Two of your pages rank for the same query, and every checklist says merge them. Merging usually helps. It is not free: you trade two middling positions for one stronger page, and you give up whatever the weaker page was quietly matching on its own.
What actually happens to rankings when you merge two pages?
The surviving URL inherits the retired page's links and relevance signals through the redirect, and usually settles at or a little above the better of the two old positions. Google's long-standing guidance is that a 301 passes signals to the target URL rather than discarding them, so the mechanism here is consolidation, not loss.
The timing is the part people misread. Google has to recrawl the retired URL, see the redirect, and then re-evaluate the survivor across every query the two pages used to serve between them. On a thin page with few inbound links that resolves quickly. On an older URL with years of link history it runs for weeks, and positions wobble in the middle of it. A dip three days after a merge is not evidence the merge was wrong.
Which page should become the survivor?
Pick the page with the stronger link profile first, then the better match to the query you actually want. Links are the slowest signal to rebuild; a title and an on-page rewrite take an afternoon.
| Signal | Favors keeping | Why it weighs that way |
|---|---|---|
| Referring domains | The page other sites already link to | Page authority built from external links is the one input you cannot edit into existence |
| Query match | Either — it is cheap to fix | Titles, headings and body copy move in one edit, so do not let them decide the survivor |
| Conversion path | The page that converts the traffic | Traffic you keep still has to do something; the better-ranking page is the wrong winner if it dead-ends |
| URL and depth | The shallower, cleaner URL | Shared by readers and easier to reason about later |
| Crawl history | The older URL, all else equal | Already crawled on a known rhythm, so the merge settles faster |
A merge is a bet that one page's authority plus the other's coverage beats two partial pages. Judge the bet against the combined clicks of both URLs before the merge, never against the survivor's own history.
What gets lost that nobody plans for?
Three things, and only the first one is on most checklists.
- Internal links to the retired URL. The redirect keeps them working, so they are easy to leave alone. Repoint them anyway: every hop is a thing that can break in a future migration, and links aimed at a redirect tell crawlers the old URL is still the real address.
- The weaker page's long tail. The second page often ranked because of one specific paragraph — a pricing objection, a regional detail, an edge case. If that paragraph does not survive the edit, those queries do not transfer to anything. Pull the weaker URL's query list out of Search Console before you delete it and check the merged draft against it.
- Whatever the retired page linked to. If it was the only internal link into some other page, retiring it can leave that page an orphan page with no path in. Crawl the outbound internal links on the page you are retiring, not just the inbound ones.
Merging two URLs and then comparing the survivor's traffic to its own previous month. The survivor should go up, because it is now carrying two pages' worth of demand. The only honest comparison is both old URLs added together against the one new one.
How do you tell whether the merge worked?
Add the clicks and impressions of both URLs for a period before the merge, compare them to the survivor over an equal period after, and give it long enough that the recrawl is finished — a month is usually the shortest window that means anything on a page with real link history.
Then look past the total at the query list. A successful merge shows the survivor holding the queries both pages had and picking up a few mid-tail terms neither could reach alone. A merge that failed shows a smaller query list: the survivor kept the head term and dropped the specifics. That is a content problem, not a redirect problem, and the fix is to put the missing material back rather than to undo the redirect.
When is leaving both pages alone the better call?
When the two pages answer different questions that only look alike in a keyword tool. A comparison page and a pricing page can share most of their vocabulary and still serve people at different points of a decision; merging them produces one page that interrupts both jobs. The same goes for a landing page built for paid traffic sitting next to an organic article on the same subject, where the paid page's whole value is that it does not read like an article.
The useful test is whether one page could satisfy both searches without getting longer. If the merged version only works by stacking two arguments end to end, you do not have two overlapping pages. You have two pages with overlapping words.
What to decide before you touch the redirect
Write down two things: the query you expect the survivor to win, and the combined clicks of both URLs over the last month. The first tells you which page to keep and what to cut. The second is the only number that can tell you later whether this was worth doing. If you cannot produce either one, the merge is housekeeping rather than strategy — and housekeeping is a reason to wait until you have something better to spend the recrawl on.