
How Readers Recognise an Item
Every time a reader fetches a feed, it compares the items with the ones it already has. To do that it needs something that stays the same for each item. In order of preference, readers use:
- the item's ID:
<guid>in RSS,<id>in Atom,idin JSON Feed; - if there is no ID, the link;
- if there is no link either, the title, sometimes combined with the date.
If any of these changes for an existing article, the reader sees a new item. It has no way of knowing that "the same article with a different ID" is the same article.
The Usual Causes
| What happened | Why it creates duplicates |
|---|---|
| The site moved to HTTPS or a new domain | IDs that were permalinks, such as http://example.com/post-1, became https://…. Every item in the feed looks new at once. |
| A new CMS, theme or plugin | The system builds IDs differently, for example from a database number instead of the URL. |
| Changed permalinks or slugs | Feeds without IDs fall back to the link, so a renamed URL is a new item. |
| Edited headlines | Feeds without IDs or links fall back to the title, so a corrected typo is a new item. |
| Tracking parameters in links | Links like ?utm_source=rss&utm_campaign=… that change per day or per fetch make every fetch look new to readers that use links. |
| Random or time-based IDs | IDs generated at request time, such as a hash of the current timestamp, are different on every fetch. |
| The same post in two feeds | You follow both the main feed and a category feed of a site. That is not a bug: the same item really is in both. |
If You Are the Reader
- A burst of old posts once usually means the publisher migrated. Mark them as read; it should not happen again.
- Duplicates on every update point to unstable IDs or changing links. Run the feed through the health check: it reports items without an ID and IDs used twice. Telling the publisher what you found is the only real fix.
- The same story from two feeds of one site: unsubscribe from one, or combine them with the feed merger, which removes items with the same link or ID and identical headlines.
- The same news from different publishers is not a duplicate in the technical sense. Each outlet wrote its own article. Some readers can group them; a keyword filter can reduce the volume.
If You Publish the Feed
The rule is simple and absolute: give every item an ID once, and never change it. The safest choice is an ID that does not depend on your domain or URL structure:
<guid isPermaLink="false">tag:example.com,2026:post-1842</guid><id>urn:uuid:2b5f1c7e-8a4d-4e0b-9a3f-6f1d2c9e7b10</id>If your IDs have always been permalinks and you now move to HTTPS or a new domain, keep generating the old value for existing items. In WordPress the guid is stored in the database when a post is created and does not change when you switch domains, unless a search-and-replace tool rewrites it: exclude the guid column when you migrate.
Also:
- keep tracking parameters out of item links, or make them identical on every fetch;
- do not change the date of an item when you fix a typo, unless you want readers to see it as updated;
- check the feed in the validator after every migration, before readers notice.
More publishing advice is collected in RSS best practices for publishers.
Questions
Why does my RSS reader show duplicate posts?
Because the reader thinks they are different items. That happens when the publisher changed the IDs, links or titles of existing items, for example after moving to HTTPS, switching CMS or adding tracking parameters to links. It also happens, without any bug, when you follow two feeds of the same site that both contain the article.
Can my reader remove duplicates for me?
Some readers merge items with identical links, and a few cluster articles about the same event. Most do not, because deciding that two different items are "the same" is a guess. If the duplicates come from overlapping feeds, combining those feeds into one with duplicate removal is the most reliable fix.
Is isPermaLink="false" required?
Only when the guid is not a URL. By default RSS treats a guid as a permalink, so a value like tag:example.com,2026:post-1842 needs isPermaLink="false"; otherwise some readers try to open it as a link.
Sources and Further Reading
- RSS 2.0 SpecificationRSS Advisory Board
- RSS Best Practices ProfileRSS Advisory Board
- RFC 4287: The Atom Syndication FormatIETF