IBM is extending its partnership with Twitter to include the social media site’s data in its Watson Analytics service, delivered in the cloud. It is also making tweets available in its Bluemix cloud developer environment.
IBM Analytics general manager Alistair Rennie (pictured) unveiled the service at CeBit, the business IT event held in Hannover.
The IT giant first announced its partnership with Twitter in October 2014, with Rennie claiming the deal would give IBM “unprecedented access to the Twitter database, going back to 2009”.
IBM, he said at the time, has trained 10,000 of its consultants to deliver business insight to clients, using the Twitter data access, while Twitter has named IBM as its “only global consulting partner”.
“We’ve had 150 client engagements so far, which we think is incredible. And they cross use cases and industries," said Rennie.
"It’s not only marketing doing sentiment analysis. It’s in demand signalling, the supply chain and customer churn models. One example is how employee retention has an impact on retail sales.”
In one group of cases, the supplier’s analytic models revealed that customers miss the relationships they have built up with sales staff, and have said so on Twitter.
IBM looked at that sentiment along with loyalty information and the financial performance of different stores and restaurants. Not only did dissatisfaction with employee turnover impact sales negatively, the dissatisfaction was most keenly felt by the most loyal and valuable customers.
Rennie said that UK customers who have taken up the Twitter data service are drawn from financial services, insurance, consumer events, telecommunications, utilities and sporting events.
Read more about IBM Watson
- Find out how Southampton University is using IBM Watson analytics
- When IBM's Watson analytics system bested 'Jeopardy!' champs in 2011, the world cheered. Now IBM must up its game to make Watson a commercial success, writes Jack Vaughan
- What is the IBM Watson supercomputer?
The Twitter data is now integrated into the supplier’s Bluemix cloud, which is a developer environment, as well Watson Analytics, a cloud-based predictive analytics and visualisation toolset made available in December 2014.
“The skillset needed to use that is similar to what you’d need for a spreadsheet, but it delivers very powerful analytics," said Rennie.
“People are not approaching this saying we’ll take two years and have a long strategy. They want to do this quickly, get an insight, make a decision. And it is not expensive – Watson Analytics is a freemium tool.
"The Twitter service is $30 per month. With Bluemix the developer can access up to five million tweets for free, and once their app is in production it is pay as you go.”
According to Rennie, this is a way for CIOS to give their business users the ability to do analytics easily on their “unlimited questions”, and conserve their data scientists for the right class of problems.
In client engagements, the Twitter data is being combined with other data sources. IBM has identified correlations between weather events, angry tweets and customer defections. The supplier said it can improve churn rates by up to 5% by analysing localised Twitter data in combination with weather data.
The supplier has also found, from the initial cohort of engagements, that Twitter is a valuable indicator of demand for the clothing industry.
According to IBM, by using psycholinguistic analytics from IBM Research to extract psychological, cognitive and social traits from Twitter data generated by influential fashion bloggers – combined with operational data, such as sales and market share information – manufacturers can better understand why some products sell well while others don't.