Content pruning has a reputation as a bold move: sort every URL by sessions, select the bottom third, delete. It is a satisfying afternoon and it sometimes works. It also sometimes removes pages that were quietly holding links, ranking for long-tail queries or supporting the pages that do earn, and the loss shows up weeks later with no obvious cause.
The better framing comes from the definition itself. Pruning is a set of five decisions about weak pages, and “delete” is only one of them.
What content pruning actually means
Search Engine Journal’s piece on the subject, written by Manish Dudharejia and published on 22 July 2020, defines it broadly: updating, consolidating or removing content that is low-performing or obsolete, not just deleting it. The full argument is in Search Engine Journal’s guide to content pruning.
The article gives two reasons it helps. Crawlers spend less time on pages that do not matter, and the site lines up better with Google’s emphasis on quality, which has been the recurring theme of core updates. Both are about the site as a whole, not about any single page you remove.
It also reports two examples: HubSpot deleting about 3,000 old posts, and one client seeing a 37% increase in organic traffic after pruning. Treat both as reported examples. They show pruning can work; they do not tell you what it will do for your site.
Why “delete the bottom third” is the wrong default
Low traffic is a symptom, not a diagnosis. A page can have almost no visits for several different reasons, and each needs a different response.
- Content decay: the page used to rank, the facts went stale, and newer pages overtook it. That is a refresh, not a deletion.
- Cannibalisation: two or three pages target the same query and split the signal. That is a merge.
- Wrong topic: the page has nothing to do with what the site is now about. That might be a removal.
- Never had a chance: thin, unoriginal or written for a query nobody searches. Removal or noindex.
A traffic sort cannot separate these. It also misses the page with three visits a month and twenty referring domains, which is doing more for the site than its analytics row admits. Delete that page without a redirect and the links point at a 404.
The content pruning decision table
Every weak page gets one of five outcomes. Pull organic traffic, referring domains and a yes or no on topical fit for each URL, then read across:
| Traffic | Links | Topical fit | Action |
|---|---|---|---|
| Declining | Any | Yes | Refresh: update facts, structure and intent match |
| Low, overlaps a stronger page | Any | Yes | Merge into the stronger page, then redirect |
| Low | Has referring domains | No | Redirect to the closest relevant page |
| Low | None | Needed for users, not search | Noindex and keep |
| None | None | No | Delete and return a 410 |
Notice how narrow the bottom row is. A page only qualifies for straight deletion when it has no traffic, no links and no place on the site. On most sites that is a minority of the weak pages, not the majority.
Gather the inputs over a long enough window. Twelve months of organic traffic smooths out seasonality, so a Christmas gift guide is not marked dead in March. Referring domains come from whichever backlink tool you already use, and topical fit is a judgement call you make page by page, ideally with the site’s current category list in front of you.
Merge and redirect: where most of the gain is
Merging is the underrated option. Sites that have published for years tend to accumulate several half-good pages on the same subject, written at different times by different people. None of them is strong enough to rank; together they would be.
Pick the page with the best links and the closest intent match as the survivor. Move the useful sections from the others into it, rewrite so it reads as one piece, and 301 the old URLs to it. Check internal links afterwards so the site points straight at the survivor rather than through a chain of redirects.
Redirects for off-topic pages should go somewhere a visitor would accept as a reasonable substitute. Redirecting everything to the homepage tends to be treated as a soft 404 and wastes the links you were trying to keep.
Noindex has its own place. Some pages exist for readers who are already on the site: tag archives, thin author pages, internal search results, old announcements. They do no harm to visitors and no good in search, so keep them live and take them out of the index rather than deleting something a reader might still need.
Content pruning SEO after a core update
Pruning usually comes up after a traffic drop, and that is where it is most often misused. A core update loss is a site-level quality judgement more than a page-level one, which is why pruning can help: it changes what the site as a whole looks like. But it is one lever among several. Our guide to recovering from a core update covers the rest.
Before you prune in response to an update, check whether the drop was really about your content. Use the Google updates tracker to see how the update moved your niche. Then, in the explorer, filter to your niche and sort by estimated traffic change across that update, and compare the sites that gained with those that lost: their size, their content depth, how tightly focused they are. If focused sites gained and sprawling ones lost, pruning off-topic content is a reasonable response. If the whole niche fell, pruning will not fix it.
The limits of pruning
The honest limitation is attribution. Pruning rarely happens alone: it comes alongside refreshes, internal linking work and the next core update. When traffic rises afterwards, you cannot cleanly say which change did it, and the reported success stories suffer from the same problem.
That argues for a staged approach. Refresh and merge first, because those are reversible and rarely hurt. Delete last, in batches, and keep a record of every URL, its links and where it was redirected. If something goes wrong, you can trace it.
Give each batch time before judging it. Google has to recrawl the changed URLs and reprocess the site, which can take weeks on a large site. Judging a pruning round after ten days mostly measures noise, and reversing it on that evidence wastes the work.
More on planning content around search sits in the strategy archive and across the blog.
Frequently asked questions
Does deleting old content improve SEO?
Sometimes, but deletion is the smallest part of pruning. Most weak pages do better refreshed or merged. Delete only pages with no traffic, no links and no topical fit, and return a 410.
What is content decay?
A page that used to rank losing traffic as its information goes stale and newer pages overtake it. The fix is a refresh, not removal, because the page usually still targets a query worth having.
Should I redirect pruned pages to the homepage?
No. Redirect to the closest relevant page. Mass redirects to the homepage tend to be treated as soft 404s, which wastes the links the redirect was meant to keep.
The takeaway Prune with a decision table, not a traffic sort. Refresh what decayed, merge what overlaps, redirect what has links, and delete only what has nothing at all.
