A regional search-visibility deliverable is useful when it lets a client decide one bounded next action from traceable evidence—not when it turns country aggregates, a one-off screen, or an unrecorded city into a promised rank or sale. Start with an official export, licensed report, or a permitted manual observation. Record what each source actually measures, the country it attributes, and the device, language, time, and query context. Keep missing evidence visible.
This guide is for an agency or in-house search team preparing a small recurring client report for a site it is authorized to assess. It does not query Google, generate traffic, create accounts, use a proxy to evade limits, or claim IPHTML results. A report can support a manual review queue; it cannot prove a city-wide ranking, demand, leads, or revenue.
Define the decision before collecting evidence
A buyer does not need a vague “rank tracker.” They may need to decide whether a regional landing page deserves a title review, whether a business listing owner should check an address discrepancy, or whether a new localized page needs a content owner. Put the decision, owner, deadline, approved site scope, query family, and permitted evidence route in the brief. If no one can act on an item, do not schedule it merely because a dashboard can show it.
| Client asks for | Deliverable can say | It must not say |
|---|---|---|
| “How visible are we in Canada?” | Country-attributed Search Console metrics for the documented date range and filters. | A literal city result or a stable rank for every Canadian searcher. |
| “What did the result look like in Toronto?” | A permitted, time-stamped manual observation with stated device/language/context. | That all Toronto users saw it, or that it caused visits or sales. |
| “Should we fix this page?” | A narrow evidence-backed review task and owner. | That a title edit will improve rank, traffic, or revenue. |
| “Can you monitor this every week?” | A defined source, fields, cadence, exception rule, and manual review boundary. | Continuous local coverage without a permitted source and acceptance rule. |
When the question is a page experience rather than a search-result decision, use the regional website QA matrix. When it is a campaign promise versus its own landing page, use regional ad and landing verification. Those are controlled destination checks; this guide owns the client-facing acceptance of search-evidence records before a narrow action queue is issued.
Choose a permitted source before considering a proxy
Use the least intrusive route that answers the stated decision. First-party Search Console exports, a source licensed for the specified report, and documented manual observations have different scopes. An official Search Console aggregate can attribute a query to a country and device, but it is not a browser observation in a named city. A manually viewed result can record a particular moment and context, but it is not a population estimate.
| Priority | Route | Useful evidence | Limit to retain | Use only when |
|---|---|---|---|---|
| 1 | Official owner export | Clicks, impressions, CTR, and average position under documented Search Console dimensions. | Average position is an aggregate metric, not an observed literal placement; query data can be anonymized or omitted. | Client has authorized access and the filters/date range are recorded. |
| 2 | Licensed report or API export | The fields and location semantics stated in that supplier's contract and export. | License, coverage, update time, and location definition may limit reuse. | Its terms expressly permit the proposed client reporting. |
| 3 | Permitted manual observation | A reproducible screen or note for one stated query, country, asserted city if applicable, device, language, and time. | One observation is not an average rank, a demand estimate, or proof of universal display. | The source rules and client authorization permit it; keep volume small and human-reviewed. |
| 4 | No collection | An evidence gap, existing client report, or a request for source permission. | No finding can be inferred from the gap. | Permission, context, or a valid source is absent. |
Google’s current spam policy calls automated queries to Google—including rank-checking result scraping or other automated Search access without express permission—machine-generated traffic and says it violates policy and Google’s terms. Do not automate Google result queries without express permission, and do not use code, rotating routes, traffic, or a proxy to bypass limits or blocks. A residential route is not permission. If an official or licensed route answers the decision, keep it.
Separate country, asserted city, and browser context
Use precise labels. Recorded country is what the approved source actually returned. Asserted city is a requested or tool-reported city context and remains an assertion unless the source documents that geography. Observation context records the device, browser/language settings, signed-in or signed-out state, query, time zone and time. None of these labels substitutes for the others.
Search Console documentation describes country and device dimensions, anonymized queries, and average position calculated from impressions. It supports report metrics only in the selected aggregation. It does not license the leap from “average position 6.2 for Canada” to “ranked sixth in Toronto today.” Keep the two fields in separate columns and never calculate a blended “rank” from them.
