Inhi Cho Suh – VP and GM of Big Data, Integration, and Governance at IBM Live With John Furrier

IBM’s landmark alliance with Twitter, Inc. is a symbol of the changes that are sweeping through the IT industry, transforming both the role of technology and who uses it inside corporations. The deal also shows how important IBM’s Watson question answering technology is to the company’s future.

Under the terms of the partnership, which was announced today, IBM will build analytics products and services around Twitter data and offer them to customers in both functional and industry vertical roles. The first fruits of the alliance are likely to come in the form of IBM Watson Analytics, a freemium service that IBM announced late last month. By folding Twitter data into the analytics engine, IBM will be able to offer modeling services that marketers and customer support organizations will probably devour.

IBM also plans to deliver enterprise applications for horizontal functions like customer support, sales and marketing, and it will train tens of thousands of its consultants to provide services to industry verticals that incorporate Twitter data. But the big news is Watson.

The announcement highlights the growing importance of unstructured data to corporate data management. As the success of technologies like Hadoop and NoSQL has demonstrated, businesses are reaching the limits of incremental value they can wring out of transactional information. The bigger opportunity lies in making sense out of data that doesn’t fit in a spreadsheet, and there’s no better embodiment of that kind of information than Twitter.

Twitter currently manages about 500 million tweets per day in 48 languages. Marketers and customer support pros are drawn to it because its off-the-cuff nature tends to produce more honest feedback than that which comes from surveys and focus groups. Those same groups also fear Twitter because negative sentiment can spread so quickly through retweets. For both reasons, they’re keen to better understand what people really mean when they tweet.

Technology entrepreneurs have been trying to analyze sentiment as expressed in social media posts since the early days of Facebook, but the task is devilishly difficult. Not only do words have different meanings in different languages, but dialects, slang, sarcasm and humor can be almost impossible for a machine to interpret.

IBM is betting that Watson can crack the code. The computer that thumped the top two all-time Jeopardy! champions in an exhibition more than three years ago is tuned to understand language and also to derive meaning from poring over massive amounts of unstructured text. Watson is essentially the opposite of a search engine. Instead of the user asking the computer for information, the computer examines the information  and tells the user what to ask.

If IBM can successfully unravel the mystery of sentiment analysis, it will have an public victory that will capture the minds of marketers and media alike. That would go a long way toward reaching IBM CEO Ginni Rometty’s reported goal of making Watson a $50 billion business by 2020. The deal also gives Twitter the opportunity to drive up its value proposition to marketers. The company has struggled to tell a convincing growth story to investors since its IPO last year. Tapping into marketing budgets is a tantalizing way to diversify beyond advertising.

It would also implant IBM with a group it actively covets: marketers. IBM is taking seriously Gartner’s forecast that chief marketing officers will outspend CIOs on technology within two years, and it has actively courted that group through research, events and products. The Twitter initiative shows how Big Data technology explodes traditional lines of IT responsibility.

Inhi Cho Suh, VP/GM Big Data, Integration, & Governance at IBM, explained the value of Twitter for engagement analysis in an interview at IBM Insight 2014 earlier this week. See the video below (19:17).

Video to social data:  https://www.youtube.com/watch?v=JANdLvJh84M&list=UUu3Ri8DI1RQLdVtU12uIp1Q#t=242

Full disclosure: The author consults with IBM’s midsize business group.