For IBM's Watson, and Social Media Analysis, Human Judgment is the Name of the Game

IBM has done an amazing thing. It’s engineered a massively powerful computer system (hardware and software) that ‘guesses’ so well it appears to be nearly human in its ability to process information. This is no small feat, but consider what it took.
Estimates indicate IBM spent 20 to 40 million dollars or more and leveraged the efforts of 25 top scientists over the course of four years to build Watson. All of this makes Watson very, very good at doing one very specific thing – playing Jeopardy.
Though Watson is great at Jeopardy, IBM worked even harder to make the system good at something far more grand than a TV game show. The true power behind Watson’s success, and the thing that’s most interesting to us here at Crimson Hexagon, is its seminal ability to approximate human judgment. In playing Jeopardy, its great successes as well as its rather comic failures all tie back to the ability to make judgment calls when interpreting information. Imbuing Watson with even a modicum of this most human ability was central to its success. There are some great parallels here between Watsonís abilities and what it takes to be truly successful when analyzing social media.
Social Media Analysis Powered by Human Judgment
Watson is a supercomputer that quickly processes massive amounts of information to make judgment calls and provide answers. Its amazing abilities (amazing for a ‘dumb’ computer) come from four years of 25 scientists teaching it how to ‘think’, and some of the most powerful and expensive computer hardware available. That said, it still made some entertaining mistakes. For example, Jeopardy question: ‘What do grasshoppers eat?’ Watson’s answer: ‘Kosher’.
This challenge – how to accurately process lots of information, decide where it fits and what it means in a given context – is strikingly similar for systems designed to analyze social media data. Most social media monitoring and analysis platforms approach the problem in much the way Watson does, but with the distinct disadvantage of substantially less applied intelligence and certainly far less raw computing power.
Their real problem is that they don’t have the benefit of being able to apply real human judgment and contextual intelligence to each and every unique analysis. This leads to significantly reduced accuracy and often a complete inability to distill social data into relevant, meaningful information that can be used to inform strategic business decisions.
In most instances only humans can do this kind of work accurately. To address this challenge you need to find a way to combine human judgment with computer speed and scale. This ‘human-infused-machine-powered’ process is at the core of the way Crimson Hexagon works.
Putting the Man-in-the-Machine
The key to making sense of this unstructured social content is to be able to apply a relevant analytical context to Twitter, Facebook and other social media posts. Crimson Hexagon doesn’t focus on basic mention-counting and simplistic measures of positive and negative, nor does it depend on a roomful of human analysts to pour over thousands of posts attempting to find patterns in the data. Instead, Crimson Hexagon’s patent-pending technology, invented at Harvard by University Professor and Crimson Hexagon co-founder Gary King, provides the best of both worlds. It takes the nuance of human judgment and applies it directly into the scale of automated computer analysis. The software allows a single human operator to ‘train’ the system using their uniquely human power of interpretation and judgment as applied to the specific context of each analysis.
In essence, the human operator transfers their judgment and intelligence to the system so that the high-speed analysis can occur as if that human’s brain were analyzing thousands or millions of social media posts rather than a ‘dumb’ computer guessing what to do with data based on generic definitions scripted into its operating instructions. Crimson Hexagon provides computer automation that’s driven and controlled by human judgment.
This human-judgment-powered process is mechanically simple but statistically accurate (97 percent accurate), freeing customers to focus their energy on harnessing insights rather than wrestling with data and technology in an attempt to simply get insights.
Unlike most other systems for social media analysis (and even the approach used for Watson), Crimson Hexagon doesn’t force a choice between human judgment and computer automation. It combines the two and delivers a uniquely powerful way to derive meaning and business value from social media data.
Commenting recently in an Associated Press story about Watson’s Jeopardy playing talents, James Hendler, a professor of computer and cognitive science at Rensselaer Polytechnic noted that, “A human working with Watson can get a better answer. Using what humans are good at and what Watson is good at, together we can build systems that solve problems that neither of us can solve alone.”
We couldn’t agree more.
What are your thoughts on this matter? We’d love to hear your perspective.
As always, we’ll continue analyzing the social sphere for interesting trends on a range of topics. In the meantime, feel free contact us at with questions or to learn how your market, brand or product could benefit from leveraging the Crimson Hexagon ForSight™ platform for social media monitoring and analysis. Want to see the Crimson Hexagon platform in action? Request a Live Online Demo.

Request a Demo

Ready to transform your business?

Get a walkthrough of Crimson Hexagon and learn how consumer insights can help you make better business decisions.