Perform a Logistic Regression to estimate the minimum bonus needed to reach 75% of a productivity target
Defining productivity incentives in variable costs of operations is a tricky endeavor.
The productivity target to be reached can directly impact the P & of the business.
One way is to use machine learning to help.
All based on the nature of the task.
The challenge is to find the right incentive policy to reach 75% of the repetitive task’s objectives.
A repetitive task could be:
- Picking
- Driving
- Packing
- Mixing
Finding the right incentives Policy Let us say you have operators who earn $5 per day bonus on a daily $64 per day salary.
This incentive structure is only resulting in 20% of the operators achieving their targets.
What should the daily bonus be to reach 75% of the target?
We could run a machine algorithm experiment by:
- Randomly selecting operators
- Implement a daily bonus varying from $1 to $20 (Incentives)
- Check to see if operators reached their target. (Target)
We can use logistic regression to find the relationship between the two data factors.
- Incentive
- Target
Then use that relationship to predict the value of one of those factors based on the other through machine learning.
Based on the data from the code in my experiment we get to a minimum amount of $15 bonus per day to reach the 75% productivity target.
However, as fancy as this process is, it is not a complete process.
Machine learning, and AI in general, are only aids to a problem not the solution.
A deeper analysis should be taken to add context to the machine algorithm’s output.
Total cost to the company (Salary + contributions + taxes) What are the trade-offs on the P & L when implementing such a new incentive bonus?
This repository code you will find all the code used to explain the concepts presented in the article.
Supply Chain Professional with international experience and a passion for data science. Connect with me on LinkedIn: Profile Profile