Connect with Us

About Us

The Decision Factor offers insightful comments and observations on analytics—from views on new technology approaches and market dynamics to the latest industry trends driving demand for faster, smarter information analysis. This blog contains personal views, thoughts, and opinions from SAP employees, mentors, and friends working in the area of analytics. It’s not endorsed by SAP nor does it constitute an official communication of SAP.


It’s the Wild West of Data Out There — But Is Unstructured Data Worth Lassoing?

Twitter feeds and Facebook status updates—some of us see and interact with them every day. If you’re in business, do you ever find yourself wanting to take those thousands of free-form, unstructured little nuggets of sentiment and opinions about your event, competition, or product launch (think about the tweet storm that goes on when Apple releases a product) and aggregate, analyze, and use them to influence your next move? Would harnessing all that information make your business better?

Lassoing Unstructured Data: The Cool Factor

I don’t know. But the “cool and interesting” factor is way up there when you see a Twitter analysis of an event or product launch. Most recently, I saw it done with SAP’s May 2011 SAPPHIRE NOW event .

We aggregated the Twitter feeds on a dashboard and saw who was making the most noise about SAPPHIRE by handle. We explored what people were talking about the most by keyword. And we saw how they felt about the event by looking at the positive and negative sentiment parsed out and categorized as such.

After viewing the results, I did get a good rearview-mirror-kind-of-shot, but I didn’t know how I would or could use this analysis in tactical future planning.

Lassoing Unstructured Data: Business Use Cases

With the ability to analyze unstructured content being the “cool” factor, I asked myself what would be a more tactical business use case for lassoing the power of unstructured data today? Customer support call logs, maintenance logs, and product reviews (think about your favorite ecommerce retail site that lets you see the customer reviews) come to mind.

If I were able to access, wrangle, and sort through the noise in all that unstructured data (free form field content usually contains a lot of jargon and shorthand) and turn into something structured so I could automatically identify what the content is about (did you hear a ‘poof’ sound when you read that?), would I be a better decision maker?

I would if it could help me:

  • Detect and parse out key concepts about my product
  • Identify phrases commonly used when talking about my product
  • Understand sentiments and other items of interest about my product

If you could extend your data view into unstructured text sources, would it give you significant insights into your business or would you put it in the “cool and interesting” category?