HR TOOLS & TECHNOLOGY
Sentiment Analysis, the new hero in town!
Sentiment-analysis software can help companies figure out what’s bothering workers—or what they’re excited about. The term ‘sentiment analysis,’ is selfexplanatory, but for the sake of it, we’ll list a definition we found on Google - “the process of computationally identifying and categorizing opinions expressed in a piece of text, especially in order to determine whether the writer’s attitude towards a particular topic, product, etc. is positive, negative, or neutral.”
Every day, humans, collectively, type out over 200 billion emails, 500 million of tweets, and hundreds of millions of texts, chats, and private messages. It’s practically impossible for one single person to stitch together these data, to figure out emotional trends or behavioral themes. This is why we have computers. For decades, researchers have been developing computer programs which can try and understand the emotions stirred up by an idea or a product from our writing. This particular field, is known as ‘sentiment analysis.’ As a matter of fact, it’s pretty popular in the world of marketing, and is commonly referred to as ‘opinion mining’. It refers to the analysis of one’s feeling (i.e. emotions, opinions, and attitudes) behind the big blur of words using language processing tools. The idea is to use computers to look beyond the veneer of construed words – are you positive, sarcastic, negative, or biased?
So, what does sentiment analysis have to do with employers and their business? You see, it all goes back to the mid2000s, when companies wanted to understand how people respond to their products, or their competitors offerings. Algorithms were being used to aggregate reviews to reveal broader insights than surveys or focus groups. This grew to the point that dozens of startups are now exclusively offering these sentiment analysis softwares to let them know how their own employees feel about their jobs.
Large corporations like IBM, Twitter, Intel, and Accenture have started instituting the software to understand how their workers feel about their jobs. The aim here is to identify problems that might easily escape a supervisor during the annual performance review.
Earlier this year, IBM started using sentiment analysis software to better retain employees in the competitive job market. The software uses languageprocessing and machine-learning
algorithms to decipher emotions from text found in open-ended questions on company surveys, comments on company blogs, and internal social networking sites.
Intel uses a similar software from Kanjoya Inc., to better understand employee frustration. The software turned out to be pretty insightful, as well. It revealed how a majority of the employees had the wrong impression that their own jobs were at risk.
Such cases highlight how imperative is it for companies to rely on such technologies. According to CIO Journal columnist Irving Wladawsky-berger, in today’s world, where workplace collaboration is the key and where technology-empowered customers can easily share what they think about a product or service, empathy is the competitive edge.
Not so long ago, Twitter hired Kanjoya, to analyze workers’ responses to company surveys about their workplace experiences. The surveys were administered twice yearly, and included only two open-ended questions. After hiring Kanjoya, Twitter started sending the survey to one-sixth of its workers every month – it also increased the number of open-ended questions. The patterns extracted from Kanjoya’s analysis platform were then shared with the executives.
Kanjoya, also advertises that its sentiment analysis tools work with Yammer, a social network acquired by Microsoft for a billion dollars. Some of Kanjoya’s product offerings include employee engagement tracking (to trace positive or negative emotions), and a search function which responds to queries which an analysis of the surrounding sentiment. A lot of companies today are more focused on analyzing employee chatter outside of the formal reviews or surveys. Now, this makes it difficult to scoop a structured response, or identify behavioral themes. IBM has for years, been collecting employees’ posts and comments on its internal social networking platform.
Called Connections, it’s available to all of IBM’S 400,000 employees worldwide. It functions like a mélange of Facebook, Wikipedia, and Dropbox, allowing employees to publish posts, comment on others’, and collaborate with one another on certain projects. IBM also sells a version of the platform, Connections to other companies. The social networking platform is integrated with a sentiment analysis tool called Social Pulse, which allows IBM to monitor posts and comments for behavioral trends and red flags.
In 2015, IBM used Social Pulse to revamp its performance-review system. Its HR department relapsed the old feedback system to create a new one in order to receive genuine responses. IBM used Social Pulse to comb through the hundreds of thousands of feedback received.
The software narrated an entirely new story: Employees at IBM were unhappy that their performances were actually graded on a curve. Within a month or two, the company introduced a new and improved method.
By widening the scope of the data accumulated via surveys, reviews, and social media posts, there’s a certain risk of violating employees’ privacy. This is why IBM limits the data-mining to posts and comments that are shared with the entire company. It bars emails, chats, or interactions in private groups.
One cannot entirely rely on sentiment analysis for a polished report. You see, the computer’s understanding of natural language, is still in its infancy. According to a research project (2016), basic analysis tools sent between developers and an open-source server software suite only had a maximum accuracy rate of 30 percent. However, when two people tried to determine the emotions expressed in 50 emails, they could only agree on three-quarters of them. These algorithms still lack the human elements, empathy. A small team of analysts routinely examine IBM’S Social Pulse, to ensure they’re sending the right trends to the management.
We’re still a long way to go, if we really want to improve our ability to understand how our employees feel. A group of computer scientists in India published a paper which suggests a new way of determining employees’ well-being: facial scans. The system utilizes images captured of employees’ faces each time they enter the building, to determine whether they’re happy, depressed, sad, or angry. The facial scans might someday help business use the date to optimize productivity and profits. Until then, there are plenty of emotion-detection technologies in the market, and to be honest, it’s not a bad start.