The HR Analytics Journey at Shell, interview with Esther Bongenaar
A few weeks ago, we met with Esther Bongenaar, Lead HR Analytics at Shell International in the Shell Headquarters in The Hague. We discussed the HR Analytics general framework at Shell and the working model in the business and decided to interview her for the iNostix blog. We discussed her background, the roots of the HR Analytics function at Shell, insights in several specific projects, the required expertise for HR Analytics and his personal advice to HR professionals on how to start with HR Analytics. It was a great interview with lots of learnings and we are delighted to be able to present Esther on our blog. Here we go with the interview…
Esther, can you introduce yourself please (background, experience,…)?
I am leading the HR Analytics team of Shell International B.V.. We are four staff responsible for predictive analytics for HR in Shell. My education and professional career focusses on applied mathematics. I have degrees in Operations Research and Mathematics for Industry. Most of my professional career was in the statistics team of Shell where I was involved in refinery optimisations, teaching basic statistics to engineers and modelling long-term energy demand for Shell Scenarios. I am good in translating business questions into theoretical problems, letting mathematics solve real life issues.
When and why did you start working with HR analytics and why did you choose to move to HR within Shell?
I joined HR one year ago and I surprised many people, including myself, with that move. Why join the soft side? The decision was made when I realized the impact I can make. HR is sitting on a vast pool of underutilised data and it is an exciting opportunity to build a team that unravels some of Shell’s challenges. Thomas Rasmussen (VP HR Data & Analytics at Shell) recruited me; he is knowledgeable and very inspiring. He convinced me that I am the right person for this job: building the HR analytics capability in Shell.
I do not regret the move to HR; I believe I am relevant to the business in my current role and I very much enjoy the work. I had to get used to the soft metrics though: e.g. what is Employee Engagement really, and how can you interpret the numbers? Temperature measurements in a refinery may include a measurement error, but at least everyone agrees on what 140 degrees Celsius means.
What is your focus with HR Analytics at Shell and why?
We add a new, quantitative perspective to discussions on new and existing HR policies; we quantify business impact of those policies. That has never been done before. We focus on the combined analysis of HR data and Business data. Business data exists in many shapes and forms: e.g. production numbers, financial results or number of safety incidents. The Holy Grail of our work is better understanding of the drivers of employee performance. But performance is very difficult to measure, especially for a highly skilled labour force.
We often use proxy variables that require careful interpretation. Performance ratings for instance do not always reflect an employee’s value to the business. Can you compare an average rating in one business unit to an average rating in another unit? Are short term performance outcomes the right way to measure business impact in an industry that has an investment horizon of decades? Analytics is a word that has different meaning to different people. I should mention that we do not do reporting. There is a separate team dealing with standard and ad-hoc reporting and another team dealing with quality of data. This allows us to focus on predictive analytics.
Is there a difference in the approach between the Shell Group and the local (business unit) level?
No, not really. We are Shell’s global centre of excellence for HR analytics. We have access to data of all our ~92,000 employees. But in our quest for performance understanding, we often do pilot studies in specific parts of the business. We need to do pilots because we do not have one, single metric for performance. You cannot compare the performance of a scientist in our Amsterdam R&D facility with the performance of a truck driver at the Athabasca Oil Sands in Canada. So some of our studies are at global level but other studies are specific to a business or location; it depends on the question and the data.
You recently hired 2 analytics experts in your team without any HR background. Could you tell us something about the profile of these people?
All candidates were measured against these three core competencies:
- Applied Statistics
- Problem Formulation
- Results Driven Attitude
Knowledge of HR was not requested?
Let me elaborate: data analysis in HR is very similar to analytics in other parts of the business. We need to bring those existing analytical capabilities into HR. So, we recruited candidates with knowledge and hands-on expertise in applying statistics to real-world problems. An analyst in our team should be able to formulate a hypothesis or statistical model that will lead to the requested insights after a conversation with the stakeholder. A pragmatic attitude is required that focuses on the delivery of insights because we often need to work around limited data availability.
