So Outside Technology Security CMS
AnQiCMS
Online Demonstration

After the garbage update, the fluctuation of rankings, first check which groups of site data

📅 2026-11-01 👁️ 0

Rubbish update after ranking fluctuations, first check which groups of site data? First check the收录面, then check the content, finally check the word ranking. First look at the two things verified that day. One, the garbage update in September is still fermenting in October, and the site master reported it as the main change this period; two, Google is contributing to the pilot settlement for artificial intelligence, about a hundred publishers get a return on the ad revenue of about one percent. After the two things are added together, the reasons for the fluctuations are divided into two categories: 'judged low quality' and 'display form changed', and the data needs to be checked separately.

First clarify the two reasons for the fluctuations.

The garbage update action targets low-quality and manipulative behavior. After being hit, the performance is a decrease in page index or the visibility of the entire site decreases, not that a word has dropped two places. The change in display form is different: the answer position, artificial intelligence summary, and reference links increase, but the page is still indexed, but the clicks are absorbed by the upstream.

The entry points for the two types of reasons are different. The former needs to check the health of crawling and indexing, and the latter needs to check the coverage of references and original content. It is easy to draw the conclusion that 'the algorithm has punished me' when mixed together, but in fact, it may just be a change in the display structure.

The first group of data to check: inclusion and crawling

View the four items in order:

Is the crawling status abnormal. Has the robots.txt configuration been changed recently, and have the rules mistakenly sealed the directory; can the addresses in the sitemap still be accessed normally?

The proportion of invalid addresses. Error pages and redirect addresses mixed into the sitemap will unnecessarily consume crawling quotas, pushing the pages that truly need to be recrawled to the end of the queue. This is especially important during the garbage update period, as crawling allocation is already compressed.

The difference between the index volume and the submission volume. The number of submitted addresses is much greater than the number of indexed pages, indicating that a large number of pages have been deemed of low value; the difference suddenly expands, usually corresponding to the time point when content is batched online.

The structure of internal entry. Anchor text management is responsible for automatic linking of keywords within the site, if the word table points a large number of words to the same few pages, the rest of the pages become islands, and the natural distribution of crawling decreases.

The data group checked again: content and words

The keyword library is the destination for this step. The keyword library management focuses on maintaining SEO keywords, and the volatility analysis needs to return to this table: first, see which type of words have dropped, whether it is brand words, general words, or long-tail problem words; then see if the landing pages corresponding to these words are still the same article.

Three types of situations are worth paying attention to. The loss of brand words indicates the visibility problem of the entire site; the loss of general words indicates that the page has lost its competitiveness among similar content; the loss of long-tail problem words often means that the page is folded into the abstract, and the display layer is replaced.

The self-inspection actions on the content side are concentrated in three places: merge or rewrite pages with high repetition, pull back the content that has been launched in bulk to a controllable rhythm, and change the expressions with old criteria. If the site has recently done a full-site replacement, it is also necessary to confirm that the rewritten text is still consistent with the word list.

How to deal with the changes on the artificial intelligence side

About one hundred publishers are settled on a small-scale ad revenue sharing settlement in the pilot, signaling that the traffic brought by artificial intelligence displays is beginning to be formally measured. The actual significance for content sites is not equal sharing, but rather that the source of citation becomes more important.

Two things that can be done: use llms.txt to explicitly list pages worth reading, so that large language models do not have to find the focus in the whole site themselves; write the paragraphs on these pages in a self-contained Q&A form, which is convenient for extraction and citation. These two actions do not change the authorization layer, and authorization is still decided by robots group.

Rankings are last looked at.

Ranking is the result, crawling, indexing, and content quality are the reasons. When the cause layer is not clear, look at the ranking, and only repeatedly adjust the title and description, and such changes are exactly the most easily misjudged actions during the garbage update period.

The criteria for recovery should also be placed at the cause layer: the number of failed crawls decreases, the index difference converges, and the display positions of long-tail problem words reappear. The ranking usually recovers after these three.

Common questions

Question: Can a large amount of content be released during the garbage update period?It is not recommended. Batch going online will change the crawling allocation and quality impression of the site, and it is more important to maintain the original update rhythm during the fluctuation period.

