Using data analytics in the workplace to measure employee performance is more objective and scientific than human decision-making – and achieves better results, right? But there are signs that conventional wisdom around the use of big data is beginning to be challenged, most recently in a new study conducted by the University of Sydney Business School.

The rise of data analytics tools for workplace planning and management has been particularly impactful in the HR profession, where organisations are using aggregated big data information about employees to make decisions about engagement, hiring and firing, onboarding and future planning, among others.

Industries using big data to make day-to-day decisions on staff management include the service industry, consulting and design companies and hospitals, as well as tech giants such as Google and Amazon.

The research pushes back against the assumption that using data analytics such as word combinations used on social media, images clicked and time spent on job-seeking site LinkedIn to measure – even predict – workplace performance is an adequate replacement for people-driven decision making.

Looking back on 2016, it’s not a stretch to suggest that our reliance on big data may have reached tipping point. Statistical analytics led professional pollsters in both the UK and the US to incorrectly bet against Brexit and the election of Donald Trump, respectively.

At The Guardian, William Davies suggests that “the ability of statistics to accurately represent the world is declining.”

There’s also growing evidence of a decline in levels of trust in statistics. The sentiment in populist thinking, Davies writes, is that “there’s something arrogant and elitist about reducing social and economic issues to numerical aggregates and averages.” Stats alone, he suggests, are a poor reflection of lived experience; not least the complexities of human behaviour.

However in private companies in particular, analytics are becoming central to day-to-day operations and increasing influential to every corner of an organisation’s operations.

And while companies see big data as more objective, faster and scientific than human decision-making, some worry it can also be inaccurate and far less nuanced than human-led decision making.

Gaming the system

Uri Gal, from the Discipline of Business Information Systems at the University of Sydney Business School has been testing the ways that some employees game workplace analytics systems: by entering inaccurate data on their productivity or using spreadsheets that paint a particular picture of their performance. His research has also shown how social media can be manipulated to create a false impression by workers.

Gal suggests that the ways that data analytics systems come to conclusions is simply not accurate enough to replace human decision-making because “inherently what they try to do is to build a simplified model of complex human behaviour.”

Why algorithms don’t make good managers

He also believes that a long-term reliance on algorithms in workplace decision making will affect management within an organisation.

“The less active people become in decision making, the less skilled they become as managers. Analytics cannot capture what is involved in managing people or showing someone how to master a skill.”

It can also lead to a breakdown in effective people management. When an algorithm is used as the basis for an action, “human beings are extraneous to the decision-making process,” says Gal. What is lost is the ability to speak with your manager to overcome a workplace issue and more broadly the potential to bring independent thinking and creativity into workplace management.

Do we need to think twice about data analytics?

Not necessarily, says Gal. But companies do need to be more critical about the way they integrate these technologies into their business. Despite the many flaws involved, companies continue to embrace big data to measure workplace productivity without thinking critically about the consequences of replacing people-led people management.