All you need is data…or not?
The field of People Analytics is revolutionizing the way we hire, retain, promote and manage employees. But what happens when metrics cannot paint the full picture?
The last decade has seen huge advances in the field of people analytics. Collecting data on employees can help answer a range of questions – from predicting who is most likely to hand in their notice, to classifying employees by personality traits, or measuring the strength of connections between teams. It seems we are only beginning to scratch the surface of applying data to the challenge of management, but have we stopped to consider what we might be missing? What if certain dynamics and situations cannot accurately be captured in the data and metrics we are collecting? Are we propagating issues that arise in the workplace, exacerbating biases in promotions and hiring, or even missing the factors that make employees the most productive?
Some examples of issues that arise when relying solely on predetermined metrics include, but are not limited to [1];
- Time Management: The most productive employees spend comparatively less time on tasks, but achieve more in a given time frame. Measuring time spent at work for employees who “clock in” and out of work for certain roles such as stacking shelves may lead to inaccurate assumptions about who is working hardest.
- Motivations: Using metrics to predict whether an employee is disengaged and at risk of leaving the company is more of a “lagging” indicator than a “leading” one. In order to understand what truly causes employees to find meaning in their work, we must seek to understand their underlying motivations. This may only be possible through qualitative research – talking to people and gathering feedback, rather than seeking data points related to their performance – when it may already be too late to intervene and prevent their resignation.
- Innovation: Employees who seek creative and innovative solutions to challenges, rather than finding safety and comfort in the status quo, may be reflective of a deeper attachment to the organization and alignment with the organization’s success. Measuring innovation and creativity with data is an almost insurmountable task. If we fail to allow for qualitative and subjective assessments of employees’ work outputs, we may rely too strongly on categorizing them into predetermined levels, rather than allowing for the unexpected.
Aside from the risks associated with relying too heavily on data for management decisions, there are additional concerns related to the use of employee data. For example – employees’ rights are called into question if they are not aware of data that is being collected [2].
However, employees may find some solace in the fact that data can more efficiently back them up in claims over pay equity and other litigation [3].
Overall, data can be a helpful resource for managers who want to better understand their employees, particularly as they cannot be present for every minute of the day. It can help them to be more objective when comparing team members, and help them recognize when they may be unfairly favoring one employee over another. However, managers need to remain cognizant of the factors that cannot be measured by the data and metrics they are collecting, and think about how to incorporate the qualitative factors that more accurately reflect the true nature of their employees’ performance.
Hi Eliza, thanks for sharing this article that alerts several issues which should be carefully considered when using People Analytics technology. I really like your three-point summary that raises some interesting and unique points regarding risks involved in People Analytics. I believe taking care of these issues will me People Analytics considerations more thorough and caring for people.
Eliza, I completely agree with your points that people should think holistically about what drives performance, satisfaction, and retention rather than collecting data that is “easy” to get and assuming it will tell you about drivers. However, I wonder if the sources you mention might be assuming that when we over-rely on data, we are over-relying on passively collected data. Even the “squishier” drivers that you mention (time management, motivations, and innovations) can be captured through data like survey output, ratings based on qualitative factors, coded interviews, etc. Your post provoked a lot of thinking for me about how to capture data on important topics! Thanks!
Eliza, thanks for sharing your thoughts. You point out the flaws of trying to squeeze everything that happens in a workplace in a data analysis and I agree with the criticism. The question I have though may be more philosophical: Isn’t everything we observe in the end just data? And if so, shouldn’t there be a way to measure it after all? In the end, our actions are merely reactions to external stimuli, paired with whatever our brain decides to sprinkle on top to process it. I think the human behavior will eventually be fully cracked, assuming we measure literally everything in our lives.
Thank you for this article, Eliza! It is indeed an interesting question: just because we have certain data, should we make business decisions just based on that data? Or should data always be amended by human intervention? I personally believe that even from a technological point of view (i.e. before we ask what’s right or wrong) we are not at the level of technological sophistication yet to make decisions solely based on data.
We should therefore focus our attention less on decision-making but more on creating better transparency and a better information base. We should continue looking for new approaches and experimenting with new kind of data. On the point around motivation: maybe at some point we’ll be able to analyze recordings of customer-facing employees interaction with customers to understand how motivated they came across to that customer. And then offer better training to employees by showing them what works well and what doesn’t based on that analysis. I believe that we have a lot of opportunities (e.g. through cloud computing and data being more centralized and more data being collected than ever) and we should continue exploring those.
Great post, Eliza. And it’s the question of finding a balance on privacy, accurate metrics and usage, that seem as important if not more than conducting the analyses itself. My sense is that before embarking on evaluating predetermined metrics, it’s crucial to set boundaries (and a value system maybe!) on the purpose of the exercise and stick to those. Having said that, I feel it’s harder to do as a collective organization than as an individual.