A comparison table earns its own page when the comparison itself is the search intent, and it belongs inside an article when it supports a broader topic the reader actually came for. The test is what the searcher is after. If people are explicitly searching “X vs Y” and there is real demand for that exact comparison, the comparison is a destination and deserves a page built around it. If the comparison is one piece of evidence inside a larger question, it is a supporting element and should sit in the article that answers that question.
The pivot is is-the-comparison-the-destination versus a-supporting-element. To apply it, check whether the comparison has its own search demand worth being a page. A query like “product A vs product B” signals that the comparison is the whole reason for the visit, so a dedicated page can fully serve it, rank for it, and act as a clear answer to a specific intent. By contrast, a buyer’s guide or how-to where a table helps the reader weigh options along the way is using the comparison in service of a larger goal, and pulling that table onto its own page would strand it away from the context that makes it useful.
This is why neither blanket rule works. Always making comparison pages spawns thin pages for comparisons nobody searches as a destination, splitting a topic that should have stayed whole. Treating tables as only ever supporting elements buries a high-demand “X vs Y” comparison inside an article where it cannot rank cleanly for the intent it would own on its own page. The format choice follows the demand, not a habit, so you decide per comparison rather than by a default.
So before you place a comparison, check whether “X vs Y” is itself a searched intent with enough demand to be a destination. If it is, give it its own page built to answer that comparison directly. If it only supports a broader topic, keep it inside the article where the reader needs it. Let the searched intent decide, and the table lands in the right place.