In an increasingly divisive American workplace, understanding employee sentiment has never been more critical to productivity and shareholder value. Do you have the resources to hear what your people are saying?
By Kevin Oakes
Americans are more divided than ever—and so are our workplaces. While this began in earnest with the #MeToo movement, it exploded with the Supreme Court confirmation hearings for Brett Kavanaugh. Every day seems to introduce new divisive issues, and our organizations are suffering as a result.
The most quantifiable aspect of this suffering is in shareholder value. We’ve seen dramatic reductions in market capitalization of several organizations as prominent executives are marched out the door due to sexual harassment scandals.
While Harvey Weinstein remains the most visible punching bag, corporate executives from CBS CEO Les Moonves to Intel CEO Brian Krzanich have exited amid allegations. The heightened scrutiny of workplace behavior has affected Nike, Lululemon Athletica, NBC, Wynn Resorts, Papa John’s, and the NBA’s Dallas Mavericks, among many other notable organizations.
Unsurprisingly, a recent study in the Harvard Business Review found that a single sexual harassment claim can be enough to dramatically shape public perception of a company and elicit perceptions of structural unfairness. When people learn that a sexual harassment claim has been made, they not only see that company as less equitable than one where no such claim was filed, but also less equitable than an organization where a claim of a different transgression, such as financial misconduct, was made.
In fact, according to the study, a sexual harassment claim is seen as more indicative of a culture problem than a bad apple problem—even compared to a claim of fraud.
Sexual scandals are merely a part of the greater issue that organizations are facing. The Kavanaugh hearings created even more division between men and women on numerous workplace issues, ironically as companies are doubling down on inclusion in the office. The divisiveness also continues across race, sexual orientation, and religious lines, creating a myriad of concerns for HR departments across the United States.
This doesn’t appear to be settling down anytime soon. While most CEOs would try to diffuse division and promote inclusion and belonging, the CEO of the country has a habit of reinforcing and introducing divisiveness. Eventually, push is going to come to shove.
The Invisible Costs of Not Listening
While a loss in stock price is easy to calculate, less employee engagement, lost productivity, strained customer relationships, and reduced product quality aren’t. These are the invisible costs that hit your company the hardest.
Two decades ago, Harvard professor Amy Edmondson wrote about psychological safety and its importance in team effectiveness and interpersonal risk taking. In 2015, Google made a splash when its internal research “discovered” that psychological safety was the single-most important aspect of successful teams (something Edmundson chuckles about). But despite the recognition, psychological safety appears to be at a low point in many organizations.
Does workforce performance suffer if some employees don’t feel safe or valued because of the actions or comments of senior associates or coworkers? That answer is obvious. The real question for organizations: What should we do to mitigate these risks? While the answer is remarkably simple—listen to your employees—the execution is decidedly much more complex.
Today, 73 percent of organizations still rely on annual surveys to uncover pressing issues and measure employee sentiment, according to research from the Institute for Corporate Productivity (i4cp). The problem is that these increasingly serious issues are raised on a daily basis, and employee sentiment fluctuates just as quickly.
With issues communicated instantaneously via social media, organizations are expected to act immediately before a crisis erupts. We need a better understanding of cultural issues that affect the workplace as quickly as possible, which means annual or even semiannual surveys simply won’t suffice.
To make matters worse, these surveys also don’t work from a design perspective. Today, more than 80 percent of companies report that they primarily use multiple-choice questions in their surveys, per i4cp research. This effectively pigeonholes employees into giving limited responses to canned questions that don’t capture what they really want to say. It’s impossible to cover the myriad issues, especially the sensitive ones, in Likert-style questions.
Maybe that’s why only 21 percent of HR leaders surveyed by i4cp say that employees are strongly heard. That’s a startling statistic given how quickly shareholder value, market perception, and employer brand can erode if companies don’t get ahead of potential employee issues.
Unlocking the True Voice of Employees
To truly understand employee sentiment, leading-edge organizations are shifting from traditional survey programs to a new listening paradigm focused on the “voice of the employee.” This transition leverages employees’ unique words, usually in the form of free-form text comments, to measure their experiences as a member of the organization.
This is possible through the advanced capabilities of natural language processing (NLP), a branch of artificial intelligence that helps computers understand, interpret, and manipulate human language. Computers have typically required humans to communicate in a specific, precise programming language to interpret data.
Human communication, however, isn’t precise. It’s often ambiguous and includes sarcasm, slang, social context, and other elements that require interpretation.
NLP, combined with machine learning, allows organizations to emphasize qualitative data instead of quantitative data, the latter of which is the focus of 95 percent of most employee surveys. Survey practitioners have long understood the benefits of capturing and using qualitative data. Analysis of free-form responses is a valuable means to identify the key issues that an organization needs to address. However, most surveys continue to rely on closed-ended (Likert-scale and multiple-choice) questions.
Despite their popularity, closed-ended style questions have significant limitations:
• Structuring the response scale: Should respondents be forced to have an opinion, or will a neutral midpoint be offered? Including a midpoint reduces measurement error but does little to inform why someone provided a neutral response. Were they truly ambivalent on the topic, or did they choose the neutral response to withhold providing a response that reflected their true feelings?
