Natasha, your comment reminded me of an interesting discussion I had with some researchers at the Leonard Stern School of Business. They had conducted a study of eBay and Amazon in order to link unstructured data (i.e. comments) to price. What they found was that sentiment analysis is a good predictor of sales. Using only the listed price they could predict sales with 55% accuracy; adding star ratings boosted accuracy to 74%. However, when sentiment data from the comments was included in the prediction variables, the accuracies rose to 90%. As you mention, companies could (and should) pull this unstructured data from anywhere online – it doesn’t have to be limited to comments made on the corporate site.

I think there is also some value of putting Social CRM into the context of a particular service instance or a transaction. For example, when a few years ago Amazon started suggesting books, it was when customer went to their site and was ready to buy. Already, you can read other people’s reviews, which are unstructured. Only number of stars and date of post are structured.

When I travelled to Europe this past summer, I booked everything no the internet, and my selection was entirely based on reading what other people said in unstructured text.

Of course, these booking channels made it very easy for customers to go to their site and write reviews there. With their switch to Social CRM, they should just write these reviews in Facebook, and booking channels should consume it. This is really about connecting the data points.

As for analytics, there are already fairly sophisticated tools out there that allow to analyse unstructred text, and make some sense of it. Again, connecting the data points, and someone analysing that text from Facebook, is a change that would need to happen.

In my opinion, the technology challenge is connecting of the data point between existing CRM and customer analytics tools with Social Networking sites, and perhaps more importanly, dealing with permissions, and data privacy issues.

The “newly” learned information will require newly designed processes to produce ROI on all these efforts. I am aware of Radian6 and quite a few other companies, including mine (www.amplifiedanalytics.com) that are working to convert unstructured data into meaningful information, and then into actionable knowledge. We all making some significant strides, however the business processes and methodologies for it’s use, are still in infancy. And that is my concern.

You are absolutely “on the money” with a “single version of truth” challenge. It is not technically all that challenging, I’ve been involved into a couple of projects focused on this, but an organizational nightmare. Yet, this is a source of the best and fastest ROI I have seen.

I think another major challenge is achieving a “single version of the truth” across the enterprise. You mention translating knowledge into measurable action which is the tough part. In order to achieve this you need to gather data inputs from throughout the enterprise and share the results with different departments. I think a lot of organizations are still very silod where the contact center does one thing, corporate communications does another, product development is disconnected from both, and the analytics division operates as a black box that gathers and shares information selectively. The lack of comprehensive visibility into the customer experience is definitely a problem.

There is a potential for tremendous benefits and cost savings for Marketing, but scalability, transformation of data into knowledge, and new processes for translating this knowledge into measurable actions, still need to take place.
Your examples of corporations adopting SM channels, while sexy and newsworthy, may prove to be uneconomical in the long run as a Customer Service operation mechanism, unless the automation of these processes and work-flows, can be automated.