Skip to content
search-engine-optimizedby Netsolution

Controlling what is shown of your page

Four published rules limit what a search engine shows of a page: nosnippet, data-nosnippet, max-snippet and max-image-preview. The first two also cover generative answers, which they stop from using the text as a direct input. Where rules conflict, the most restrictive one applies.

An indexed page is not a page shown in full

Between letting a search engine do as it likes with your text and taking the page out of the index, there is a middle ground, documented, and oddly little used. Four rules live there, all written in the same place, and two of them took on a new scope once generative answers began drawing on pages.

The nosnippet rule prevents a text snippet and video preview from being shown in all forms of results, including the AI-based overviews and conversational mode, and also prevents the content from being used as a direct input for those answers.

officialSource : Google Search Central, “Robots meta tag, data-nosnippet, and X-Robots-Tag”

The second half of that sentence is the most important thing on the subject, and it is recent in the documentation: the rule does not merely hide the snippet, it prevents the content being used as a direct input for those answers. It is, to date, the only published lever acting on that use while leaving the page in the index.

The max-snippet rule limits the number of characters in the snippet and also limits how much of the content may be used as a direct input for generative answers; the value 0 is equivalent to nosnippet and the value -1 lets the engine choose.

officialSource : Google Search Central, “Robots meta tag, data-nosnippet, and X-Robots-Tag”

The setting is therefore gradual rather than binary, which matters to a publisher who wants to stay visible without handing over the whole answer. Two particular values are worth remembering: zero means complete removal, and minus one hands the choice back to the engine.

The trade-off nobody presents as one

Restricting the snippet reduces what a reader can judge before clicking as much as what a generative answer can reuse: the two effects are inseparable, since no published rule lets you target one without the other.

hypothesis

The status says what it is. The documentation does not present that consequence as a trade-off: it describes rules and their effects. We add that the two effects are linked by construction, since the same rule produces both. A publisher who cuts the snippet so as not to be reused also cuts what their reader could judge before clicking.

Protect your content from AI: our plugin adds an anti-AI directive to all your pages, with no effect on your SEO.

The published rules that act on use by generative answers are the ones governing the snippet. Promising one without the other announces an effect no documented directive produces.

Excluding a passage rather than a page

The data-nosnippet attribute excludes a portion of text from the snippet, on span, div and section elements only; any value given to the attribute is ignored, and an unclosed tag extends the exclusion to everything that follows.

officialSource : Google Search Central, “Robots meta tag, data-nosnippet, and X-Robots-Tag”

This is the finest of the four instruments, and the least known. It keeps a descriptive snippet while withholding one precise passage: the opening of a subscriber article, a price that changes often, a figure that loses its meaning out of context.

Two implementation traps are flagged by the documentation itself, and both are silent. A value written into the attribute is ignored, false included, which surprises anyone who thinks they are switching the exclusion off. And a tag left open extends the exclusion to everything that follows, to the end of the document.

That limit does not apply where the publisher has separately granted permission for use, for instance by supplying the content as in-page structured data or through a licence agreement.

officialSource : Google Search Central, “Robots meta tag, data-nosnippet, and X-Robots-Tag”

That exception deserves a second reading from publishers who mark up their pages heavily: a restriction set in the rules can be lifted by a permission granted elsewhere, and in-page structured data counts as one.

The rule that decides when two directives disagree

Where robots rules conflict, the more restrictive rule applies: a page carrying both max-snippet:50 and nosnippet will show no snippet at all.

officialSource : Google Search Central, “Robots meta tag, data-nosnippet, and X-Robots-Tag”

On a real site, directives come from at least three places: the template, an HTTP response header, and sometimes a plugin of the content management system. The published rule settles it without ambiguity, and it always leans the same way — the strongest restriction wins.

Three once-common rules are no longer used by Search: noarchive, since the cached page feature no longer exists, nocache, and nositelinkssearchbox, whose feature has gone.

officialSource : Google Search Central, “Robots meta tag, data-nosnippet, and X-Robots-Tag”

Tuning the snippet rather than removing it, sorting which pages need protecting

hypothesis

Effort
moderate
Time to effect
weeks
Risk
low
Durability
lasting

The effort lies in the sorting: deciding which pages deserve a restriction takes longer than writing it. The delay follows the next crawl. The risk deserves two notches, because a directive set by mistake on a template empties the snippets of an entire site, and the drop in clicks that follows resembles anything but its cause. Durability is medium: these rules change, three of them already declared unused.

What does not work, or no longer works

Setting noarchive in the belief that it does something. The documentation writes that the rule is no longer used, and two others are in the same position.

Then looking for a directive that would exclude generative answers without touching the snippet. The published rules govern both together.

Finally, restricting the snippet without rereading your structured data. The documentation explicitly provides that a permission granted elsewhere escapes the limit.

Sources

  1. Robots meta tag, data-nosnippet, and X-Robots-Tag specifications — Google Search Centralaccessed on
  2. AI features and your website — Google Search Centralaccessed on

Block revalidated on

Frequently asked questions

Can you stop an AI reusing your content without leaving the results?
Not with the published rules. The same rule governs the snippet and the input to generative answers: nosnippet removes both, max-snippet reduces both. The page stays indexed and shown, but without descriptive text, which changes what a reader can judge before clicking.
What is data-nosnippet for, rather than nosnippet?
For excluding part of a page rather than the whole: a paragraph reserved for subscribers, a figure you would rather not see quoted out of context. The attribute only works on span, div and section elements, and an unclosed tag extends the exclusion to everything that follows.
What happens when two rules contradict each other?
The more restrictive wins, and the documentation gives the example: a page carrying both max-snippet:50 and nosnippet will show no snippet. That is the rule to know before writing directives in several places, between a template, an HTTP header and a plugin.
Is noarchive still worth setting?
No. The documentation writes that the rule is no longer used, the cached page feature having gone. Two others are in the same position, nocache and nositelinkssearchbox. They do no harm, but their presence mostly signals a configuration nobody has reread in a long time.
Does structured data escape these limits?
The documentation covers that case explicitly: the limit does not apply where the publisher granted permission separately, notably by supplying the content as in-page structured data or through a licence agreement. Restricting the snippet without reviewing your markup therefore leaves a door open.