Groundbreaking big-data research produced by the United Arab Emirates University (UAEU) suggests that, when it comes to keeping employees, it’s less important for them to be happy – and more important for them to feel they are liked.
A study led by Dr. Jose Berengueres, Assistant Professor in UAEU’s Department of Computer Science and the university’s Advanced Analytics Group has assessed “happiness data’ from 4,000 people working in 34 companies across Europe, in order to predict staff turnover and identify precisely what causes the smile that makes them want to stay.
The results of the research collaboration between myhappyforce – an app-based platform that measures, and aims to improve, happiness in the workplace – and the UAEU team will be presented at the pydata.org conference at Barcelona-based management education school ESADE on 19th May, and the prestigious 2017 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, Sydney, Australia, where the topic will be a key area of focus.
The study is one of the largest and most comprehensive data-collection projects on global happiness ever conducted, and has produced the biggest dataset of its kind on the subject to date which has been open sourced by UAEU. As governments, businesses, and institutions look to leverage workplace happiness with a view to increasing productivity and reducing employee churn, it has produced some eye-catching and thought-provoking findings.
Through developing a model capable of predicting whether a worker is likely to quit their job in the next three months with 75% accuracy, the research team found that the most reliable predictor was not whether an employee says they are happy, but “likeability” – the ratio of likes versus dislikes received by their comments on workplace social forums, dictated by factors they cannot control. Meanwhile, the findings also suggest that the notion of ‘Monday morning blues’, as the Western working week begins, may need to be revisited – as other days of the week actually bring a greater amount of gloom.
“Happiness is considered to be good for business and good for governments,” said Dr. Berengueres. “On the other hand, a significant body of research shows that being happy all the time is not necessarily natural or healthy, and can be exhausting. To further complicate matters, happiness is subjective, hard to define, and difficult to measure.
“Of all the factors that influence happiness, work - or the lack of it - is the single most important predictor of happiness. The World Happiness Report states that being unemployed causes people to be miserable. But, through this study, we aimed to look at this topic in the context of those who have jobs.”
The team – comprising project lead Dr. Berengueres, UAEU research assistant Guilem Duran, and myhappyforce’s chief technical officer Daniel Castro – modeled happiness data collated by myhappyforce from thousands of workers over a two-and-a-half-year period and used it to predict turnover, before conducting deeper investigations, using more than 100 different measurements, into the characteristics of this data which are most closely linked to an employee leaving their job.
The app allowed employees not only to log how happy they felt in their work, but also to send a suggestion or comment to a company forum once a day – in the form of “water-cooler chat”, according to Dr. Berengueres – and view and ‘like’ the reaction from peers. “Messages on this forum were a channel for employees to send feedback to management without fear of repercussions, and a way for management to know what their workforce is actually thinking,” said Dr. Berengueres.
The strength of the results allowed the team to group employees into two distinct clusters: “A-type employees”, who account for 75% of the workforce studied and have a high “likeability ratio”; and “B-type employees”, who did not provide feedback and therefore did not see their comments receive many ‘likes’. The study found that B-type employees were three times more likely to quit in the following quarter of the year than their A-type counterparts.
“Another surprising finding was that the indicators that better predict employee churn are not related to individuals, but are those that relate an employee to their peers’ metrics. We were also able to measure how a particular day of the week affects employees’ self-perceived happiness – contrary to popular belief, Tuesday and Wednesday, rather than Monday, are the least happy days, by a thin margin, and Saturday is the happiest day.”
According to Dr. Berengueres, a company’s employee-churn is connected to, and governed by, “a series of features”, some of which are the sole preserve of the individual employee, such as their personal happiness. However, he said: “None of the top features identified through this study are completely internal to the employee, on the contrary they depend on extrinsic factors that are outside the employee’s control, such as the number of likes they receive and the relative happiness of their peers."
“If we hypothesize that a happy workforce is correlated with less churn, so are the features that generate happiness. This leads to the conclusion that, when it comes to work, happiness is not as much of an ‘inside job”, as is sometimes claimed, but is actually significantly dependent on factors outside the control of the employee.”
Daniel Castro further added that: “This study not only confirms that people need to share and validate opinions or thoughts with their peers, but also shows how important it is for organizations to invest resources in listening to and understanding how employees relate in the workplace”.