Google Ads sitelink audits for agencies: review coverage, descriptions and the scope of each client change

Define the client's navigation objective

A sitelink review should begin with the pages that help the client's intended visitor. An agency may be promoting evening courses, a specialist product range or a service with separate booking and pricing information. The audit can establish selected configuration facts, but it cannot decide which navigation choices best represent the current offer without a business brief.

Ask what the campaign promises and what the visitor may need next. A course campaign could need a schedule, course details and enrolment information. Those destinations should come from the client's actual site and current business instructions. Inventing a generic pricing page because many businesses have one creates a plausible-looking proposal with an unsupported destination.

Define success for the review in practical terms: the intended campaign has relevant selected links, known descriptions and an accountable destination owner. This is different from promising that every asset will show or that the account will obtain a fixed improvement in click-through rate. A structural and content review can be valuable without a numerical guarantee.

Keep scope visible from the start. A client asking about one campaign does not necessarily want an account-wide rewrite. The agency should know whether the requested work concerns existing link content, campaign coverage, reuse of a known asset or a new campaign attachment. These choices determine the evidence and operation needed.

Confirm who can approve the proposed page choice. The campaign specialist may know the account, while the client controls whether a course is open for enrolment. A short agreed brief reduces the chance that the writer prepares useful-looking links to offers that the business no longer wants promoted.

Separate asset content and application

A sitelink asset contains content such as its link text, descriptions and final destinations. An application relationship associates that asset with a particular account or campaign scope. A stored asset without the relevant relationship is different from an applied asset, even if its content is complete.

Google's sitelink guidance describes links to specific pages and explains that an asset can have multiple associations. It also distinguishes account, campaign and lower-level application. The agency should preserve this distinction when reviewing a shared asset or explaining the scope of an attachment change.

Do not describe every stored sitelink as currently serving. Configuration, policy, scheduling, auction selection and recorded activity are separate subjects. The current audit concerns selected presence and description completeness. A report about which link actually received activity needs the appropriate performance evidence.

Likewise, removing one campaign relationship does not necessarily delete the asset. Another campaign may use the same content, and the account may still hold it for future use. A readable proposal should say whether it changes content or changes one association, instead of using the ambiguous phrase remove the sitelink.

Read the core coverage check

ppctor's extension reader obtains enabled account asset relationships and enabled campaign asset relationships. The core coverage check treats a sitelink at account level as coverage for its selected campaign review. Where that type is not found at account level, it examines the selected campaigns for their own applied type.

This is a type-presence check. It does not establish that every campaign has a complete useful set of links, that each destination fits the offer or that all descriptions are present. Those are separate questions. A passing sitelink coverage row should not become a claim that the client's entire navigation strategy has been reviewed.

The core card also covers other declared asset types, including callouts, snippets and logos. The sitelink subject remains one part of that broader surface. Do not infer a missing sitelink merely because another type causes the overall core card to report an issue.

Current source selection includes enabled relationships without excluding automatically created sources. An agency should inspect actual content and provenance when those details affect the decision. The coverage result itself does not certify that every observed link was manually authored by the account team.

Preserve the selected campaign period

The campaign coverage population is based on enabled campaigns with recorded impressions in the selected date window, excluding the reader's declared Local Services and Smart channels. The extension reader has a default thirty-day window, while the audit passes its selected date range into that read.

A campaign that did not serve in that window can therefore be absent from the core campaign coverage population. This is not proof that a planned campaign has no sitelinks or is ready to launch. A question about a future launch needs a relevant current configuration read rather than a broad interpretation of served-campaign coverage.

Account asset relationships remain relevant even when no campaigns are currently enabled. The audit deliberately reads them independently rather than skipping every extension check in a dormant account. This preserves the difference between an account-level stored application and the campaign population selected for recent coverage.

Do not describe the description-completeness population as identical to the served-campaign coverage population. The applied sitelink set also includes the enabled campaign relationships returned by the relationship query. The agency should use the evidence boundary for the particular finding rather than assuming every card shares one date-filtered set.

