Felix Schmirler is a London-based People Analyst and Insights Specialist at Penguin Random House where he’s in charge of various analytics projects in the HR department.
Felix holds a bachelor’s degree in Psychology from the University of Heidelberg and a master’s degree in Organisational Psychology from University College London.
People Analytics = Solving Sudoku
For me, people are the most exciting part of my job and the business world in general.
Before working as an Analyst, I studied psychology in Heidelberg and London. Right from the start, I had a great fascination with the interplay between people and business. My first work experience was in Learning and Development but I soon pivoted more towards the measurement side of Organisational Psychology: employee behavior, satisfaction, performance, and selection.
Psychology is a quite data-driven subject and many people don’t know that statistics is an integral part of it. However, during my undergraduate degree, my interest in statistical methods was rather low. It was really my Master’s degree in London that got me excited about statistics, empirical research, and, eventually, data analytics.
I tried to find a challenging project for my dissertation and investigated to what degree people modulate their voice pitch when they talk to friends or corporate clients. One of my findings was that women lower their voice pitch when they speak to corporate clients, while men did not show any change in voice pitch on average.
Working with data is like playing Sudoku – it’s about solving complex problems coupled with analytics and creativity. I enjoy using statistics or new technologies to find better answers or even ask new questions that could not be asked before.
When I came to London, I worked as a business psychologist, providing psychometric tests for large companies that used our insights for selection and talent development. We analyzed performance data and developed personality profiles that allowed our clients to predict which candidates are more likely to be high performers within their specific environment.
In 2019 I eventually worked as a freelancer for a few months before starting my current position as a People Analyst and Insights Specialist at Penguin Random House in London.
Opportunities & Challenges in People Analytics
I’m in charge of various analytics projects in the HR department which show how broad the term people analytics is.
Sometimes it can be a challenge to get some basic metrics for reports such as
- How many people get promoted each year by function and department?
- How many people receive a salary raise each year by function and department?
- How many people leave the company each year by function and department?
This is often not because the data doesn’t exist but it can be distributed across different systems and needs some manipulation before you can pull the data in the way you want and properly capture individuals who move teams.
Once you have the right data in the right format you can start to analyze patterns in turnover, career paths, tenure etc to create insights that can help the organisation to plan better or to improve for example career conversations between managers and employees.
Another big topic in people analytics is experiential data: Surveys. New Joiner Survey, Leaver Surveys, Engagement Surveys, etc. At PRH we have started a few years ago to send out an inclusivity survey to better understand belonging within the company and if people experience their work-life differently depending on their department, their origin, or their socioeconomic background.
Another recent project I’m working on is the creation of a uniform database to better evaluate and visualize data.
It’s more challenging to change human behavior than existing IT systems.
I have made the experience that it is tough to change existing systems because there are many dependencies, technical as well as human. It always involves a lot of effort and resources to improve established processes.
Furthermore, many analysts often underestimate the difficulty of changing human behavior. The added value of such a change is often hardly visible at first. Therefore, it’s beneficial that you know the business and build up good relationships with your stakeholders. You have to understand what motivates and triggers people.
Analyse Trends not Individuals
Employee data is one of the most sensitive data in each company and I understand that some people are concerned about what might happen to their data. Therefore, many security measures are essential. I can assure you that very few people have access to employee data and that security concerns are front and center of every analytics project or system change.
In my daily work, I’m not interested in individuals.
My daily practice has a strong research focus and that I look at employee data to find important trends. As mentioned before, I want to find answers to questions. I am not interested in individuals. It’s not my intention to predict when person X will leave the company with, let’s say, a 75% probability over the next months, so I can communicate this to his/her manager as early as possible. Such use of data is wrong in my opinion and would hurt employees who might not have any intentions to leave while also diminishing trust in the organization. But quite a few larger organizations work with these kinds of models, more so in the US than Europe though.
My main priority is to provide objective information for decision-makers. If we could predict accurately that turnover in team A will be between 25% and 30% next year we can plan for it and if we expect turnover in team A to be much higher than team B, we can start asking more questions to understand why. This insight does not affect the individual but it helps the company to plan better. I want to go even further and say that people analytics can help to identify opportunities to respond to employee needs such as career development.
Limitations from GDPR
I haven’t noticed any limitations so far regarding the strict data privacy regulations in the EU and personally think they are a great step to give the individual more power to decide about the use of their data. Some of the data we analyze, are collected entirely voluntary and anonymous. All other data I analyze for organizational purposes is pretty much what is on your employee file, so the only information we as a company have the right to own and use (e.g. start date, time since last promotion, etc.) but even this information is only used on an aggregate level. We do not access data outside of the employee file like social media.
In general, I don’t see much risk in data itself, but in the resulting actions. You must be careful that the findings you communicate are not used in a discriminatory way towards applicants or employees. I believe the fear of misuse of insights overall is somewhat overrated. But when you start to automate decisions such as in automatically scored assessments or individual turnover risk scores, things get complicated. At least from my experience, however, HR is very cautious in general when it comes to risks for data privacy or discrimination as they are already used to dealing with people’s most sensitive.
Anyway, I think if you understand the mechanisms behind Data Science and Analytics, you’re less afraid. But you should always be able to explain why you need people’s data and what it is used for.
Create Value from Data
While HR is probably not perceived as a very data-savvy department, senior decision-makers within HR already have a big appetite for more analytics and they are asking many of the right questions. The challenge is really where to start because you have to start working within an »organic« system of data sources and processes which has not been set up to do all the things you have in mind. Something that helps your data quality and could help to make some friends is automating processes even if it is with Excel. Some other projects might cost more effort.
One thing where I feel that I can add value is when it comes to answering questions about differences and trends. While often reporting can be enough to understand one business area differs from another, it can be hard at times to interpret more fuzzy and complex data. The ability to deploy statistical analysis in order to make more informed statements on the relevance of trends and differences is certainly something that seems to be appreciated.
Future development of People Analytics
The biggest driver for innovation in people analytics is the availability of emerging technologies. From research, we know for example that, social media data are excellent sources to predict the personality traits of applicants. Better than self-report tests in most cases where social desirability bias can distort results. However, I’m personally 100% against taking private data from Facebook and Co. for assessment purposes. It is unethical to put private and professional life into a prediction model, even though the result (prediction of your work) can be more accurate.
Another interesting topic is AI in candidate assessment. Companies like HireVue are already using AI to score video interviews. Overall this could be exciting because it allows you to bring more objectivity into the process. Still, it is a risky project if algorithms and models are not diligently constructed.
I don’t see the future of people analytics as an episode from Black Mirror in which AI has all of your data to monitor and control your life in real-time. Instead, I believe that large organizations will widely adopt machine learning approaches to predict turnover, NLP approaches for survey analysis, and create some other exciting AI-based tools such as internal career platforms or chatbots. Advanced analytics can also play a big role in identifying inequalities. In the end, it is still down to humans to interpret findings and take action – at least for the foreseeable future.
