1. Applying the HR analytics process
Now that you've learned about the general process for HR analytics, you can apply that process to a real-life question: where are your best hires coming from, and where could you get more of them? Success in business relies on how well your workforce performs, and your workforce is largely a product of who you hire.
2. Applying the process to recruiting
Using the first two steps of the general process described in the last video, the groups you'll be comparing are employees who were hired from different recruiting channels,
3. Applying the process to recruiting
and the statistics you'll calculate are how good the hires were, or the average quality of hire for each group.
4. Quality of hire
How do you measure quality of hire? There isn't one accepted way to compute it for every type of employee, but you can usually use metrics such as retention, hiring manager satisfaction, job performance, or "time-to-productivity", which is the amount of time it takes for the employee to become fully productive.
How exactly you combine these measures together into a single metric is something to discuss with stakeholders, including those in your talent acquisition department. In this chapter, you'll be working with these four variables. One is a measure of attrition, and performance_rating and sales_quota_pct are measures of job performance.
5. Calculating the attrition rate
Attrition rates can be calculated in a few different ways. For this course, you'll only be looking at one snapshot in time, so you can use a simple formula - number of employees who left the company, or attrition, divided by total headcount.
In the dataset for this chapter, attrition is coded as 1 if the employee left, and 0 otherwise. This means that an equivalent formula for attrition is to take the average of the attrition variable. This is the formula you'll use in the exercises.
6. Review of tools from dplyr
The analysis in this course will build on the tools introduced in the introductory course on the Tidyverse. These functions fit nicely into the pattern of identifying groups with group_by() and summarizing information about those groups with summarize().
In this example, you can see that group_by() is used to split the recruitment dataset into groups, and then summarize() is used to apply a function - the maximum of performance rating - to each recruiting source. arrange() is used to sort the result by highest performance.
7. New tools
This course will also introduce some new tools from the tidyverse. One of them is count(). The count() function from dplyr returns the number of rows corresponding to each member of the specified group. If you don't specify a group, count() returns the total number of rows in the data. This function is useful for exploring categorical data. Here, count() tells you which recruiting sources are in the data, and how many applicants were hired from each one. If a recruiting source is missing from your dataset, count() can help you catch that.
8. Let's practice!
Now it's time for you to find where the best hires come from.