Keyword-heavy resume vs plain resume: a small test on 15 applications

Spent an evening tailoring a resume to one posting. I copied every keyword I could find and it read like a robot wrote it. Then I sent the plainer version to ten roles and tailored only the summary line. Results so far are mixed: the plain version got two replies from ten, the heavily matched one got one from five. Way too small to draw conclusions. Has anyone tracked this properly, and does the company size change things?