If you were tasked with scraping a website that frequently changes its structure, how would you adapt your approach?