So Outside Technology Security CMS
AnQiCMS
Online Demonstration

AI answers mention the brand but do not provide a link, should the mention rate and citation rate be counted separately?

📅 2026-11-01 👁️ 0

Does the AI answer mention the brand without providing a link, should the mention rate and citation rate be separated? They should be separated. The mention records whether the brand name appears in the answer text, and the citation records whether the answer includes a link pointing to the source page. They are two types of events, either to treat 'being mentioned' as 'being引流', or vice versa, to treat the exposure without a link as zero value.

Mentioning and referencing are two different events

Four combinations may occur simultaneously in a Q&A session: only mentioning, only referencing, both mentioning and referencing, and neither. Statistically, the first two types of actions need to be counted separately because their corresponding subsequent behaviors are completely different: mentioning affects cognition and search association, while referencing affects click-through rate and source credibility.

The habits of different engines in giving links vary greatly. There has been a study that analyzed more than ninety responses, where responses of software queries have all named manufacturers, but mentions with links to the brand's own page only accounted for about 40%; another study covering more than 30,000 AI responses gave a lower scale, and the gap between the two engines was nearly twenty times. Placing these two numbers on the same table can show that looking only at the citation rate would judge a large number of real mentions as invalid exposure.

What three types of misjudgments will merging statistics bring?

Statistical methods Direct consequences Better alternatives
Only record citation rates Underestimate visibility, misjudge content direction Parallel recording of mention rate and citation rate
Only record the mention rate Overestimate the ability to attract traffic, and no one looks at the landing page Additional landing page visits and conversions
Merge into a 'exposure' Unable to locate the problem in content or source Establish separate accounts by query type

Also pay attention to sample differences: One decompression locks a type of interface form under a response, and the other is a massive response across engines, the two口径 are inconsistent, can only say their own conclusions, can not subtract.

The ledger must record these columns at least

Account by query dimension, not by article dimension. Each round of testing has a fixed set of questions, each question is repeated multiple times and then averaged, at least leaving: query text, query intent type, engine name, whether mentioned, whether linked, linked to which domain, test date, sample number.

The intent type is very important. Transaction and navigation queries are more likely to generate clicks, while pure definition queries are more likely to be mentioned but not cited. Averaging the two types of queries does not guide topic selection or page design.

Two things that can be done on the content side

First, write the conclusion clearly. Paragraphs should be self-contained, numbers should have sources, and terms should be explained; this type of content is more likely to be summarized into an answer and may also carry the source when summarized.

Second, make the site readable. Optimization of the generative engine is not just about stacking keywords in the text, the site should also be indexable by the model side, AnQiCMS will automatically generate site LLMs.txt, delivering the site structure and key content summaries to the readers; the AI writing ability of the back-end is responsible for supplementing the conclusion paragraphs and question-answer structures in batches, solving different links at different upstream and downstream links.

Common questions

What does it mean that the mention rate is rising but the citation rate is not?The content has entered the candidate set, but the source attribution has not been established, focus on checking whether the landing page is easy to extract and cite.

Is a mention without a link valuable?Yes, but it is difficult to attribute it directly to the visit, it is suitable to use brand search volume and direct call visits to assist in judgment.

How often should it be tested in one round?A fixed set of questions is recorded weekly or bi-weekly, the results of the same question itself will fluctuate, and the single result can only be seen as a direction.

Writing methods that need to be avoided

Do not treat 'being cited' as a rigid goal, nor should the lack of linked mentions be considered a failure; each indicator has its own decision-making meaning, and only by setting separate baselines can the trend be observed.

Related articles

Where does the difference lie in the content selection system of government portals and corporate official websites in terms of audit permissions and backup?

The emphasis of the content selection system of government portals and corporate official websites differs: the former prioritizes pre-publication review, sensitive word filtering, and permission stratification, while the latter prioritizes publication efficiency and column management. The backup range should cover data and static files, and recovery should be practiced at different rhythms. This article provides a checklist for the two types of sites.

2026-11-01

Static site generator and content system mixed, which pages are suitable for pre-generation

Static site generator compiles content into files during the build phase, with extremely low read cost, but it needs to be regenerated every time it is changed; enterprise content systems are suitable for pages that can be changed at any time and have permission. When both are used together, allocate according to the type of page: documents and landing pages are pre-generated, comments, forms, and member content remain on the dynamic side, and then use pseudo-static rules to unify the address style.

2026-11-01

Only leave security fixes and not stop supporting, when should the scheduling of the runtime environment be?

PHP version enters the stage of only providing security fixes, which does not mean ending support, but functional defects will no longer be fixed. The scheduling of runtime environments should be traced back from the end date of active support, not the final end date. This article explains the differences between the two stages, the trigger points for scheduling, and the backup and regression preparation before upgrading, as well as the different rollback methods under Docker and panel deployment.

2026-11-01

Several hours after the patch is released, someone starts probing. How should the affected version range be used to determine the emergency sequence?

After a few hours of patch release, if there is detection traffic, don't rush to change the content. Use the affected version range to confirm whether you are within the scope. Take the example of an unauthorized local file inclusion, the version of the fix and the span of the historical range determine the emergency queuing order: confirm the version, tighten the entrance, upgrade, backup and acceptance, and leave evidence for each step to review.

2026-11-01

Only about 10% of sites have installed llms.txt, can this number indicate that it is useless?

Only about one-tenth of the sites have installed llms.txt, can this adoption rate number indicate that it is useless? No. A low adoption rate indicates that it is not yet a common practice, and an unclear relevance indicates that it is not a lever for citation volume, neither of which equals invalid. This article separates the sample and scope boundaries of two studies, provides a way to judge it as a reading path rather than a ranking method, and explains how to cross-verify whether it has been read using crawl logs and site maps.

2026-11-01

The core security version fixed over twenty issues in one go, where did the rollback patch stop on which branch?

When the core security version fixes more than twenty vulnerabilities at once, it is first necessary to clarify the statistical scope, and then to see where the patch追溯 stops on which branch. This article takes a round of repairs made public at the end of this month as an example, explaining the difference between the number of fixes and the number of announcements, the minimum support branch, and why versions below the minimum do not receive any patches, and provides the disposal sequence for sites still on the old branch and the acceptance list for self-built systems during emergency upgrades.

2026-11-01

That week, the old branch was stopped just as the new version was released, what should the site running Drupal do first?

During the week of the major version release, the old branch just reached the end of support. For sites running Drupal, what should be done first? First, accept the backup, then plan the upgrade path. The official schedule compressed the two time points into adjacent days, effectively compressing the upgrade window. This article explains why the scheduling collided, what level of backup is considered acceptable, the order in which module compatibility should be checked, and what options are left after the discontinuation.

2026-11-01

Vulnerabilities exploited in the wild must be reported within 24 hours, who should submit this report

Vulnerabilities exploited in the wild must be reported within 24 hours. This obligation extends to manufacturers and software maintainers who place products on the EU market, not all website operators. This article explains the effective time of the new regulations, early warnings, detailed announcements, and the three levels of time limits for final reports, who is considered to be the party that needs to report, and how unregulated sites can organize their emergency response procedures using the same timeline.

2026-11-01