What five blogs got from content pruning, and two did not

QuickBooks deleted just over 2,000 posts from its Resource Center, more than 40% of the whole library, and organic traffic rose 44% by the next tax season.

That is the result everybody quotes when they argue for content pruning. It is well documented, and it is also the best case in a set of published results that includes two sites that saw nothing at all.

We have not run a pruning project on our own blog that we can report numbers on, so this article does not pretend to. Instead it takes the published cases, puts their figures side by side, and then sets out the method we use when a client asks whether to cut.

What the published content pruning cases actually show

Pruning helped most where a site had thousands of pages and a large share of them got no search traffic. Where the cut was small, or the pages were already ignored, it did very little. Here are the published cases with figures attached.

Two things stand out. The sites that moved the needle cut a large proportion of their pages, and the sites that cut a sliver (RecipeLion's 3 to 4%) got nothing. Kaylee Pope at Prime Publishing put it in three words, "literally, nothing changed", which is a more useful data point than most of the success stories.

The second thing is attribution. Brafton's 108% rise came alongside a new blogging programme. HubSpot's Aja Frost says plainly that pruning was one of several growth efforts and gives no traffic figure at all. The QuickBooks case is the cleanest because the 20% jump came within weeks of the deletion, before anything else had time to act.

Will Waggoner, who ran the QuickBooks project, is honest about the conditions it needs:

"The sites that benefit most from pruning have thousands, if not 100,000 or more pages." Will Waggoner, SEO lead at QuickBooks, via Animalz

For a UK business with a 60-post blog, that line matters more than the 44%. Pruning a small blog is housekeeping. It will not double your traffic, and anyone who promises it will is selling the QuickBooks number without the QuickBooks conditions, like the other SEO services that are mostly theatre.

Why deleting pages can raise traffic at all

Deleting pages raises traffic when it removes pages that were competing with, diluting, or burying the pages that could rank. There is no credit for being smaller. Google's own helpful content guidance, updated in October 2026, lists "removing a lot of older content primarily because you believe it will help your search rankings overall" as a warning sign, and adds in brackets: "No, it won't."

So the mechanism is not freshness. The mechanisms we can actually see in the cases are these.

1. Internal link equity. QuickBooks had thousands of internal links pointing at posts nobody read. Waggoner's view was that cutting those links "gave the remaining ones even more equity".
2. Cannibalisation. When four posts answer the same query, Google picks one, and often not the one you would choose. Merging them gives it one clear candidate.
3. Crawl and indexing speed. HubSpot saw crawls fall 20% and URLs crawled fall 38%, and new posts started getting indexed within minutes instead of hours. Google says crawl budget is mainly a concern for sites with over 10,000 pages that change daily, or over a million pages, so for most UK SMEs this third effect is marginal.

If your site has 80 pages, the first two still apply. The third does not, and a good chunk of what you read about pruning is written for sites a hundred times your size.

Step one: the inventory in Search Console

Start with a 12-month export of the Search Console performance report, broken down by page, with clicks and impressions. Twelve months, not three, because the performance report defaults to the last three months and a three-month window hides every seasonal page you own.

Set the date range, switch the table to Pages, and use the export button. Put the result in a spreadsheet with one row per URL. Then add three columns from other tools: external backlinks pointing at each URL (from Ahrefs, Semrush or Search Console's own Links report), internal links to the page (from a Screaming Frog crawl), and the date the page was last substantively updated.

HubSpot's audit added one more filter that is worth copying. Before judging a page on organic traffic, it removed from consideration any page with over 1,000 users from all sources in the period. A page that gets no clicks from Google but is linked from your email footer or your sales deck is not a candidate for deletion, whatever Search Console says.

Now sort by clicks. On most blogs we audit, the shape is the same: a short head of pages that earn nearly everything, then a long flat tail. The tail is where the decisions get made.

Step two: the four buckets

Every page in the inventory goes into one of four buckets: keep, update, merge, or remove. The point of forcing a choice is that "leave it for now" is not an option, because leaving it for now is how the tail got there.

The thresholds depend on the site. QuickBooks used under 100 visits in six months plus no conversions. HubSpot used under 450 organic users across the period, which it worked out as under 80 organic views a month. RecipeLion used under 10 views in two years, which is so low that it only caught pages Google had already given up on, and that is probably why nothing changed.

