[concretemarble]

Your logic is flawed in that you didn't clarify the definition of "small". How small are we talking about? This is a simple statistical question.

Calculate the number of sexual harassment reports of all US companies normalized by employee counts, and check if the number of sexual harassments at Blizzard is statically significantly higher. If so, then it is plain reasonable to suspect Blizzard has certain issues that cause this.

Whether the females are broadly and uniformly harassed is not a good indicator. As long as the rate is significantly higher than other company, it could indicate some problem and is worth investigation, even if the total number of harassment cases are small compared to the total female employees.