Our goal is to find how different plugins affect performance of WordPress website. The more plugins you have - theoretically - the more complex plugin execution is. Therefore we should not judge plugin performance without looking into "how they work and interfere with other plugins". That's why I conduct test on an empty WordPress site with no other plugins first and then - same test on the WordPress site, that has many pluings installed. Both sites are located on the same server, they run same PHP version (unless we are benchmarking different PHP versions).