The following are my notes on a presentation from James Newswanger (IBM) called “IBM and Twitter: What’s In a Tweet?”. This was live-blogged, so there will be typos.
IBM has become very proficient in social media analytics, this began with text analytics and then with Twitter.
More than once, a single Tweet has created billions of dollars worth of influence on financial markets.
Sept 21, 2015 Hillary Clinton posted a tweet about price gouging on Twitter, pharma index dropped 5% in 3 hours.
Late October 2014 IBM and Twitter announced a partnership where IBM could have access to all tweets – a unique relationship.
What IBM is Doing
- The Social Silo – clients know social listening, mentions, brand management
- The Gap – social monitoring vendors focus on social as a silo
- The Opportunity – rarely is info merged across: internal and external data, structured and unstructured, without data burden
Have decided to focus on Twitter, using Watson analytics.
Every single tweet holds 150 unique pieces of metadata that can be extracted through access to Twitter’s API.
A lot of junk in tweets, need to clean.
- any account with no followers is generally spam
- IBM looks at Klout, volume of tweets, followers, post history, referrals, retweets
- Have not come up with one algorithm to create one weighted score
A global retailer uses the combination of internal and real-time public data, like weather, competitors promotions, Twitter feeds, economic data and news to identify strong, even counterintuitive demand signals
Result – company reoriented its massive supply chain to deliver merchandise based on real-time forecasts
Listening Value Adds
Influencer analysis – used to identify influential people for sponsorships or advocates for company/brand/cause
Microsegmentation – use unstructured SM data to detect life events and refine segmentation — ex. “Told my mother I got a new job today” (life events)
Watson Personality Insights – build new segments based on an understanding inherent personality traits to grasp attitudes and traits.
- Convert the way your organization thinks about social platforms
- A tweet is >140 characters; look behind the tweet to undertsand what’s hidden in the meta data
- Leverage social media as a leading indicator and value add appliances
- [Sorry missed last one]