12 Oct 2022

Text analysis: how can it support internal comms practitioners with employee engagement?

Employee feedback is a powerful tool for any organisation, but it can be difficult to gather, analyse and act upon this information. Here, Relative Insight’s Einat Korn, global head of people, outlines how, through text analysis and comparative studies, internal communicators can uncover the drivers of workplace culture and increase job satisfaction.

Einat Korn sq.jpg


It’s no surprise that engaged employees are more productive, with their positivity – or otherwise – reflecting on the customer experience. But how can you tell which elements of your company culture, policies or benefits are making an impact?  And how can you encourage valued employees to stick around?

One way is via employee feedback, but the crucial element is being able to effectively and efficiently analyse engagement survey results – and then make business decisions based on these findings.

For optimal effectiveness, it’s important to leverage both quantitative and qualitative data given that the latter can give you a deeper understanding of "why" behind these feelings.

Compared with other employee data sources such as exit interview transcripts and company reviews, your survey results have the potential to generate real business change. What’s more, if you were to compare all this internal data to your external Glassdoor reviews, then you’d soon be able to spot dichotomies in opinion.

But how do you process all this data?


The data is already at your fingertips

Today, it is possible to use technologies such as advanced natural language processing to properly utilise text data from all employee data streams. While this mass of unstructured text data can seem overwhelming, the truth is that most businesses already have it at their fingertips and accessing it efficiently can uncover the drivers of engagement – generating actionable insights which can be used to improve culture and experience.

In addition, comparative text analytics can enable teams to compare two or more sources of text to surface the statistically significant differences and similarities – pinpointing the words, phrases, topics, grammar or emotions that are more prevalent in one data set over another.

It can be useful to think in terms of office location – for instance, comparing the employee engagement survey results of global teams to understand the reasons for differences in turnover rate and satisfaction. Similarly, it can be valuable to drill down by team or function – comparing different parts of the business to understand the drivers of experience.

Another example is using this approach to compare your own company reviews to those of competitors so you can uncover weaknesses, strengths and opportunities for improvement. It’s also wise to segment feedback using associated metadata points such as age, team or office location to develop a 360-degree understanding.
 

Assessing employee perceptions through use of words

Looking specifically at Netflix, HubSpot and Evernote – three companies known for their employee perks – we gathered publicly available review data from Glassdoor, and used text analytics to compare and analyse this data to gain an understanding of how these companies are perceived by their own workforces.

We found that employee reviews praised HubSpot for coming through on the perks it advertises, using emphatic intensifiers such as "truly" to convey how impressed they were. Words such as "support", "grow" and "mentorship" also appeared 3.3 times more in this data set, while the words "valued", "belonging" and "authentic" appeared more in this data set than for those of the other two organisations.

That said, HubSpot employees were more likely to refer to the steep learning curve they experienced when joining.

Meanwhile, Evernote employees were significantly more likely to talk about colleagues in their reviews, while Netflix employees were 63.3 times more likely to discuss freedom, responsibility and independence. However, in comparison to competitors, Netflix employees were also 3.8 times more likely to discuss long hours.

These organisations could take this analysis even further by comparing the public data to their own survey results, reviews and exit interview transcripts. It would then be possible for internal communications professionals to identify weaknesses or gaps in the employee experience and respond accordingly.

The beauty of this type of analysis means an unbiased view of how any organisation stacks up in their industry and when compared to competitors.

Being able to benchmark yourself in this way also offers a unique insight into how potential job candidates view you, helping the organisation to attract and retain top talent whilst also shaping policy and employee experience.