Never optimise before need

Code development should favor early deployment and code maintenance over performance until optimization is proven necessary by actual measurement.

The old saying that 90% of the code only runs 10% of the time is very true. I will add that you would be surprised where the 10% that runs 90% of the time actually is in a software project. So much of the code is contributed by the OS and libraries that you just can't be sure where optimization is needed until you prove it.

Earlier deployment and more maintainable code is better than perfectly optimized code.

Make it work as best as possible, just well enough to deploy, deploy then start finding where optimization is needed. Refactor to match performance goals.