Was it a big challenge to find such experts? (see job description here)
One of the challenges in the recruitment process was that there are not many experienced people in the market because it is a new discipline. We made the choice to look for people with strong analytical capabilities rather than an HR background as described. We need to deliver counter-intuitive insights to HR leadership and you could even argue that that is best achieved with a fresh, non-HR pair of eyes. We are a small team because we are not involved in HR reporting; we focus on predictive analytics. In Shell, we have the luxury that we can draw specific expertise (e.g. statistics, business or HR) from other teams when needed.
Any HR or Business results you were able to achieve with HR analytics?
Engagement as a driver of performance is high on the company’s agenda. It is not only our achievement, but we have delivered plenty of evidence to support that link. Besides the traditional approach of showing correlations, we put a lot of focus on causality, i.e. the directional relationship between engagement and performance. We actually demonstrated that engagement is the driver of performance. In general we are assessing whether the HR policies we have in place drive towards better employee performance. It is difficult to share the results of our work but I can share some of the topics we are looking into:
- What are the measurable effects of (variable) pay on performance?
- Should we thrive for increased job tenure?
- Do we employ expatriates, our most expensive employees, in places where they add the most value?
- To what extent is management influencing individual employee engagement?
- Can we quantify the effect of our Diversity & Inclusion policies?
These questions are very relevant to our business. For inspiration on new projects, we engage with our HR and business stakeholders, other companies and we also cooperate with universities.
What do you see as next steps using HR analytics?
There are still so many opportunities. Plenty of unanalysed data available for us; I feel like a child in a candy shop. For us the next step is to engage more with Shell’s operating businesses. What does the business want from HR? By now relevant stakeholders in HR know about us and they send their requests. I want more studies initiated by business stakeholders. Something else I would like to expand into is the analysis of unstructured data; think of text in resumes or goal setting documents. For the longer term, I would like to focus more and more on predictive studies.
Did you separate HR Reporting from HR Analytics?
Yes, but of course there is an overlapping area of advanced reporting / basic analytics; think of ratios and correlations. We are creating lots of correlations and it serves us in deriving better understanding of the data that we are dealing with. It also generates inspiration for further analysis. But we are training the HR community in basic statistics and in due time it will allow us to focus more on predictive studies.
What would you change if you had to start again?
My job is predictive analytics; that may explain why I struggle to give a meaningful answer to this question. There are things we have done well and things we have learned. I can even predict that we will learn more lessons in the future. I see lots of opportunities and I prefer to put my focus there.
What advice would you have for HR professionals who are planning to start with HR analytics?
Assuming that data of good quality and some level of reporting is already available, I would suggest investing in obtaining analytical skills. At least educate yourself in critically viewing and evaluating reports/analytics presented to you. I have come across a lot of poor analytics wrapped in a nice story. Studies are either mathematically poor or solving non-existing problems; e.g. a low correlation remains low no matter how appealing the story and employee turnover is only an issue in specific businesses. Know your data and analytics. Preferably get your hands dirty in the data. Then find yourself a credible 3rd party analytical supplier or invest in people that have statistical knowledge. You may call me biased. Without predictive modelling capabilities you will never get beyond reporting.
Thanks a lot, Esther, for this great interview!
- The HR Analytics Journey at Maersk Drilling, interview with Peter Hartmann – HR Analytics Manager
- The HR Analytics Journey at Sears, interview with Ian O’Keefe – HR Analytics Manager
- The HR Analytics Journey at ABN-AMRO, interview with Patrick Coolen – HR Analytics Manager
Interested in using predictive HR analytics as a key component in your HR strategy? Get in touch with CEO and co-founder Luk Smeyers for more information ([email protected]) or via Google+. Or follow iNostix on Twitter and/or Facebook for exciting international articles on HR analytics. And…don’t forget to subscribe for this blog (2 to 3 articles/month, no spam, 100% quality content!).