Happyforce is, as far as we know, one of the few platforms in its category whose data has supported a peer-reviewed publication with an openly available dataset.
The study
Berengueres, J., Duran, G., & Castro, D. (2017). Happiness, an inside job? Turnover prediction using employee likeability, engagement and relative happiness. Proceedings of the 2017 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM '17), Sydney. DOI: 10.1145/3110025.3110132
A collaboration between UAE University and Happyforce, using 2.5 years of anonymised platform data â 34 companies, more than 4,300 employees, more than 221,000 happiness votes, plus interaction data from the anonymous forum â to rank employees by turnover risk.
Key findings
The top three turnover predictors were behavioural and relative, not attitudinal and absolute: likeability (ratio of likes received on one's own comments), posting frequency and relative happiness (individual level normalised by the company average). Precision@50 = 80% on the test set.
Average happiness was not a significant predictor. Being at 6/10 in a company that averages 8 is a very different signal from being at 6 in a company that averages 5. Context is the information.
The anonymised dataset was published openly, enabling independent replication â a level of transparency unusual for commercial workplace data.
How it fed the product
These findings are the reason our analytics privilege deviation, trend and within-company comparison over absolute benchmarking. It is not an aesthetic preference: it is what the data said.
The base we calibrate on
Since 2014, Happyforce has measured 288 organisations across multiple countries: 82,901 employees reached and 65,763 activated (79%). Sustained median participation across organisations is 41%, with the top quartile above 56%. That accumulated history is what calibrates the models.
A note on dataset scope. These figures refer exclusively to organisations measured continuously through the Happyforce platform. They do not include one-off research initiatives, such as the annual World Happiness at Work survey we run with the World Happiness Foundation, whose scope is reported separately.Keep reading
Limitations of the model â where this evidence stops.
The Happiness Index: what it measures and what it doesn't â the indicator the study was run on.
What are the Happyforce Scores? â how these findings shaped the model.