Read description completeness accurately

The reader obtains sitelink content directly from the asset resource and matches it to the enabled account and campaign applications. The description check examines whether each selected distinct applied asset has both description fields present. An incomplete result identifies one or both missing lines.

This check is hidden when the selected applied set contains no sitelinks. Absence belongs to the core coverage subject, not to a description test of objects that were never selected. The lack of a description card should therefore not be interpreted as proof that every possible sitelink in the account is complete.

Filled fields are a structural observation, not an assessment of useful writing. A repeated vague description can satisfy field presence while doing little to explain the page. The agency should review the actual offer and resulting content when the client asks for improvement rather than merely filling two boxes.

The description check does not read every destination's content or establish that a particular ad displayed the descriptions. Keep the conclusion narrow: the selected asset's lines are present or absent according to the read. Delivery and business usefulness remain separate evidence.

Choose the correct counting unit

The account description reader uses a set of distinct applied sitelink asset identities. If one asset is associated with several selected campaigns, it contributes once to that unique-asset count. This differs from counting campaign associations, which would count each relationship separately.

Identical text does not prove identical asset identity. Two separately created assets can share a link label while differing in destination or descriptions. Conversely, one shared resource can appear in several campaigns. The agency should not deduplicate by visible wording when the requested unit is the asset or association.

The missing-description total counts selected incomplete assets, not the number of missing lines. One asset missing both fields contributes one incomplete asset and two absent lines. A readable summary should state which unit it reports so the client can judge the actual work required.

A campaign coverage total is another unit again. It counts selected campaigns lacking a declared type under the check's account-first logic. Do not merge this total with incomplete-asset counts. The agency can present both findings clearly without implying that every missing campaign needs a newly invented asset.

If the user requests a list limited to a few examples, separate that preview from a complete count. Showing five link labels does not establish that five is the full selected population. The reporting plan should preserve collection coverage and the requested counting unit when producing a short conversational answer.

Handle failed and unmatched reads

Failures in the extension section are recorded and rendered as an unavailable check rather than silently turning fallback arrays into missing coverage. Both relationship and content reads are required for the description comparison. A failed asset read should not become an assertion that every applied link has empty descriptions.

The current implementation has a narrower unmatched-row caveat: if an applied identity is returned but its content row is absent from an otherwise successful asset collection, it assigns a generic label and treats both descriptions as missing. That is a fallback classification, not two directly observed empty fields.

If such a generic unmatched entry drives a proposed edit, resolve the actual resource and content first. The agency should not tell the client that it verified blank descriptions when the content row itself was missing. Current exact target evidence is necessary for a defensible change.

Preserve partial knowledge in the report. Knowing that an enabled relationship exists does not reveal its unread content. Knowing that a content asset exists does not establish an enabled association. The next query or owner review should address the missing side of that relationship rather than repeating unrelated account checks.

Review inherited and campaign-level links

The account-first coverage result can be useful while leaving business relevance unresolved. An account-level course catalogue link may make sense broadly, while a campaign promoting one workshop may need a specific enrolment destination. The client should understand whether the audit found inherited type coverage or an actual campaign-specific relationship.

Adding a campaign-level sitelink does not automatically detach the account-level pool. The provider's hierarchy allows relevant associations from the same branch to remain eligible together. An agency should not promise exclusive campaign control merely because a new campaign attachment was created.

If the client wants an inherited link prevented from appearing in a particular scope, investigate the actual supported relationship operation and its wider effects. Removing an account relationship can affect other campaigns. Do not substitute a campaign detachment that leaves the inherited association untouched and describe the request as fully completed.

The current compact coverage reader does not provide a full inventory of lower-level sitelink associations. A missing account or campaign type does not establish that no group-level association exists. A complete hierarchy question needs the relevant supported relationship sources, with the exact account branch preserved.

