A New Page for Every Keyword? Check You're Not Sabotaging Yourself First
Familiar scene: you find a new keyword, close cousin to one you already have a page for, and your hand just reaches for "create new URL." The logic seems sound — more pages, more coverage. In practice, it's often a road in the exact opposite direction.
I've seen it over and over: sites cutting their own visibility by breeding duplicate meaning under different phrasings of the same query. And it's not about Google being too lazy to spot the nuance — it's that you're creating competition between your own pages. Google ends up not knowing which one to show, so it either flip-flops between them or shows both, weakly. That's cannibalization. It's not a "not enough content" problem. It's an architecture problem.
Search Console first, decisions second
Before you create a new URL for a similar keyword variant, go into Search Console and check whether the site already ranks for that query. If it does, the question isn't "should I make a new page" — it's "which page is already sitting there, and why." Often it turns out the existing page is already picking up impressions across dozens of similar phrasings. You just never saw it, because you were only watching the one keyword the page was originally built for.
SERP overlap — your main test
Second step: compare the results for your original keyword against the candidate. If the top results share the same competing pages at 70-80%, that's a strong signal — Google treats these queries as the same user need, just worded differently. In that case, building a second page isn't expanding your reach. It's splitting one strong signal into two weak ones. High overlap isn't a "proceed with caution" situation — it's basically a direct answer: don't make a separate page, expand the one you've got.
If the results are genuinely different — different page types ranking, different intent behind the query (say, one keyword is "buy," the other is "compare") — then a separate page is justified. It's serving a different need, not a variation of the same one.
Expand, don't multiply
When keyword variants solve the same user problem, the right move is to fold those phrasings into the existing page: expand a section, add an FAQ block, deepen a paragraph. This isn't "less effective" than a new page — it's exactly what strengthens the weight the page has already earned, instead of splitting it in half.
Thin pages built for every sneeze of a keyword variant are, functionally, doorway pages in new packaging. They used to be made by hand for every city and region. Now they're made for semantic variants of one intent. Nothing's changed underneath: the user and the search engine see almost the same thing wrapped differently, and nobody wins from that.
Architecture is a division of roles, not a page count
A good site is built so that every page has a clear, non-overlapping job. Not "page for keyword A" and "page for keyword B, which is basically the same thing" — pages that solve genuinely different problems for genuinely different audience segments. This applies to content sites, SaaS, e-commerce alike, same principle: if two pages could swap places in the results and nothing would change for the user, you don't have an asset. You have a problem.
Before your next content sprint, spend an hour not writing something new, but auditing: which keywords are already covered, where overlap is high, where pages actually duplicate each other in meaning. It's more boring than publishing something new. But that boredom is what saves you months of fighting your own cannibalization later. If you want to work through this on an actual project, we can go over it during an audit or consultation.