Question: How to judge whether it is a penalty or a display change?Check the index volume. A significant reduction in the index is closer to low-quality judgment; stable index with a downward click is mostly a change in display form.

Question: Should we change the title and description first?Don't change it yet. Changing the title when the cause layer is not clear will mix two variables together, making it impossible to attribute later.

Question: Does llms.txt help with fluctuations?It solves whether the model can read the key page, without changing the search side index judgment, and belongs to the complementary action of the display side.

Related articles

Is the proportion of spiders blocked in training increasing, how should robots be grouped?

The proportion of blocked crawler training is rising. According to the samples checked on the day, the statistics of the robots.txt files of nearly five thousand top-level sites show that 14.3% are blocking GPTBot, and 19.2% are blocking at least one AI training crawler. It is recommended to divide the grouping writing by the purpose of crawling: search crawling, retrieval citation, and model training should be authorized separately, expressed by different user group rules, and llms.txt is only responsible for navigation without participating in authorization.

2026-11-01

Will mixing error pages and redirect pages in the site map waste the crawling quota?

The site map has error pages and redirect pages mixed in, which will waste the crawling quota. According to the standards verified on the day, the crawling party will consume the quota to request these addresses, and then discard the results, which is equivalent to spending the quota on addresses that will not generate indexes. Before submission, error pages, redirect addresses, and pages marked as not participating in indexing should be removed, and the target address of the 301 redirect should be written into the sitemap, and the last modified time should be filled in with the real modified time.

2026-11-01

Should the old site follow the new major version and default theme launched before the end of the year?

The major version and the new default theme are scheduled to be launched by the end of the year. Should the old site follow? The release plan confirmed on the day puts the official version on December 8th, and the new default theme will be released at the same time. The judgment sequence for the old site is three steps: first, confirm how long the current branch has for security support, then check the dependencies of the template and extensions on the new version, and finally see if the deployment side can roll back. Sites with expiring security support should not drag the window to the end of the year.

2026-11-01

When 90% of the extension plugins are having problems, the site should first tighten which layer?

When a problem arises largely due to extension plugins, which layer of the site should be tightened first? The public data checked that day wrote that about 90% of the vulnerabilities in the WordPress ecosystem come from plugins, and the median window from disclosure to bulk exploitation is only a few hours. The sequence of tightening is to first manage the number and source of extensions, then discuss the patching speed, and finally rely on content review and backup as a last resort: the core version update frequency is low, and extensions are the layer that changes every day.

2026-11-01

AI answer the source of reference changes every round, what is left on the content side?

The change of reference source every round is the norm, not an exception. According to the criteria checked that day, only 30% of brands can remain visible between two consecutive answers, and 86% of the references come from brand-owned or nearby sources. There are still three handles on the content side: write paragraphs as independent extractable units, concentrate the pages worth reading in the owned domain and use llms.txt for navigation, maintain consistency and freshness with content review and scheduled publishing.

2026-11-01

Is it really easier for AI to cite statistical numbers in the main text?

According to the guidelines checked on the day, the visibility of content with statistical figures in AI answers is increased by 30% to 40%, but this is not an income that can be obtained simply by adding numbers. The numbers must have a scope and a verifiable source, and the paragraph should be able to stand independently; otherwise, isolated numbers will actually reduce credibility. The focus on content is determined by the keyword library on what questions to write, and content review ensures consistency in the scope of reference.

2026-11-01

What do the terms 'references' and 'share' refer to when measuring the performance of AI search?

A share refers to the proportion of answers containing the brand out of all answers, based on the statistics of the day, the model share is the proportion of the brand appearing in the artificial intelligence answers; the citation rate looks at the frequency of the brand appearing in the source of the answer. The two sets of standards answer different operational issues and cannot be substituted for each other, the data must be able to fall on the interface or background data for verification.

2026-11-01

When the form is flooded by robots, should we add human verification first or collect fields first?

When forms are flooded by bots, the field collection is done first followed by human verification. According to the samples checked that day, in the first eight days, the proportion of real people among the hundreds of contact form submissions was very low. The pricing of captcha service at a scale of thousands indicates that bypassing verification has already become industrialized, while edge-side condition filtering can block about three to five percentage points of the bottom flow. The order of processing is to tighten fields and condition filtering, then integrate reCAPTCHA, and finally leave a content review.

2026-11-01