Interpret the separate campaign cards

The campaign-level extension card asks whether the campaign has any of its own declared core types or relies on account assets. Its pass condition is not a requirement that every core type be present. A campaign with one qualifying type can satisfy that broad own-extension check while a specific sitelink question remains open.

The campaign sitelink-description card checks whether there is at least one own sitelink with both lines. That differs from the account description check's all-selected-assets completeness test. A campaign card can therefore have a positive result without proving that every associated sitelink is fully described.

Explain which card answered which question. Inherited account sitelinks are not broken solely because they are inherited, and campaign-specific content can still be worth preparing. The agency should distinguish an operational preference for relevance from an established missing-setting failure.

Do not overread a summary label. The underlying observations and predicates determine what is established. If the client asks whether all links in one campaign have descriptions, use a report whose returned set and condition answer all, rather than treating an at-least-one finding as the requested total.

Select destinations from the actual offer

Choose pages that support the campaign's visitor journey. A schedule page can help someone deciding whether a class fits their availability; an enrolment page can support someone ready to register. These purposes should come from the current brief and actual destinations, not a universal list applied to every client.

Ask whether each linked offer is available in the campaign's relevant market and period. A broad account link may promote a service unavailable to the audience of a specialist campaign. The coverage reader does not assess this commercial fit. The agency needs business confirmation where it changes the proposed content.

Keep promotional and evergreen links separate in the project brief. An expired offer can remain as a stored asset even after one campaign stops using it. Review the actual relationships before changing shared content or proposing a replacement for a link that another current campaign still needs.

Do not invent the final URL from the link label. A label called Course Schedule does not prove that /schedule is a valid route on the client's site. Use a supplied or appropriately verified destination, and ask for the specific missing URL when the available context cannot establish it.

Consider the relationship between the main ad destination and the proposed alternatives. The agency can ask whether a visitor needs more detail, a different course choice or a direct next step. This is original planning based on the brief, not evidence that the provider returned a recommendation for the client's page structure.

Write useful link text and descriptions

Link text should tell the visitor what the page offers. Descriptions should add useful factual detail rather than repeat the label twice. For a course schedule, the lines might explain the relevant session options using the client's confirmed information. The assistant can help draft alternatives from that brief without performing a new account query just to brainstorm.

Distinguish suggestions from existing account content. A creative reply can say here are possible labels based on the supplied offer. A statement that these are the current labels needs an actual read. The agency should keep the two modes clear when moving from discussion to a live proposal.

The standard sitelink creation helper requires link text within its declared twenty-five-character limit and a final URL. Its convenience fields allow the description lines to be supplied. Provider validation still applies to the resulting request; the helper's acceptance of preparation is not a blanket compliance certificate.

Review the complete selected wording, including any factual promises. Filling both lines is not permission to add unavailable discounts, guaranteed outcomes or unsupported qualifications. The connected owner should see the actual text before confirming, and the agency should preserve the factual source in its handoff.

When a client rejects a proposed description, update the complete draft rather than retaining the rejected line as an unseen default. The conversation may contain several alternatives. Preparation should identify the selected version unambiguously so the approval cannot accidentally include an earlier suggestion that the owner already discarded.

Establish destination evidence

The creation helper normalises supplied URLs through the landing-URL function and checks their public-page form. A domain-style input can be converted to HTTPS where that validation permits it. This is useful input handling, but it does not crawl the page or prove that it delivers the promised content.

Separate configured destination, page availability and business correctness. The agency can verify that an asset stores the requested URL while the website owner confirms that the relevant schedule, offer or form is ready. Both observations matter, but they come from different evidence.

A page change and an asset change are different tasks. If the client updates the schedule at the same URL, the asset may not need a new destination. If the desired page changes, the supported edit must describe the actual resulting setting. Do not manufacture an Ads write merely because website work occurred.

For shared links, inspect the impact of any proposed destination change. A new page suitable for one campaign may be unsuitable for another association. The current campaign creation or attachment tools can support a targeted new asset where appropriate; they should not be described as silently rewriting a shared resource.

