Tech Insights: When HR Approaches Data the Way Marketing Does

May 6, 2014

HR analytics usually looks at numbers, but unfortunately, a lot of relevant information is not numeric.

To understand textual data, our friends in marketing use sentiment analysis tools like Sysomos and Meltwater. These give them a handle on what social media is saying about their company and products.

The tools do exactly what the name suggests: they analyze the sentiment of a comment — is it positive, negative, or neutral?

Yes, this IS relevant to HR

While sentiment analysis is a sophisticated use of computers, it is not quite the dark art of artificial intelligence you might suspect. In fact, it is more a matter of looking for positive words such as “great” and negative words such as “terrible.” The tools have some understanding of context and are remarkably good at dealing with the awful spelling and abbreviations that suffuse social media.

Is this relevant to HR? Some employee survey vendors, like IBM Kenexa and Towers Watson, offer sentiment analysis of employee comments. HR should expect that leadership will no longer be satisfied with a document of collected comments; they will want a formal analysis summarizing what the comments mean.

Sentiment analysis can also pick out themes, generate word clouds, and create “buzz graphs” (which draw thin and thick lines between keywords to show how they are connected.) One of the cooler tricks is being able to summarize the content of a long comment in meaningful sentences.

What is interesting

  • Analytics just keep rolling along and spreading to new places; love it or loath it, but we have to master it.
  • Recognize that marketing is a great role model for HR in analytics. In fact Freshbook’s Chief HR Officer Dave Wexler says HR should stop modeling itself after finance and look to marketing for inspiration.

Where the real value is

  • Of all the things that employees write, what percentage of that are comments on an annual survey? Maybe using sentiment analysis on survey comments is just the tiniest tip of the iceberg.

Imagine what will happen when we have analytical insight into all the documents, reports, emails and text messages employees write. Will this be a good thing?