You're the director of human resources at a mid-size tech firm. The company's founder comes to you one day complaining, that some of the company's best engineers have been leaving for other companies, taking some good ideas for new products with them. How do you stop the bleeding? The first thing you want to know is, why people are leaving. What's more attractive at other companies? More pay? Better benefits? Nicer offices? Shorter commute? Maybe if you knew specifically why they think they'll be happier at another company, you can do something to make them happier at your company. But by the time you know that someone is unhappy, it's too late, they're gone. You need to know that a person wants to leave before they know it. How? Through data science. With the help of data scientists, you can create a predictive model that looks at a variety of factors about an employee and flags those individuals who are likely to be thinking about leaving. You start by identifying employee characteristics, such as the number of years worked in your company, education level, salary, benefits, number of promotions received, health issues, and so forth. The specific characteristics are probably different for each type of company or industry. Next, you compare the characteristics of people who leave with those of people who stay. Your statistics software can build a model that takes in the historical data, and identifies those specific characteristics that are associated with an employee leaving. If your data gathering capabilities are really good, you might even be able to look at their social media posts to see if there are mentions of being unhappy or wanting to work somewhere else. It takes some time to train the model. But soon, you'll be able to predict with reasonable certainty who is next likely to go. You can then offer these individuals some incentives to stay. More pay, trips to conferences, opportunities to publish their work, onsite childcare, whatever is appropriate for their needs, and your particular industry. This is just another way, how the marvels of modern data science can help your company be successful.