Record an unresolved page question before seeking approval. If the agency has the domain but not the enrolment route, the missing fact is that route. Asking a precise question keeps the flow useful. Guessing a plausible path creates an avoidable error that valid URL syntax would not detect.

Decide whether to reuse an existing asset

An existing suitable asset can be attached to the intended campaign without creating another copy. The tool checks that the selected resource belongs to the account and matches the requested asset type. It also needs readable content or a name for the owner to review.

Text similarity alone is not enough to establish suitability. Confirm the relevant destination and descriptions when the client asks to reuse a specific link. A convenient label in a compact inventory can hide different content. The proposal should refer to the actual selected resource through validated identity.

The helper checks existing enabled campaign relationships. If the requested attachment is already present, it avoids creating that same relationship again. This is a scoped duplicate check, not a claim that the account contains no other assets with similar content.

Reusing an asset preserves its shared identity. Future content changes to that asset can affect its associations. The agency should record this shared ownership when the client chooses reuse. A campaign attachment does not create a private copy of the underlying content.

A newly created asset can still resemble an existing one. The creation helper's declared behaviour is not a universal content deduplication service. If avoiding parallel copies is part of the request, establish the relevant existing candidates first and choose reuse deliberately rather than claiming that every new creation searched all similar content.

Prepare supported creation and attachment

The standard campaign-extension creation action supports a sitelink with the supplied link text, optional description fields and final destinations. It creates the asset and attaches it to the selected campaign using the returned relationship plan. The proposal should display both effects rather than implying that only a detached library object will be made.

The helper preserves existing campaign extensions. A new asset does not replace every current sitelink, and the operation does not automatically detach inherited account relationships. This additive behaviour matters when the user asks for a new link while expecting the rest of the current navigation to remain.

Creation and attachment are submitted together in this helper's declared mutation request. That local implementation detail should not be expanded into a promise that every multi-action conversation is one atomic provider transaction. A larger proposal can involve separate writes with their own outcomes.

The dedicated convenience path is for campaign scope. A request for account-level creation, shared content modification or a lower-level association needs the actual available supported action and its validated fields. Do not advertise the campaign helper as a universal editor for every sitelink setting.

Distinguish detachment and content removal

Detaching an existing campaign sitelink removes the selected enabled campaign relationship while retaining the asset. Other campaign or account relationships are separate. The owner should see a readable instruction that identifies the campaign and selected asset instead of a vague deletion statement.

The helper checks the current selected relationship and preserves the remaining enabled set in its verification plan. If the requested direct relationship is absent, there is no direct association to detach. This does not prove that the content cannot remain eligible through an account-level or another supported branch relationship.

If the user wants the link removed everywhere, establish every relevant association and confirm the actual supported operations. A campaign-local request should stay local. Do not quietly broaden it to account scope merely to produce the desired appearance in one campaign.

Content deletion is another operation with its own availability and constraints. The current detach helper does not promise it. An agency can use precise language to explain what will remain after the change, making the owner's approval more useful and avoiding an incorrect success summary.

Follow an invented agency campaign review

This example is invented. Birch Shore Media manages Cedar Vale Courses. The Evening Classes campaign is covered by an account-level sitelink type, but the client wants a current enrolment link specifically associated with that campaign. The account coverage pass does not establish that the intended campaign-specific destination already exists.

The specialist checks the supplied course brief and obtains the confirmed enrolment URL. The writer drafts a short factual link label and two useful description lines. This creative preparation uses the client's context; it is not presented as retrieved account content or guaranteed performance improvement.

The client asks for a new campaign sitelink rather than reuse of a known shared asset. The proposal names Evening Classes and displays the selected label, both descriptions and final URL. It says that the new asset will be created and attached while existing extensions remain.

The connected owner confirms. The supported helper creates the asset, attaches it to the campaign and checks the declared content and relationship using the returned identity. The final technical result establishes those settings. It does not say the old inherited links were removed or that the new descriptions appeared in a particular auction.

