By no means do we want to encourage you to think about work during your summer holiday. Instead, take a step back from the office, switch off your work phone and lock your laptop in the deepest and darkest drawer in the house. Time to head to the beach with your favourite cooled beverage and a fun book! We collected three easy-to-read books with strong messages to help you take a fresh view on data-driven recruiting. You won’t be thinking about work, we promise. They’re all ideal reading material for under the parasol!
Book 1: The End of Average by Todd Rose
When measuring doesn’t measure up
In the 1950s, the American airforce measured the limbs of over 4,000 pilots with one goal in mind: designing the optimal cockpit for a fighter plane. All pilots were measured on more than 140 dimensions, including thumb size and distance between the ears, in order to determine the optimal size of the cockpit.
This optimal size was determined by taking the average of every dimension and making the cockpit fit the average-sized pilot in each of these dimensions perfectly.
However, there was only one problem: even when the pilots were allowed to diverge from each of these averages by 30%, still none of the pilots would be within this 30% on every dimension!
This is just one of the many examples of problems with averages Todd Rose mentions in his book.
The problem here is that the size of human limbs are only weakly correlated. In other words, someone with tall legs does not per definition have a large torso. Furthermore, this weak correlation does not only apply to human limbs but also to psychological characteristics such as IQ and personality traits. Note that this contradicts common knowledge.
We usually assume someone with a high IQ excels in all cognitive tasks, while in reality this is not the case.
So what does this have to do with HR and recruitment? Well, a lot:
- To which extent does designing the optimal cockpit differ from writing a job description? I.e., when we write a job description, are we perhaps too frequently assuming the preferred skills of the candidate to be correlated? And if we assume these skills are not correlated, how many potential candidates would remain suitable for the job?
- What does an average assessment score truly tell us about the candidate? Even if two candidates have the same score, they can still differ a lot in terms of their skills.
- When we present results, such as average vacancy lead time and turnover rate, how representative are these numbers for individual vacancies and employees?
Luckily, the optimal cockpit problem had a simple solution: make the seat adjustable to the dimension of the pilot. Rose presents many more solutions that can help to answer the problems mentioned above. But, to not spoil the fun of reading the book, I’ll not discuss them here. In short, it is a must-read!
How to present information that allows people to make better choices
On which vacancy on a career website do job seekers apply to most? For most of you, this won’t be a surprise: it is the vacancy on top of the vacancy list. Even if there might be a better vacancy for the job seeker on position nine, the job seeker will still be more likely to apply for the vacancy on top.
Furthermore, application probabilities seem to drop about 50% when moving the vacancy two positions down. Hence,
the choices of job seekers are not solely determined by their needs but also by the way information is presented to them.
When we “nudge,” we try to use this to our advantage, as nudging is about how to present information in such a way that someone will make better choices, without jeopardizing his/her freedom of choice.
Some systems are designed so poorly, they almost force us to make wrong decisions. This book shows us how to avoid this, by explaining the kind of heuristics people tend to use when making decisions, and by giving many examples of how systems were adjusted to adapt to these heuristics.
Note that a “system” here is not per se an app or website, but it could be the system describing how to display the food in a company canteen, or how to best onboard new personnel. Both examples given by Laszlo Bocks’s book Work Rules!
Nudging provides HR and recruitment with a powerful tool: it provides a way of successfully implementing HR strategies without forbidding employees to do something their own way. Although the book does not explicitly mention analytics, it is entangled with nudging: how else will you measure whether the strategy had the desired effect?
If you want to have some idea about machine learning, but not too much
Of course, the list would not be complete without a book about machine learning – and we wouldn’t be true to our promise of recommending only fun reads if this book was written in a scientific mumbo-jumbo language. If you are tired of people starting to spit out mathematical formulas or telling you “it’s a black box” whenever you ask them about what machine learning or any of its algorithms entail, this is the ideal book for you.
Domingos takes the challenge of explaining the various algorithms and perspectives in machine learning but in a non-mathematical and novel way.
As more data scientists start crunching HR data, having some idea of what they are doing will be essential, especially when it comes to translating statistics to action and fuzzy business objectives to quantitative analysis.
This book might not take you all the way, but it does provide a useful map through the field of machine learning, without the hassle of understanding the mathematics.
Is your summer holiday yet to come, and you want to get a real taste of recruitment analytics? Check out the posts below!