NEWS28 October 2011

HP tests predictive capabilities of social media data

Data analytics North America

US— Hewlett-Packard (HP) has announced promising results for a pilot programme that combined social media and company-held data in a bid to predict customer behaviour.

Piloted by HP customers in the media and entertainment industries, Project Fusion combined two sets of data: unstructured (Amazon reviews, customer surveys, support logs and other natural-language text) and structured data (customer support tickets, sales transactions and customer demographics).

HP Labs developed new analytics technology to convert the unstructured data to a structured format, which allowed human analysts to use standard data-mining and statistical tools to analyse the two sets together.

In one case, HP said ‘social signals’ were found to predict customer support tickets with 90% accuracy, and in another case different social signals were shown to be “highly correlated” to sales.

HP’s VP of customer intelligence Prasanna Dhore said: “The results are significant enough to act on. Applying this methodology can improve both the customer experience and the bottom line.”

Read more at the HP Data Central blog.