In March of 2011 Piers Morgan was flying to Minneapolis for AGT auditions. The plane didn’t land in Minneapolis but turned back and returned to New York. After few hours of trying to get to Minneapolis, Piers wrote in his Twitter “Be ashamed Delta. This is shocking even by your standards. And you can all relax at Delta now because the gates of Hades will freezeth over before I darken your pitifully incompetent doors again.”
That’s right, he documented the entire ordeal on Twitter generating 46 tweets over the course of few hours. Back in 2011, Piers Morgan had almost half a million followers (more than 4 millions today) and 46 tweets about Delta flight disaster were retweeted 700 times. It is difficult to under-estimate the impact of such a feedback, real-time, on the airline’s reputation. Should Delta had a system with text analytics capabilities in place that would’ve allowed the airline to monitor social media and identify messages with negative sentiment and messages that can be classified as describing potential issues, generating alerts and notifications, the PR disaster could have been avoided or its impact minimized by the timely and appropriate response.
I remembered this story from 2011 as I was watching online broadcasting of a closing General Session of IBM Insight 2014 conference on Wednesday, October 29th last week. Chris Moody, VP of Data Strategy of Twitter on stage, started his speech with: “Twitter data has unlimited value and near limitless application… We know what the world is thinking at any point in time around any particular topic… In the future any significant business decision will have Twitter data as an input” and went on explaining how Twitter data can help in any aspect of life and business, from medical research to finding out if the commercial fryer machine needs maintenance (the one usually installed in the places like McDonalds). Apparently, when people complain on Twitter about “soggy fries” it is a single biggest indicator that the machine is malfunctioning and needs maintenance even though the operator of the machine does not see the problem. So if you are a manufacturer of some commercial equipment and think that Twitter cannot help you to get customer feedback, it is time to hire a data scientist and draft your analytics strategy.
If you are a manufacturer of a specific gadget, you can learn from Twitter what the customers who bought your device think about its features, what exactly they don’t like, what features they wish the device would have, you can even incorporate analysis of Twitter data into your production planning and inventory management.
What’s important – all feedback is real-time. Twitter data can be used in the decision-making and business analysis related to customer service, supply chain, brand management and product development.
On that note, IBM and Twitter announced a new partnership, allowing IBM customers to take advantage of Twitter data through a set of IBM solutions and tools:
- Twitter data will be integrated into IBM Cloud analytics tools, such as Watson Analytics and Bluemix platform-as-a-service which now includes Watson Developer Cloud
- Twitter and IBM will work on a set of solutions for the enterprises, the most obvious and first on the list to start is integrating Twitter data into the IBM ExperienceOne customer engagement solution to help marketing professionals to hear the voice of a customer
- 10 thousand IBM consultants will be trained and certified to use twitter data in analytics solutions.
This new partnership with IBM was made possible by the recent Twitter acquisition of Gnip in April of this year. With Gnip capabilities,Twitter can now stream its data through the “Twitter firehose” provided by the Gnip platform to other applications, including IBM Cloud analytics, to mine insights from the data, analyzing half a billion tweets a day.
With Watson being part of the system, a customer should be able to ask the system “Do people like Moto 360?”, and get back not only the results of the sentiment analysis of what the world thinks about Android Wear Smartwatch in general, positive vs negative feedback, but a break-down of what features most of the owners of the gadget like and what they don’t like, and where customers who like the watch live.
The benefits of the partnership and the inclusion of Twitter data as one of the inputs into the variety of analytics applications is obvious and I am looking forward to learning about the solutions that IBM is going to build with the use of Twitter data.
There are still some questions that I will be looking to answer though. For example, as ECM professional, I want to know what is the retention policy for Twitter data, how long the data is going to be stored in IBM cloud data stores and how far back in time we are looking for historical tweets? Gnip platform provides “realtime data as well as access to every publicly available Tweet dating back to the very first Tweet from March 21, 2006”. Historical data is important for predictive analytics and trend analysis.
I would also like to see a clear explanation of how the applications that use Twitter data will be managing retweets and how/what value will be derived out of retweets. Back to the Piers Morgan example in the beginning of the post, – 46 tweets were retweeted 700 times, simply based on the popularity of the author of the original tweets. For Delta, it does not mean that 700 passengers didn’t like Delta services, it simply demonstrates the extend of PR damage. In the other example, if I went out for a lunch with a friend and we both really didn’t like soggy fries we both had, I can simply retweet my friend’s message about the fries, confirming that I hated those fries too. Would be interesting to see how IBM will design handling of the retweets in the solutions analyzing Twitter data along with other data sources.
Meanwhile, the news about Twitter/IBM partnership were retweeted hundreds of times.
Leave a Reply