# Metrics for changing a process

I am trying to automate a manual process in my organization. There's one component which has a very repetitive coding. So i decided to automate it. Before jumping on to the automation, I want to quantify my reasons for automation. So i decided to come up with a set of metrics.

### Problem

Consider 3 sub-components within the main component.

Every sub-component can have two parameters:

• Frequency of repetitiveness of code(rated 1-5; 5 being highest) and
• manual effort (rated 1-5; 5 being highest).

Depending on these two parameters, I can devise an automation index.

### Some facets of the question:

1. how to devise the automation index which gives me a rating so as to help automate a particular sub-component?
2. Do i multiply or add the values from frequency and effort?
3. Are these parameters sufficient for my scenario?

• I think you need to also share your equation – New Alexandria Dec 9 '14 at 14:57

I want to quantify my reasons for automation

The best reason for automation is cost reduction. This is your ultimate metric.

Calculate how much time your automation + execution takes, and how much time execution without automation takes. The difference is your reason. If you can show how automation provides cumulative settings over time, you'll get a management buy in without problems.

I would expect automation and execution effort to be higher than normal work in the beginning, and decreasing over time first to the zero balance, and then to the savings.

And yes,

So i decided to come up with a set of metrics.

If you cannot make sense of metrics, ditch them.

If I got your question correctly, you are trying to quantify the quality of the program code. That's better to do with automated code quality tools, like Sonar:

http://www.sonarqube.org/

You can set it up and see what indexes of code duplication you'll get for the components you mentioned. You can use the indexes as justification for code optimization.