From a company’s prospective, employees are fixed cost resources. While they’re working, they’ll contribute to the product that they’re working on, but what about when they’re not working? A simple way to test is to come up with estimates for a product, use a weighted system such as a system which factors in experience, accuracy in estimation, and etc., to come up with the estimates. Then you apply changes to that worker and see what the effects are. If you see an increase in productivity keep it, if you see a decrease in productivity remove it. Now in order for the experiment to properly execute, you need to control for all the other factors that might affect the worker. As in, you need to prevent anything other than the said changes to affect the worker, this way, if there is a change, you know exactly what caused it. If you introduce too many variables at once, you have no way to figuring out what contributed to what. The combination of factors that might’ve worked for one worker might not work the same way for the next, which is why in order to best figure out what works and what doesn’t, you need to limit the things you’re introducing.
I think open communication is key to employee optimization, you need to figure out what works best for the employee. If at any time the employee feels that there are reasons for him to not openly communicate, then all forms of communication will start failing, and you’ll no longer have the insight to optimize the employee as you did before.