The following steps will help to guide efforts to implement a performance improvement. The key is to only make a small change with each iteration and run a performance test to ensure that change resulted in an improvement.
Establish a performance target (goal posts)
Build an automated performance test
Run the performance test to establish a baseline (the current performance data)
Make one change to the process
Run the performance test again to measure the impact of the single improvement against the baseline
If the change results in a measurable performance improvement, then keep it
Repeat steps 4 to 6 until the performance target has been met
For bonus points when you're ready to deploy to production:
Run the performance test against production to establish a production baseline
Deploy the changes to production
Run the performance test to measure the impact of the improvements
Provide the performance improvement results to your Product Owner and bask in the shower of compliments coming your way for a job well done!
Working against a baseline and having a defined target will ensure that you are not prematurely or over optimizing your process.