A worked synthetic client report
The following fictional report is for Northstar Bikes, an authorized client. It contains invented values solely to demonstrate a handoff. It is not a Google export, a live observation, a measured search result, or evidence about IPHTML.
| Evidence ID | Source and permission | Recorded country / asserted city | Device / language / time | Metric or observation | What the client may decide |
|---|---|---|---|---|---|
| N-01 | Official owner export; permitted | CA / no city asserted | Mobile / language NOT_APPLICABLE (not a GSC dimension) / query “northstar commuter bike” / page /ca/commuter / Oct 1–28, Pacific reporting date | 120 impressions, 3 clicks, 2.5% CTR, and 6.2 average position—hypothetical aggregate values only. | Content owner reviews whether /ca/commuter answers the query; do not call 6.2 a Toronto rank. |
| N-02 | Permitted manual observation; scope recorded | CA / Toronto asserted by approved test context | Desktop / en-CA / signed out / query “northstar commuter bike” / 2026-11-17 14:00–14:05 America/Toronto | One result card for /ca/commuter was visible; screenshot reference retained. | Reviewer checks title and destination relevance; do not infer impression share or conversion. |
| N-03 | Licensed export; permission verified | CA / city field absent | Mobile / en-CA / supplier timestamp recorded | Supplier reports its documented visibility field. | Compare only under that supplier's documented semantics; retain license reference. |
| N-04 | No approved source | CA / Vancouver requested | Device and language not recorded | MISSING—not a result. | Ask for a permitted source or mark the location out of scope. |
| N-05 | Permitted manual observation | CA / Montreal asserted | Mobile / fr-CA / 2026-11-17 10:15 America/Toronto | Expected page was NOT_PRESENT in this one observation. | Manual review: check query, locale, page eligibility, and source context; do not report “absent in Montreal.” |
The report intentionally does not combine the country aggregate in N-01 with the literal screen in N-02. It also preserves MISSING separately from NOT_PRESENT: missing means no usable evidence was supplied; not present means one permitted context did not show the expected item. Neither is a zero metric or a city-wide absence claim.
Accept records with a small evidence worksheet
For every row, record source type and permission, source reference, query or query group, target URL, reported country, asserted city (or blank), device, language, time and timezone, metric semantics, observation state, reviewer, and the only allowed next action. Use a fixed vocabulary:
PRESENT: the stated observation contains the item; it says nothing about all users.NOT_PRESENT: the stated observation did not contain the expected item; it creates a manual review, not an absence claim.MISSING: source, context, or observation was not supplied; it is not zero.GSC_AVERAGE_POSITION: aggregate metric only; never rewrite it as a literal observed rank.
Download the offline Python 3.9 worksheet and run python3 regional-search-evidence-deliverable.py. It contains six synthetic records and no URLs. It rejects malformed states, treats permission and missing device/language context as holds, and demonstrates that a country aggregate, a one-off observed result, a missing record, and a non-present observation produce different dispositions.
country_export: REPORT_COUNTRY_METRIC_ONLY
manual_context: REPORT_OBSERVED_CONTEXT_ONLY
not_present: MANUAL_REVIEW_NOT_ABSENT
missing_context: HOLD_CONTEXT
unknown_permission: HOLD_PERMISSION
missing_evidence: MISSING_EVIDENCE
The script is a report-classification aid, not a scraper, rank checker, traffic generator, or city-location verifier. It sends no request, reads no credential, resolves no name, changes no setting, and cannot establish demand, clicks, or sales.
Issue a narrow manual action queue
A useful recurring report ends with a small queue whose action matches its evidence. For example: “content owner: compare page intent with the documented Canada query group by Friday”; “listing owner: verify the approved address source because an authorized report conflicts”; or “analyst: request device/language context before interpreting N-04.” Avoid instructions such as “improve Toronto rank” or “fix sales.”
| Evidence condition | Queue state | Permitted next action |
|---|---|---|
| Official country export shows a relevant page-query pattern with documented scope. | REVIEW | Human checks page intent, title, and client-approved content facts. |
| One permitted observation differs from expected regional presentation. | RECHECK | Repeat only with approval and the same recorded context; compare a control. |
| Permission, device, language, time, or source semantics are missing. | HOLD | Request the missing evidence; do not create a performance finding. |
| Source is official, licensed, manual, or no-proxy alternative meets the need. | KEEP_SIMPLE | Use it; do not add proxy collection. |
For source-rights, coverage, freshness, and reconciliation decisions, see the public-data source selection guide. For commercial scope, acceptance, and capacity before offering a recurring report, see market-monitoring service feasibility. Neither replaces the record-level distinction between a client report metric and a specific search observation made here.
When an authorized client has a documented location-specific evidence gap: first ask whether an official export, a licensed source, a manual review, or no collection resolves it. Only if a permitted, small, human-reviewed observation still needs a residential network context, review IPHTML residential proxy information and contact the team with a redacted scope: approved domain, country (not assumed city precision), device/language context, evidence source, frequency, owner, and stop condition. Confirm current terms and suitability before any test. Do not send credentials or use a route to bypass source controls.
Sources checked November 17, 2026. Google Search Console’s table guidance documents country/device dimensions, anonymized queries, and how aggregate average position is calculated. Google’s Search spam policy identifies automated queries, including rank checking without express permission, as machine-generated traffic. Google’s Terms of Service prohibit abuse and automated access that violates machine-readable instructions. These sources support the evidence boundaries above; they do not establish search demand, rank, client outcomes, IPHTML capability, pricing, or legal advice.