An extension might feel fine on one page and slow down a different kind of site. Measuring only the initial load could miss what happens during normal browsing.
How would you describe the test scope to a specialist? Would you include several extensions running together or start with an isolated baseline?