A recent Forbes article, Should Big Data Be Used To Measure Employee Productivity?, includes some interesting insights on a study done by Society for Industrial and Organizational Psychology. The study specifically looks at how employees respond to big data/machine learning when applied to employee productivity.
According to Forbes,
“Research shows that employees don’t like large scale, intense data mining programs. Such programs may lead to decreased job satisfaction, well-being, organizational commitment, attitudes about fairness, and performance by providing a sense of invasion of privacy. However, there are certain conditions where electronic monitoring and data mining can actually increase motivation, job satisfaction, commitment, and performance.”
As a part of the push to improve employee productivity, big data and machine learning are being added to the time tracking process. Getting people to fill out timesheets is already a struggle. So how do you avoid making the problem worse by making a disliked activity even less popular with data mining?
The study suggests five ways that you can improve employee acceptance when implementing data mining:
1. Make a plan – For the collection of time data, this step has always been important – big data or no. You need plan out why you’re collecting specific data, how it will be used and how it is relevant to the employees and their teams. When employees understand why data is being collected, they will be more willing to provide it.
2. Be transparent – Again, this step is not new to the list of best practices for implementing a time tracking process. Make sure you officially document and communicate how the data will be used (e.g., corporate policies and procedures), and avoid policies that would use the data as a stick. Timesheets are not performance reviews, so it’s important not to confuse the two.
3. Offer control – With respect to the time tracking process, big data and machine learning are especially helpful for improving the speed and ease of filling out a timesheet, as well as dramatically improving the accuracy of the data. The tools to make this happen should go directly to the employees themselves. It should be up to the people filling out timesheets to decide whether and when to use these data analysis tools to make their jobs easier.
4. Give positive feedback – Probably the best feedback you can get as someone who has to fill out a timesheet is, “Hey, I don’t need you to fix anything on your timesheet. It’s perfect!” The elimination of tedious corrections is a big bonus. Beyond the short-term benefits, it is also good to show employees how the improved data produced better decisions that led to direct benefits for them, their teams and the company.
5. Consider individual differences – Big data and machine learning algorithms cannot assume one size fits all. Some assumptions will produce great results for one employee, while providing garbage for another. A good data mining process should cover a variety of use cases and improve over time based on analysis of past results. This includes continual analysis of the different ways that various employees and teams carry out their work and enter their time.
Implementing data mining in your time tracking process shouldn’t make filling out timesheets stressful, or even less popular. Rather, it should reduce (or eliminate) the time spent entering data and increase employee satisfaction.