The agency records the remaining account-level pool and the website owner's destination responsibility. If the client later asks for exclusive campaign navigation, that is a new scope question requiring the appropriate hierarchy review. The original additive action should not be retrospectively described as an exclusivity configuration.

A later performance request uses the correct source and grain to inspect actual observed activity. The structural receipt remains useful evidence of what was added, while commercial results remain separately measured. This keeps the campaign review accurate from creative brief through execution and follow-up.

Review clear owner-approved steps

Show the campaign and selected asset action in readable rows. For creation, display the complete content and destination, followed by the attachment effect. For existing attachment or detachment, use the asset's readable name or content and the exact intended campaign relationship.

One confirmation can cover a supported sequence with each action clearly shown. If the complete request cannot be prepared safely, explain why smaller reviewed proposals are needed before writing. Do not silently create only some of the requested links and report the whole set as complete. The owner must know every intended effect before live execution, including creation and attachment relationships.

The pending message states that no changes have been made yet and that the proposal expires in fifteen minutes. The connected owner's confirmation authorises the displayed supported operations. Asking for draft labels or a coverage explanation should remain a conversation without a live write.

Keep relevant notes visible. Existing extensions retained and asset itself retained describe real resulting behaviour. Internal resource references and provider payloads are not needed for ordinary client review. Clear formatting should reduce the reading burden while preserving target scope and complete requested content.

Verify fields and the exact relationship

Creation readback checks the declared asset fields and final URLs through the returned created resource. It also checks the enabled campaign association for that actual asset. This separates stored content from successful application and avoids assuming that one accepted response proves every intended leaf field.

Existing attachment and detachment use a current relationship read and verify the expected remaining enabled asset set. The internal identity is exact even when the client output uses names. A readable receipt can therefore be precise without displaying raw IDs.

The verified result establishes the declared configuration. It does not guarantee policy approval, auction selection, displayed descriptions or increased conversions. The agency should give the technical milestone appropriate prominence while preserving the separate delivery and commercial questions.

A later coverage check may observe the new type or complete descriptions within its selected population. That is another scoped structural result. It still does not establish that every lower-level association was reviewed or that the client has exclusive control over the eligible link pool.

Reconcile uncertain attachment or creation

A delayed Slack result or incomplete readback does not establish that creation failed. The asset and relationship may already exist. Use the durable execution record and current provider state before proposing another creation, so an uncertain message does not become duplicate account content.

Distinguish the two milestones when reconciling. An asset may be known while its relationship remains unresolved, or a relationship observation may establish current application while some content field remains unread. The next check should address the actual uncertainty rather than repeating every operation blindly.

For several requested actions, retain their separate outcomes. One successful attachment does not prove that another new asset was created, and an unresolved later check does not establish rollback of earlier accepted work. The client should receive the actual confirmed and unresolved scope.

Do not infer failure from a refreshed audit that uses another population. A new association can be present while a served-campaign coverage card omits a campaign without activity in its window. Reconciliation should inspect the actual written relationship, rather than using a different audit predicate as a substitute for its status check.

Coordinate agency recovery ownership. One specialist should reconcile the attempt and communicate the result before another prepares overlapping edits. A corrective change or restoration remains a new deliberate proposal with current identity and the required owner approval.

Read actual performance separately

When the client asks how sitelinks performed, select the relevant supported performance grain and outcome definition. A campaign total repeated beside each asset does not measure each link's individual contribution. Relationship-level and asset-level results can also answer different questions when content is reused.

Preserve the distinction between activity associated with an ad containing assets and activity on a particular asset. A field's readable label should not hide the provider's actual measurement. The reporting framework needs a compatible validated plan for the requested metrics and dimensions.

Use completed date bounds and consistent primary outcome definitions where those metrics are selected. A value metric can include legitimate primary non-purchase actions; purchase-only reporting is narrower. The presence and description checks do not choose these definitions automatically.

