Consolidate two mid-performing pages when they target the same intent, and keep them separate when they serve genuinely distinct intents. The pivot is shared-intent versus distinct-intent, and it is the only thing that should decide this. When two pages chase the same query, they split the authority, links, and relevance signals that should be pooled behind one URL, and they often compete with each other in the results so that neither ranks as well as a single merged page would. One stronger page beats two middling ones precisely because the strength stops being divided.
Cannibalization is the symptom to look for. If both pages surface for the same searches, swap positions over time, or read like two attempts at answering one question, they are fighting over the same intent and the consolidation case is strong. Merging them lets the combined content, links, and history concentrate on one page that can actually win the query, instead of two that each fall short because the signals are spread thin.
Keeping them separate is correct when each page answers a genuinely different need that can win on its own. Two pages that look related can still serve distinct intents, an overview versus a how-to, a comparison versus a buying decision, a general topic versus a specific sub-question with its own demand. Forced into one page, those distinct intents get muddied and the merged result serves each searcher worse, so they belong apart. The test is not how similar the titles look, it is whether a single searcher would be satisfied by one page or whether two different searchers each need their own.
So before merging, check intent overlap directly: look at the queries each page ranks for and ask whether they are answering the same question or two different ones. If they share intent and compete, consolidate them into one stronger page and redirect the weaker URL. If each serves a distinct intent that can win on its own, leave them separate and sharpen each one toward its own searcher.