- Our approach
- About us
An Australian media company had identified that the amount of rolled over holiday entitlement year after year, was reaching into millions of potential liability for the business. They wanted to understand why and how people were accumulating such high levels of unused holiday at the end of every year.
Uncover subconscious insights
For this challenge, the data was readily available. Holidays allocated, holidays taken and day accrued were all collected annually by the HR Department.
The challenge was to identify any patterns that might exist and seek to answer a number of questions.
Were there any days of the week, days of the month or days annually that we could identify using behavioural science data techniques?
And were the patterns evident in every department or just specific departments or was it their age or tenure with the company that made the difference ?
Using the power of cutting edge mathematical, big data and predictive analytics techniques, our behavioural data scientists can take out the guess theory and statistically prove, with a hierarchy, which of the various factors are the most important in driving stockpiling. We asked the data questions like this;
Is it men or women who are most likely to stockpile ? is this the defining characteristic ?
Is it a particular age group that are more likely to stockpile ?
Is this a demographic profile combination of men with over 5 years tenure ?
Or is is a function of which department they work in ?
Or is it a combination of those above ?
Our behavioural data scientists, built an algorithm that could create lots of mini-experiments or variations and different ways of looking at the data (hundreds/thousands) and then seeing how well the calculations predict what actually happens on a data set that the algorithm itself hasn't seen. Essentially it removes 'anecdote' and provides us all with 'empirical evidence' of the key drivers of stockpiling holiday allowance.
From the data analysis, we identified that employees were Chucking a Sickie the day after regular Bank Holidays, such as Australia day. Not only that, the employees were working in pairs and gaming the system. Quite often, one employee would cover their friend’s sick day and get paid double time for the ‘favour’. The ‘sick’ friend would then return the favour later in the year!
We also identified that this was a universal problem. The single most powerful “explanatory factor” of the number of days accumulated is the Operating Division.
This suggests very strongly that it is a cultural issue within the divisions and therefore open to be improved by a behavioural approach. The second most important factor is the Tenure Category and there is some evidence that after 10 years the effect is different than from 0-10 years.
Separating the two gave a more satisfactory model. So in summary, there was clearly a cultural and behavioural issue in Editorial and Operations that we needed to address, as these two areas were the biggest stockpiling groups. This was compounded by people who have been in the business for 10 years or more. It wasn't simply because these teams have more men than women or older or younger people, it's actually the operating division/s themselves that are creating the problems.
(We’ve) decreed two more compulsory leave days – the Aus day and the Anzac day one as well”
HR Director, Australian Media Company. (Action taken post behavioural insight)