• Single-item reliability: With closed-ended questions, single-item scores aren’t as reliable as a cluster of related items. As a result, many closed-end survey tools use groups of items or indexes. These multi-item dimensions, like “job satisfaction” or “engagement,” are constructs that aren’t that useful to derive specific insights from, or to determine a path to action. In fact, post-survey focus groups are frequently used to “dig deeper” to gather meaning in the words of employees.
• Understanding the message: One of the most important limitations of the close-ended, scaled response question is the difficulty of understanding what message the survey participant is communicating. The most often used closed-ended scale type is the evaluative scale. Using an evaluative scale (5 = Excellent and 1 = Poor), respondents are asked to assess their experience of an event, practice, or phenomenon comparing what they expected to what they received. The problem with using this type of scale is a lack of any standard describing what “good” means. One individual’s “satisfied” may be another’s “dissatisfied.”
• Utility: Identifying an issue and knowing what to do about it are two separate steps on the road to organizational improvement. Scaled response questions may be useful to identify an issue in the organization, but generally the response options provided in a scaled item aren’t designed to surface innovative possibilities, a strength of the open-ended question format.
The descriptive open-ended comment (“If you have any thoughts, please write them here”) and the prescriptive open-ended question (“What one or two things could we do to make the company a better place to work?”) are common. In response to these relatively uncomplicated questions, organizations typically gather thousands of comments from their employees.
How to Organize the Open-Ended
The challenges associated with carefully processing (reading, categorizing, and analyzing) open-ended comments, particularly as the volume increases, aren’t insignificant. As a result, these types of questions have historically received insufficient attention. They’ve been relegated to a minor add-on role and generally sit at the end of surveys.
For most companies, the greatest emphasis when analyzing survey results has been placed on questions with responses that are readily quantifiable (scaled items). Comment data, because of the limitations in making it quantifiable, has received dramatically less attention.
This is a huge miss. Employee comments can have immense value when it comes to understanding the health of an organization or the risks that may be lurking. Open-ended comments are a self-directed, authentic source of feedback and ideas. In other words, employees can comment on areas that are important to them, as opposed to the survey designer.
Employees take advantage of the free-form format found in open-ended questions to provide input in their natural voice, unconstrained by the requirement to select from predefined response options.
Appropriately, high-performance organizations are turning to robust NLP analysis platforms to help. These platforms perform multiple tasks on each text body, including sentiment analysis for positive, negative, and neutral assignment; theme detection for major categories and subcategories; statistical outlier targeting; and highlight generation of significant issues.
Adding in the extra element of machine learning provides a continual improvement cycle, where results are constantly refined to advance accuracy and relevance. This ensures that the results are contextually accurate and meaningful to HR and business professionals.
While employee comments at the end of a survey are a common source for qualitative feedback, unstructured input can originate from various sources:
• Onboarding surveys and new hire surveys
• Stay interviews and exit surveys
• Alumni surveys
• Coaching sessions
• Focus groups
• Upward feedback and 360 feedback
• “Always on” channels that allow employees to initiate feedback or respond to broadcast response opportunities
• Social site analysis
Accessing multiple sources allows organizations to take wide snapshots of sentiment that they can benchmark. For example, what are people saying on Glassdoor about us versus our competitors, or versus what people are saying internally? This provides a more complete picture of the employee experience.
Beyond Engagement: Employee Experience & Employer Brand
Employee experience (EX) is a broad, relatively new measurement concept that focuses on how employees experience the organization. This includes their feelings about their interactions with policies, practices, and activities in the organization. A favorable employee experience is motivating, while an unfavorable employee experience is demotivating.
Employees who are satisfied with their EX are more likely to respond favorably on Net Promoter Score assessments (NPS), have higher levels of engagement and commitment to the organization, and represent a lower retention risk.
Organizations want to provide a positive employee experience because having an overall favorable EX combined with strong performance capability is the path to high performance.
Employee experience measures typically examine the following:
• Engagement level
• Employee NPS
• Personal development and career advancement prospects
• Relationships with peers and supervisors
• Work-life balance
• Pay and benefits
• Reactions to organizational policies and practices
While employee experience is important for your current workforce, it’s just as important for future workers. More HR professionals, particularly those in talent acquisition, are being tasked with overseeing the employer brand.
The advent of social sites like Glassdoor and Fairygodboss provide new transparency for prospective employees to assess what life is like at a company. A strong employer brand and skilled recruiters aren’t enough to lure top talent anymore. CEO ratings, employee reviews, and understanding an organization’s bigger purpose all factor into the brand someone wants to work for.
Nothing destroys that brand quicker than the illicit behaviors of senior leaders. If you want to be truly strategic to your organization, start by creating strategies to understand and potentially mitigate the risks that are lurking beneath the surface.
The cost, whether it’s calculated by stock price or other metrics, is simply far too big to ignore.
Kevin Oakes is the CEO of the Institute for Corporate Productivity, the fastest-growing and largest corporate network focused on the practices of high-performance organizations. He has been a leader in the human capital field for the past two decades, and was previously the founder and president of SumTotal Systems.