In the past few years, savvy companies have been using text analytics
software to analyze positive and negative phrases appearing in social media and other electronic posts to figure out what customers think about their products, service
and policies. For example, firms in the travel industry can monitor comments posted on websites for users’ opinions about hotel properties, or
perhaps about specific features such as housekeeping or food service.
Less well-known but a growing reality is the use of this “sentiment
analysis” technology by companies to gather information about their own employees. Vendors of text analytic software confirm that clients are
now using these tools to gauge employee sentiment, although users aren’t yet willing to talk publicly about internal applications.
Why are companies venturing into this area, which at first blush might seem fraught with privacy concerns?
Reasons for mining employee communications vary. One of the most basic is dealing with litigation. In a court case where thousands or even
millions of emails must be combed for threads of information, the power of analytic software can help make an onerous task less burdensome.
Offering greater promise may be more broad-based applications such as mining emails, intranets, surveys, internal portals and performance review
systems to gain a clearer picture of employee sentiment and engagement. ( Employee engagement is the extent to which an employee is enthusiastic about his or her work and committed to furthering
the organization’s goals.)
“Any textual form of expression can be analyzed,” says Elizabeth Charnock, CEO of Cataphora, a Menlo Park, Calif., firm providing analytics
software, “including public forums like Yahoo Finance.”
For instance, a routine review of emails might reveal that among staff assigned to a specific project, the number of negative words or phrases
has increased significantly over a three-week period. That might alert managers to revisit the project’s status and spend more time communicating
with key staff and addressing their concerns.
Gauging Employee Satisfaction
If used judiciously, this kind of analysis may provide a more effective tool for assessing key factors such as job satisfaction than internal surveys
or other traditional methods. In mining unstructured data, HR leaders and management teams can obtain insights on the degree to which employees
are engaged. By gaining a clearer picture of employee sentiment, companies may identify areas where employees are dissatisfied and design
strategies for enhancing engagement and in turn, improving productivity, employee retention and customer service.
Clarabridge, a Reston, Va.-based provider of text analytics software,
notes in a white paper that the happier employees are in their jobs, the more likely they are to engage positively with management, peers and
Vendors point out that once the software used to analyze employee sentiment is in place, it automatically gathers key data and prepares
reports without taking up staff time. In eliminating manual processes and removing dependence on structured surveys, the software increases
efficiency while providing useful info on a routine and dependable basis. In the process, insights can be drawn from a variety of complex data across
the entire enterprise, including multiple locations and different languages.
In addition to fostering employee engagement, the insights gained from front-line employees who have direct contact with customers can be
used in revising sales procedures or developing training programs.
Jeff Catlin, CEO of Lexalytics, a provider of text and sentiment analysis software based in Amherst, Mass., notes that while some companies
monitor employee communications on external sites such as Facebook, few take a “watchdog” role in this process. The ultimate purpose is not to
keep tabs on individuals, but rather to assess attitudes reflected by multiple employees, and in turn use the data in promoting positive
The challenge at present seems to be determining how best to use this emerging technology.
“Many companies are interested in doing it, but are trying to find the appropriate line in the sand,” Charnock says. “That is, no one cares about an
employee having a bad day or being in a cranky mood. But broad and/or deepening morale problems over a specific policy, for example, are a
‘No One is Singled Out’
She says that analyzing such information is good practice and doesn’t infringe on individual privacy.
“If the majority of your employees really disapprove of something, it is to the advantage of both the company and the employees for
management to be aware of that,” she says. “And, by definition, no one is singled out in such a scenario.”
An important distinction is the degree to which information is public in the first place, according to Catlin. He points out that tweets on Twitter
are clearly external and public-facing, but internal email lists and wikis are less so; while information posted there can potentially make its way
anywhere, it is primarily intended for internal audiences.
Robert Ellis Smith, publisher of the “Privacy Journal” newsletter in Providence, R.I.,
says that a major consideration with employee sentiment analysis is whether information is gathered on an aggregated or individual basis.
“Be very careful about respecting individual privacy in reviewing employee sentiment,” he says. Aggregated analysis should be less threatening to
all concerned, but even then employers should be open about their practices, he notes. “It’s the employer’s obligation to make it clear that this is
being looked at on a cumulative basis,” Smith says.
Still, employers who embrace the technology should be prepared for at least some level of employee backlash.
“It will definitely be seen as Big Brother even if it’s only used in an aggregated fashion,” says Catlin. But he says it’s good for companies to get
employee push-back, because “it makes the companies think about the right way to use emerging technologies.”
He adds that the right of companies to know what’s happening on their networks should be balanced with the knowledge that everyone
complains to one degree or another, so monitoring individuals should not be the aim.
Employers should “aggregate the data across the users and figure out the gist of the complaints,” Catlin says. “That way, the company can fix
actual weaknesses rather than targeting the messenger.”