Content Pruning Case Study: We Removed 103 Pages

Key takeaways

• On 21 September 2026 we removed 103 zero-traffic pages from LiveAgent, which is 2,784 URLs across 27 locales. Daily impressions in Google rose to roughly three times their earlier level within days.
• The results are early. We want a few more weeks of data before we call it a win.
• Google says deleting content just because it is old is not a thing. Remove pages that are unhelpful, not pages that are old.
• An index bloat report finds pages that were shown but earned little or no traffic, which makes the first list much faster to build.
• Work in rounds, measure each one, and decide page by page whether to keep, improve, or remove.

Table of contents

• What is content pruning?
• What did we do in this content pruning case study?
• What happened after the first round?
• How does an index bloat report find pages?
• What does Google say about removing pages?
• How do you decide what to remove?
• What would count as a result?
• When should you not prune?
• How Web25 helps
• FAQ

This content pruning case study is short and honest. Most websites only grow. Every year brings new articles, new landing pages, and a few pages that say almost the same thing as older ones. Few teams ever try it. We tried it on LiveAgent, a website we work on, and below is what we did, what we saw, and what we still do not know.

What is content pruning?

Content pruning is the practice of removing, merging, or rewriting pages that no longer help your site. These are usually old articles nobody visits, pages that cover the same topic as another page, and pages that have had no impressions in search for months.

The reasoning is simple. A page that never appears in search and never gets a visit is not helping a customer. It only makes the site bigger, not better. Deciding what to keep is part of running a good site, just as deciding what to publish is. Most editors spend their time on the second decision and almost none on the first, which is how a site slowly fills with pages nobody wants. A regular review, even twice a year, keeps that from happening.

What did we do in this content pruning case study?

On 21 September 2026 we removed 103 zero-traffic pages from LiveAgent. Each page exists in 27 locales, so that came to 2,784 URLs. The idea came from Viktor Zeman, who suggested that pages with no impressions for months are not helping anyone. We agreed, and we went through the list section by section instead of cutting in one big sweep. A redirect map was set in the same pull request, so removed pages sent visitors to a useful page. This was the first round.

To build the list, we used the index bloat report in AmICited, a tool from FlowHunt. We used it to export and filter the pages, and it saved us a lot of time. Working through a site of this size by hand in a spreadsheet would have taken far longer.

The decision itself was still ours. A report can show which pages earn little or no traffic. It cannot tell you whether a page is worth keeping for another reason, such as a service page that converts, or a guide that earns links. We reviewed each section with that in mind before anything was removed.

What happened after the first round?

Impressions in Google went up fast. Reading the Search Console chart for LiveAgent, daily impressions sat at roughly 100,000 to 125,000 for most of the summer. By the end of September they reached about 375,000 a day, and daily clicks rose from around 800 to around 2,400. These figures are read off the chart, so treat them as approximate.

The timing is close but not clean. The climb started around 19 to 22 September, so it overlaps the removal date, and we cannot separate the removal from other causes such as seasonality, new content, or a change in how Google shows results. We treat this as a promising sign, not a finding. We are watching impressions, clicks, and indexed pages in Google Search Console over several more weeks and comparing them with the same period before the change.

If the trend reverses, we will say so. Web25’s guide to measuring AI Overview traffic covers how to read these Search Console numbers carefully.

How does an index bloat report find pages?

AmICited describes its index bloat report as a way to find URLs that were shown but earned little or no traffic, or that stopped appearing. It sorts them into two groups. Zero-click URLs were shown in the chosen window but fell below a click threshold. Went-dark URLs were seen before but are absent in the window.

You can adjust the thresholds, such as minimum impressions, maximum clicks, and the date of the last impression. The report lists each URL with its impressions, clicks, current position, and the dates it was first and last seen, and it can show Google’s URL Inspection result. It keeps Google and Bing apart.

The report does not delete anything. Its recommendation for each page is to keep, improve, or review. That fits the way pruning should work. A tool narrows a huge site down to a list, and a person makes the call.

What does Google say about removing pages?

Google is clear about one thing. In August 2023, Google’s Danny Sullivan said on social media that deleting content because you believe Google does not like old content is not a thing. Older content can still be helpful.

He also added that a page unlikely to rank well may be fine to remove, and that on a massive site removal could help Google crawl other content better. So the line is about why you remove a page. Removing a page because it is old does nothing. Removing a page because it is unhelpful, duplicate, or never seen is a different matter.

That is the test we used. A page does not go because of its age. It goes because it has had no impressions for months, covers ground another page already covers, and helps no visitor.

How do you decide what to remove?

Treat every page as one of three outcomes: keep, improve, or remove. Start with the pages that have had no impressions for a long period, then check each one against a short list.

• Does the page still get clicks or impressions? If yes, leave it alone or improve it.
• Does it earn links from other sites? If yes, keep it or redirect it to the closest page.
• Does another page cover the same topic better? If yes, merge or redirect.
• Does it serve a purpose that analytics will not show, such as a legal page or a sales page? If yes, keep it.
• Is it thin, outdated, and unlikely to be fixed? If yes, remove it.

Work section by section, as we did. A small round is easy to review and easy to reverse if you find a mistake. Do not remove a whole group at once just because the report lists it.

What would count as a result from content pruning?

Before you start, decide what success looks like, because impressions alone can mislead. A rise in impressions is a good first sign, but it is not the goal. The goal is a site where the pages that remain get found and used by real people.

We are watching three things over the coming weeks. The first is impressions and clicks across the pages that remain. The second is the number of pages Google reports as indexed, which should fall as removed pages drop out. The third is whether our key pages, the ones tied to real enquiries, hold or improve their position. If impressions climb but the key pages slip, that is a warning, not a win.

Write these checks down before the first round. It stops you from picking whichever number looks best afterward. It also gives you a clear reason to continue or to stop, which matters when you plan another round.

Common mistakes to avoid

The most common mistake is removing pages in bulk without a review. The second is judging the result after a few days, when search data is still settling. The third is skipping the baseline, so there is nothing to compare against. A fourth is treating it as a one-off clean-up, when it works better as a regular habit, done in small rounds.

When should you not prune?

Do not start content pruning on a site that is small and young. If you have only a few dozen pages, improving them is usually a better use of time than removing any. Do not prune pages that earn links or bring even a trickle of qualified visitors. And do not prune when you cannot measure the result, because you will not know whether it helped.

Also be careful with expectations. Our own content pruning case study is early, one site, and unproven. Content pruning is a way to keep a site tidy and useful. It is not a shortcut to rankings, and nothing here promises that it will raise yours.

How Web25 helps

Web25 builds and maintains websites, and LiveAgent is a site we work on. We link to FlowHunt and AmICited because they are the tools and the idea behind this round, and we credit Viktor Zeman for the idea. Weigh that when you read our view, and test the approach on your own data. If you want help reviewing a large site, see Web25’s website maintenance plans or the notes on how AI is changing search.

FAQ

What to do next

Open Google Search Console, find the pages with no impressions in the last several months, and sort them into keep, improve, or remove. Start with one section of your site. Then book a discovery call if you want a second pair of eyes on the list.

---

Written by Adam Khaled, Founder, Web25. Last updated October 2026.