Extracting insights from employee surveys using text analysis

Knowing what employees think about their working environment is vital to keeping them motivated and productive. It’s never been easier for people to change jobs, with the Great Resignation becoming a self-fulfilling prophecy. In this landscape, acting on insights from data sources like employee surveys is vital for organizations.

It’s with this backdrop that a large software company used Relative Insight to analyze open-ended responses from its most recent employee engagement survey. The survey used a combination of closed and open-ended questions.

It asked employees to rate different aspects of the company, then quizzed them on reasons for their responses. While it’s rare that questionnaires feature this many open-ends, to really understand what respondents think you need to give them the chance to express themselves.

Relative Insight quantifies the differences between free text responses from varying respondent groups. This process highlights what matters most to each segment, allowing you to identify actions that you can take to address their concerns.

For example, by using Relative Insight’s survey analysis tool on employees’ overall rating of the organization, the company uncovered a vast difference in the way satisfied and dissatisfied staff talked.

Even those who rated the company highly overall gave some critical feedback, yet this was constructive and rational. They were infinitely more likely to use the word ‘challenges’ – meaning this didn’t appear in dissatisfied staff’s responses – and were specific about what these were.

Respondents with a negative overall experience were more emotional. They used words related to ‘disappointment’ and ‘worry’ infinitely more, while expressing ‘frustration’ 10.3x more.

Using this insight, the company’s people analytics team recommended that its HR staff connect with dissatisfied employees on an emotional level before trying to treat their complaints rationally.

Assessing employees based on length of service

The Relative Insight platform enables you to segment responses through different data points. This might be information your organization already has, or data you’ve obtained using closed survey questions.

One of the ways the software firm grouped staff was through length of service. Splitting and comparing responses based on number of years within the organization revealed fascinating insights around learning and development opportunities.

People who had been with the company between six and 10 years were positive about them. They discussed the business giving them the chance to ‘develop’ 6.8x more, and were 4.8x more likely to talk about ‘skills’ they’d learned. This group had clearly benefited from the company’s learning and development programs.

“The company encourages and supports its employees in their career growth, offering opportunities to learn and develop new skills.“

In contrast, employees who’d been with the company less than two years complained about learning and development. They talked about ‘limited’ progression within the organization 5.1x more, relating to both development programs and promotion.

With regards to the latter, this group of staff thought there was a lack of fairness surrounding career advancement. They complained of ‘nepotism’ infinitely more and also used the word ‘merit’ infinitely more – within the context that promotions weren’t made on merit.

“There are very limited career development opportunities within the organization. Promotions are often given to people based on nepotism and favoritism, rather than merit, which makes it difficult for me to advance.“

The people analytics team suggested that the organization’s L&D function evaluate how its offering has changed over the past 10 years. This would identify why there’s a discrepancy between long- and short-tenured employees. They also suggested implementing better-defined career development pathways to avoid accusations of favoritism.

Managers discuss flexibility 8.4x more

The organization also captured staff seniority within the survey. Comparing how employees at different levels talk about the business provided another opportunity to learn about engagement levels.

The most surprising variation came between mid-level and manager-level employees. Broadly, only managerial responsibility separates these two groups, therefore you’d expect their outlooks to be similar. However, there were some key differences, particularly when it came to the benefits each group wanted.

Mid-level employees focused on tangible benefits available to them. They were infinitely more likely to discuss these, mentioning offerings like ‘retirement plans’, ‘paid time off’ and ‘health insurance’ as desirables.

“They should offer better benefits such as health insurance, retirement plans, and paid time off.“

Despite managers being one rung up the career ladder, their priorities change significantly. They crave ‘flexibility’ rather than more tangible benefits. This group used the word 8.4x more than their mid-level counterparts, with many asking the company for to incorporate this as a way of supporting their work-life balance.

“The company can improve its salary and benefits package by offering more flexibility and options for work-life balance, such as remote work options or more generous vacation time.“

Based on these findings, the company looked at ways to offer managers greater flexibility. It also examined their workloads, given the shift in desire for flexibility when compared to mid-level employees.

Delving granularly into employee surveys through multiple data points

As well as segmenting employees into groups, the Relative Insight platform can also split survey responses by individual metadata points. This means you can drill deeper into specific focus areas. In this case, the software firm wanted to know the differences between good and bad management within the organization.

First, the firm isolated responses to the following questions from its employee survey:

Out of 10, how would you score your line manager (with 1 being very poor and 10 being exceptional)?
Please state why you gave your line manager this rating.

It then used the platform to split the responses into negative and positive feedback by utilizing this scoring. This crystallized the attributes behind good and bad management within the business.

Employees who rated their managers highly cited ‘regular feedback’ infinitely more, while also describing them as ‘supportive’ 5.0x more. They also talked about managers’ ability to develop a team’s people and processes, using words related to ‘improvement’ 3.4x more.

“My line manager is supportive and provides constructive feedback. They are also open to hearing feedback and making changes to improve our team’s performance.“

This availability contrasted with managers given bad ratings. Employees stated that these managers were ‘unresponsive’ infinitely more and used the word ‘lack’ 7.3x more, referencing a lack of communication and feedback. It’s clear these staff members feel their managers aren’t there for them.

“They’re often unresponsive to my needs and concerns, and there is a lack of communication and support from them.“

These results chimed with managers’ statements about work-life balance. Based on this feedback, the organization’s people analytics team undertook further analysis of managers’ workloads.

They looked to understand whether managers are sacrificing work-life balance to be available for their team members. Or, are managers unavailable to staff the ones who feel like work is invading their personal life?

This ongoing investigation will determine whether the managers rated poorly need support with prioritization and time management, or whether all managers have an unsustainable workload that can only be met by working out of hours – and therefore their duties need to be redefined.

Getting the most from employee engagement surveys

By utilizing the Relative Insight text analysis platform, an enterprise software firm analyzed, quantified and visualized free text responses to its employee surveys faster and in greater depth than ever before. This enabled it to act on the insights uncovered from the survey immediately, demonstrating to employees that the company takes feedback on board.

Interested in how you can replicate this type of analysis for your employee surveys? Try the Relative Insight platform for yourself and learn how you can maintain and improve employee engagement.