Do not promise a fixed uplift from adding descriptions or more links. The client can approve more useful navigation based on the offer and observe subsequent results. A before-and-after comparison should acknowledge other account changes and the available evidence rather than assigning every movement to the most recent asset addition.

Continue the sitelink conversation in Slack

A useful opening request asks whether the selected campaign has its own sitelinks and which returned links lack descriptions. The answer should distinguish own and inherited coverage, the selected asset set and any unread detail. It should not present an at-least-one campaign result as proof that all links are complete.

A follow-up can ask for wording ideas from the already supplied course brief. The assistant can produce creative suggestions without a new provider call solely to brainstorm. If the user asks what the current sitelinks say, retrieve or reuse the relevant actual content evidence instead.

Retain established campaign and destination context when preparing a requested edit. Ask only for genuinely missing information or ambiguity. The tool still needs current exact target checks where required, but the owner should not have to repeat every fact from the preceding conversation.

Keep normal mentions before the final feedback text and controls, and keep progress messages visible without a user mention. Avoid raw IDs and technical mutation objects in client replies. The target, complete content, relationship action and actual result are the useful information.

Reuse relevant data without broad scans

The extension reader obtains its independent base collections concurrently and matches assets and relationships locally. It also gates additional logo reads according to relevant observed channel needs. Those logo surfaces should not be described as additional sitelink hierarchy coverage merely because they belong to the same module.

Retained content can answer an explanation request without another account scan. A new exact attachment request needs the relevant current campaign relationship and selected asset validation. The difference is useful: reuse evidence for interpretation while collecting the live-state observations required for the operation.

Local dependency planning can establish that an existing campaign is the parent without a provider request for unrelated prerequisites. A campaign sitelink request does not require reading every keyword, Merchant product or Analytics event. Focus additional collection on the fields and relationships that answer the client's actual question.

Efficiency does not mean removing result checks. Avoid redundant work while preserving current ownership, type and relationship validation. Do not promise a universal call count or instant latency. The client needs an accurate scoped result, while infrastructure details remain behind the scenes.

Leave a useful agency handoff

Record the campaign, selected asset identity internally, readable content, confirmed destinations and intended application scope. Include whether the action reused an asset or created a new one, which relationships remain and the actual verified or unresolved result.

Assign website readiness and future content maintenance separately. A shared enrolment link may need review when the course schedule changes. Another campaign owner may rely on the same asset. The handoff should identify shared use before someone later treats the content as a private campaign object.

Preserve the client's approval boundary when the brief changes. An earlier proposed destination or description should not be executed after it has been withdrawn. Retained context is evidence of the history, not permission to ignore current instructions.

Set a practical review trigger for temporary content. The agency may need another check when the client changes the enrolment period or withdraws an offer. A stored asset can remain technically complete while its message becomes obsolete. That later business review should use current instructions and a fresh supported proposal.

Make the final conclusion specific enough for another operator to continue. State whether the selected campaign gained a new link, reused a known shared link or lost one direct association. This concise outcome is more useful than a general sitelinks optimised label that conceals what changed and what remained eligible.

Use the agency audit workflow for account context and the focused sitelink audit guide for this subject. Related policy review remains a separate task when an actual policy observation affects the selected asset.

Answer sitelink audit questions

Does one shared asset count once or once per campaign?

The account description check counts distinct selected assets once.

Does coverage prove the descriptions are complete?

No. Type presence and description completeness are separate checks.

Does a campaign description pass mean all links are complete?

No. Its current own-link condition is at least one complete sitelink.

Does campaign attachment remove inherited account links?

No. It adds the selected relationship without that broader removal.

Does detaching delete the asset itself?

No. The current campaign detachment retains the content resource.

Does verified creation guarantee the link will show?

No. Configuration and actual delivery remain distinct.

Should uncertain creation be replayed?

No. Reconcile the recorded attempt, asset and relationship first.