A text analytics study by Seth Grimes

Seth Grimes recently conducted a text analytics study that included vendor and consultant experiences deploying voice of the customer (VOC) solutions. Seth provides great insight into the things to consider when assessing these solutions. He has some really good charts from a survey he conducted on VOC and some lessons learned from early deployment experiences.
 

Seth found in his study that when evaluating the use of text analytics, it is important to focus in an area where value can be created quickly.  An initial pilot is critical to establishing this value, with clearly defined goals focused on strategic objectives. A common mistake is going after the biggest problem first. This is not conducive to winning fast and winning often. Proven value is a predecessor to putting the solution in front of executives. Successful providers in this space have established an internal champion and ensured that the business was ready to take action on the data. Failure to do so puts a pilot at risk of never moving to production.
 

Once value is proven, these solutions have become viral in that they expand to address issues in other areas of the business.  From a deployment perspective, collecting requirements from all potential internal customers ensures that deliverables meet the need. On the front end, enabling the analyst is critical due to the time and effort it takes to develop dictionaries. Scope must be managed aggressively and both structured and unstructured data should be addressed. These findings are consistent with our experiences. Value, once proven, is the springboard to broader use of text analytics. I'd be interested in the experiences of others.