The same page can be reachable at http and https, with and without www, with tracking parameters, or via a category and a tag. Without a canonical tag, Google may split link equity between those variants or index the wrong one.
A canonical tag is a single line in the <head>: <link rel="canonical" href="https://example.com/page"/>. It tells crawlers 'this is the master version — treat all the others as duplicates of me'.
Every indexable page on a modern site should self-canonicalise. Get it wrong and organic traffic quietly leaks to a URL variant you didn't mean to promote.