Save a working version of the code so that you can get back to the "last known good state."

Measure the system to find hot spots.

Determine whether the weak performance comes from inadequate design, data types, or algorithms and whether code tuning is appropriate. If code tuning isn't appropriate, go back to step 1.

Tune the bottleneck identified in step (c).

Measure each improvement one at a time.

If an improvement doesn't improve the code, revert to the code saved in step (a). (Typically, more than half the attempted tunings will produce only a negligible improvement in performance or degrade performance.)