Identifying sentiment and searching for patterns
Employee sentiment analysis is, like all organisational initiatives, about driving maximum improvements cost-effectively. This requires quality data. Therefore a strong methodology and pedigree in collecting sentiment data, to avoid magnifying edge cases or ignoring critical patterns, is key. Identifying patterns across the workforce is ultimately the goal, otherwise you risk overinvesting in areas that will not deliver satisfactory returns.
Identifying patterns obviously takes skill, but even at the level of individual employees, it’s actually quite hard to interpret employee sentiment without experience and expertise: in a study focused on computer-mediated sentiment analysis, the authors found that two humans trying to determine the emotions expressed in 50 emails could only agree on three-quarters of them. Correctly identifying sentiment is obviously the first stage of sentiment analysis, which is why many organisations work with external partners who can do it accurately.
The importance of objectivity
External partners also bring objectivity, although let’s remember that there’s no such thing as complete objectivity. Objectivity is very important to sentiment analysis. Those who analyse sentiment must not be too close to the subjects or the situation, or their biases and views can skew datasets, misidentify patterns or poorly contextualise opinions due to pre-held beliefs or ideas. This is why internal measurement of employee sentiment, while useful for giving a steer, may deliver data that is less accurate.
Let’s give an example: what may be dismissed as non-pertinent by an internal manager (because the employee is widely seen as a cynic, for example) may get flagged by an external organisation, corroborated with other data and passed to leaders because it implies a real and tangible risk of employee flight. External organisations may be more able to present data ‘as is,’ because they are removed from the structures, people, narratives and ideas that end up receiving criticism or praise.
Why a robust methodology is essential for accurate data
It’s also really important to ask questions in the right way, otherwise you tend to get the answers that you’re expecting and want to hear. Just as scientists conducting experiments must avoid leading subjects and lawyers must avoid leading witnesses, survey designers or those collecting sentiment in person must ensure the way they ask questions does not imply that a certain answer is desired. The sentiment should come solely from the employee.
Of course, this requires expertise. And it’s not just individual questions but the methodology of the sentiment gathering in general: are there ‘control’ questions to enhance confidence in the conclusions and to help identify anomalies? Don’t forget that employee sentiment changes day-to-day, and it’s important that the effects of ‘bad days’ and ‘good days’ are controlled for. It’s not that good days and bad days are irrelevant, but they must be contextualised in the right way to give the right conclusions.
Driving change the right way: one step at a time
It’s also important to remember that, while you’re focused on collecting accurate data, driving leadership action is the ultimate goal. Talent Daily spoke to hundreds of Heads of Engagement, who said their biggest problem was putting engagement problems on the agenda and getting leaders to do something about them. That’s why it makes sense to work with external providers who peg their sentiment analysis to a scaffolded approach to improvement, so that easy next steps can spiral into significant long-term improvements.
Investors in People, for example, uses employee sentiment to help organisations understand where they are in many areas such as wellbeing and innovation, so they can better understand what their next step is to performing even more strongly in these areas. This treats continuous improvement as a going concern, as a journey, and it’s much easier to get buy-in for smaller steps on a journey than for transformational initiatives. Using an established framework also maintains a strong connection between employee sentiment and clear organisational goals, which can help prevent a disconnect between the needs of employees and discussions in the boardroom.
Ultimately, any insight into what employees are feeling is useful to organisations. Simply talking to your staff and finding out how they feel on a day-to-day basis can mould leadership behaviours for the better. But for genuine insight into employee sentiment across the workforce, with major patterns identified and courses for improvements clearly marked and based on proven validation, working with external partners makes sense for many organisations.