Our rule of thumb for a small UK business blog is to look hardest at pages with under 12 clicks in 12 months. One click a month is not a page that is working, but it is not automatically a page to delete either. Run it through the flowchart below first.

flowchart TD
    A[Page in 12-month export] --> B{Clicks in top half of site, or steady impressions?}
    B -- Yes --> K[Keep]
    B -- No --> C{External backlinks, or linked from sales or email material?}
    C -- Yes --> D{Still accurate and on-topic?}
    D -- Yes --> K
    D -- No --> U[Update at same URL]
    C -- No --> E{Another page on the site covers the same queries?}
    E -- Yes --> M[Merge into the stronger URL and 301]
    E -- No --> F{Seasonal, or published in the last 6 months?}
    F -- Yes --> W[Wait, re-check next audit]
    F -- No --> G{Topic still relevant to buyers?}
    G -- Yes --> U
    G -- No --> R[Remove: 301 to nearest page or 404]

Step three: redirect rules that hold up

Every merged or removed URL that has links, bookmarks or traffic from anywhere gets a 301 to the single most relevant surviving page, and that redirect stays in place for at least a year. Google's site move guidance says to keep redirects "for as long as possible, generally at least 1 year", and in practice there is no reason to ever remove them.

The target has to be a genuine replacement. Google's consolidation documentation describes a redirect as a strong signal that the target should become canonical, and says to use it "only when deprecating a duplicate page". A post about VAT thresholds redirected to a post about payroll is not a duplicate; it is a page Google will treat as gone.

For pages with no links, no traffic and no close relative, a 404 is fine. Google's HTTP status guidance says indexed URLs returning any 4xx code are removed from the index, and that it treats 404 and 410 the same. Do not redirect orphans to the homepage in bulk. QuickBooks did this for pages with no natural destination and got away with it on a very strong domain; on a small site it reads as a soft 404, which Google says will "continue to be crawled, and waste your budget".

Three housekeeping rules go with every redirect batch:

• Update internal links to point at the survivor, so you are not routing your own navigation through redirects. HubSpot wrote a script for this.
• Avoid redirect chains. If page A already redirects to B and you now retire B, point A straight at C. The same rule is in the technical SEO afternoon.
• Release in batches and watch for a week between them. HubSpot's developer ran the archive and redirect scripts in stages with about a week of traffic monitoring in between, and that is the sensible pace for a site a fraction of its size.

The mistakes we see most often

The two pruning mistakes that cost real traffic are deleting pages that other sites link to, and removing seasonal pages because they looked dead in the wrong month.

Links first. HubSpot pulled every page's backlink count before deciding anything, and any page with 60 or more referring links (about 350 of the 3,000) was reviewed by hand instead of by a freelancer working off a rule. A page with ten links from trade sites and no traffic is a page to update or merge, never to 404, because the 404 throws the links away. Dario Zadro's ecommerce client in the Orbit Media round-up set about 4,000 product pages to 404 and watched the number of ranking keywords fall, even as the surviving pages ranked higher.

Seasonal pages second. A post on Christmas opening hours, Self Assessment deadlines or summer holiday childcare will show nothing in a three-month window pulled in June. This is why the inventory uses 12 months, and why the flowchart has a "wait" branch. The QuickBooks team excluded new content that had not had time to rank for the same reason; a six-month-old post is not a failure yet.

Two more, briefly. Changing dates on updated pages that have not substantively changed is something Google's helpful content page names directly as a search-engine-first practice. And pruning without a baseline means you cannot tell what happened; it is the same discipline as reading outcome reports rather than activity reports. Take a screenshot of the 12-month clicks total on the day you start, then compare the same window a quarter later. Taylor Cimala at Digital Third Coast ran a prune at the same time as an HTTPS migration and a domain change, and now cannot say what any of the three did.

When pruning is the wrong tool

Pruning is the wrong tool when the blog is small, when most pages get at least some clicks, and when the real problem is that nothing on the site is good enough to rank. Cutting 15 posts from a 60-post blog will tidy the sitemap and change little else; the fix is usually the right 15 service pages, not fewer posts.

The published cases are consistent on this. The gains came from cutting 15% to 50% of large libraries where the tail was genuinely dead. A site in that position usually also has a cannibalisation problem and a crawl problem, and pruning fixes all three at once. A small site in that position just needs better pages, which is a content marketing problem, not a pruning one.

If you are not sure which situation you are in, the 12-month export answers it in ten minutes. Count the pages with under 12 clicks. If that is under a fifth of the blog, update the weak ones and move on. If it is over half, you have a QuickBooks-shaped problem and the method above is worth a weekend.

The rule we apply: delete a page only when you can name the page that should rank instead. If you cannot, the problem is not that you have too many pages.

Where content work sits in a full retainer is set out in our guide to what a UK SEO agency should cost and deliver. If you would like the inventory run and the four buckets filled for your site, our SEO audit covers it, including the redirect map and the technical checks that go with it. Get in touch and tell us how many pages you have.