just not a realistic way to position this new classification of toolset.


Okay, back to the “Sentiment by Media Source” screen. This demo was of a fictional brand/company of course but it was real data. What did it show you?

“How you are perceived on Huffington Post vs. Twitter vs. Crooks and Liars”.

Think about what that means for a moment.

Yes, there is an audience to care about. And possibly ones not to care as much about… perhaps?

The Teragram acquisition made by SAS is paying dividends in this platform on the textual analysis side.

The SAS Media Workbench was shown and one could classify minutia such as wicked.

Example: wicked

Wicked Good vs. Wicked Ways vs… uh… WICKED WICKED

While the tools provided cater to the larger enterprise entity, there is going to be feedback loop and learning in the tools (SaaS has that unique property) that filter down to the SME and possibly SMB markets. Before that takes place, an agency model will be explored as a competitive differentiator.

This is a SaaS play from SAS. Say that three times really fast.

The data sets come from SAS’s own web/api crawlers normally but this was driven with data from a third-party data aggregator for the purpose of the demo. Yet, my take away is that this SaaS approach will lend to flexibility outside of a purely SAS vertical integration.

The demo shown didn’t cover location based services or geocoding analysis but there is a call center email by geographic location analysis being done. As I think in telecom terms (often) for new technology, this is welcome news.

The questions from the crowd were ones you might expect:

When can I play with this? What about internationalization? Who is using this now? What is the algorithm for “influential”?

My real question wasn’t how the SAS platform is digesting the social data but what will enable using the SAS tool in a social capacity? That’s where I see companies like Live Math (best pitch I got at SxSW 2010) coming into play.