HP Discover 2012: Analytics provides 10x return on investment, claims HP SVP

News 4 Dec, 2012

Hardware giant claims businesses can no longer afford to throw away unstructured data.

HP has used the second day of its Discover conference to talk up the importance of analytics, claiming that every $1 businesses invest in data mining tools will generate at least $10.66 in return.

"Throwing away [unstructured] data is like throwing away money.[HP has] every way you need to take information, keep it around forever and monetise that information," Colin Mahony, SVP and GM of HP Vertica, told attendees.

By 2015, information is expected to nearly double (48 per cent) and 90 per cent of this will come in unstructured form via videos, tweets, Facebook posts and emails.

But HP is adamant that using packages from subsidiaries such as Autonomy and Vertica can generate huge benefits to enterprises. It is possible to increase revenue, improve efficiency and to provide better value for customers, Mahony claimed.

The SVP gave a couple of examples of the particular verticals that can benefit from analysing unstructured data.

A large telco would be able to sift through customer details to see how many customers are using a particular phone with less than two months remaining on their contract, for example. This is important information as the business can take steps to retain those customers.

However, HP can take the queries even further. The telco would be able search to see the experience of these customers and for a query to see if any had dropped calls in the past month.

They would also be able to gauge if customers have had negative interactions with the company, through chat, emails or phonecalls.

“Bring all that together and you can run queries in analytics and predictive models and can make a dramatic change to the churn rate. For a telco, the ROI is much greater than $1 to $10 dollars.”

The healthcare sector is another industry which can benefit. A healthcare insurance company used to take 3-6 months to detect fraud, Mahony noted.

However, it is now possible to analyse and extract relevant information from data in less than three minutes, he added.