I second the suggestion of Mojo::DOM, however if you're trying to scrape google search results I suggest investigating their various APIs rather parsing results.