One of the major criticism levelled against those in the human resources function and those manning it is that oftentimes their activities and interventions are so detached from business. This has led other professionals in other functions to believe that this function is of no value. Here is how the human resources function can add value through applying science to the practice of human resources.
1. Predicting sales performance – You can predict which sales representatives or business development officers will sell more and what amounts on average are they likely to sell using regression analysis. This is how we are doing it. We test the sales representatives through psychometric tests (cognitive and personality tests) and collect historical performance data for each of your sales representatives. We eliminate the effects of seasonal influences in the data until we get an accurate prediction; how much variation in sales can be attributed to actual individual sales representative skills and effort. Once we are happy with the accuracy we develop an algorithm for you which you can use every time you are hiring a new sales preventative. This algorithm will tell you whether the candidate will be a good performer or not. Using the same algorithm we can include other dimensions which help us predict which sales representatives are likely to leave your organisation and when. You can use the same algorithm to predict which potential employees are likely to leave your organisation at the hiring stage. The advantage of this combination is that although a sales representative is identified as a good performer, you can establish if they will only stay for six months in your organisation. This gives you the option to weigh the risk of hiring them.
2. Predicting fraud/disciplinary risk – Using integrity, personality and cognitive tests we are able to predict, with a fair degree of accuracy, which employees are likely to steal, get involved in disciplinary issues and other counter-productive behaviour before you hire them. This also applies to your current employees who have not yet committed these offences.
3. Predicting who will make errors (Bank Tellers for example) – For banks, you can also predict which Tellers are likely to make errors before you hire them. This is done by testing your current employees and collecting as much data about them as possible. Using regression analysis, we will be able to build an algorithm that can do the prediction with a fair degree of accuracy.
4. Predicting driver accident risk – Using regression models you can predict which drivers are likely to be involved in accidents before you hire them. If you have already hired them and they are found to have non-compensable deficits, you re-assign them to other functions that minimise the risk or damage they are likely to cause to property and human life. We recently developed an algorithm that predicts which drivers are likely to be involved in accidents. Our model has a high prediction accuracy of 75%. This was after testing our clients’ drivers using the driver psychological test (Driver Plus) simulations and collection of each driver’s accident data. I am sure a number of organisations have employed drivers who are costing them huge sums of money through careless driving. These losses can be reduced if you apply such algorithms to avoid hiring bad drivers and identify those that may need to be re-assigned to other roles. In the same analysis we discovered that being a holder of a Defensive Drivers’ Licence does not reduce your risk of accidents. This could point to the need to restructure how the defensive driver training and testing is done.
5. Predicting customer service complains for Call Centres Staff and all other customer interfacing staff. You give your current staff personality, integrity and cognitive tests and collect historical performance data about them. Again this information is already in the organisation. This will assist you in building a tool that will detect trouble makers before you hire them.
6. Workforce planning– Using the same regression analysis techniques and other techniques, you can predict the number of staff you require in each role. We sit down with you and establish the critical factors that you consider (or should consider) when right-sizing all your departments and business units. These critical factors, which are defined as headcount drivers, are then used in the analysis to predict your optimal staffing levels.
7. Remuneration – This is one area where statistical methods, if applied correctly, can bring a lot of value through calculation of metrics such as pay compression, compa- ratios and range penetration. You can also establish, using regression analysis, if there is a link between individual performance and salary. If there is a link, you can quantify how much of the variation in salaries can be explained by individual performance.
8. Training effectiveness – The best way to measure the effectiveness of your training is to use randomised experiments and use valid statistical tests such as the t- test to check if your training is effective. A comparison of performance before and after training will give insights on how effective the value addition you are trying to infuse in your employees is effective or not
The ideal situation is for human resources departments to measure the success of their interventions by using business experiments. This is the pinnacle of application of science to human resources. Other fields are way ahead and human resources is lagging behind. This is because it is a profession that tends to attract people who are deficient in the sciences and in some cases, rejects who have failed in other departments. This is a tragedy for this profession with so much potential to add value to the business. This same tragedy is perpetuated by Universities and Colleges that seem to think those who can’t do well in other fields should find refuge in such an important function. It is no wonder some business people have started calling this department “the refuge department”. It houses those in distress, with limited capacity for independent thinking.
The biggest challenge in applying science to human resources is that the majority of human resources professionals have a low numerical ability limiting their ability to apply mathematical thinking to business problems. This problem can be solved by infusing non- human resources people into the human resources function. I have assembled a team of talented graduates in Applied Maths, Statistics and Operations Research to assist me in applying science to human resources; a project that I started 4 years ago. Using my own background as a Psychologist (psychology is a science; the way we were taught and trained, not sure about the post 2000 graduates from our local universities). Another hurdle in applying science to human resources is that the human resources function does not keep enough business and human resources data to enable advanced analysis for the benefit of the business. The majority of human resources reports are narrative reports that add no value to the business and no one reads them.
Please note that I only selected a few examples but there are so many areas where we are already applying science to human resources to ensure that decisions made on human resources issues and business are all evidence based.
I appreciate you reading this article. My views are meant to stimulate debate on issues affecting businesses and I welcome contributions against or in support of these views.
Memory Nguwi is the Managing Consultant of Industrial Psychology Consultants (Pvt) Ltd, a management and human resources consulting firm. Phone 04-481946-48/ 481950/ 2900276/ 2900966 or cell number 0772 356 361 or email: firstname.lastname@example.org or visit our website at www.ipcconsultants.com