Insights / Search diagnostics
Why Is My Website Indexed but Not Ranking on Google?
A diagnostic guide for service businesses.
Your page can be indexed and still perform poorly for the searches that matter to your business. Start with one page, one intended search and one market. Confirm Google's indexed version, inspect page-filtered Search Console data, and compare the page with the task a prospective customer is trying to complete. Repeated indexing requests do not diagnose that gap.
This guide provides a working method for choosing the next investigation. It is not a ranking experiment or a report of client results. The examples below are hypothetical; the decision framework is Layerly's proposed method, grounded in the linked Google documentation.
1. Define what "not ranking" means
"Our SEO isn't working" can describe four different situations:
| Situation | Evidence to collect | Investigation |
|---|---|---|
| The intended page is not indexed | Indexed URL Inspection result | Crawl access, indexing decision and canonical selection |
| The page is indexed, with little relevant search visibility | Page-filtered Performance report | Query fit, page usefulness and discoverability |
| The page appears but attracts few clicks | Query-level impressions, clicks and result context | Search intent, competing results and presentation |
| People visit but do not enquire | Consented analytics and request records | Audience fit, offer and contact journey |
On smaller screens, scroll the table horizontally.
Choose the row before choosing a fix. A broken email link can obstruct enquiries without explaining a ranking gap. An indexed page can also rank well for a query that brings the wrong audience.
Write a testable problem statement: "Our bookkeeping service page is indexed, but we have little recorded visibility for searches about outsourced bookkeeping for UK consultancies." That statement identifies a page, an audience and a service. "We need to rank for business" does not.
Google's ranking systems assess many signals and relevance to searches. An index entry is not evidence of a competitive position for your intended query. [1]
2. Inspect the exact page, not just the homepage
Open URL Inspection in the correct Search Console property and enter the complete service-page URL. Save the inspection date, indexing status, last crawl and Google-selected canonical where available.
The indexed report describes Google's stored view. The live test checks the current page and cannot determine Google's eventual canonical choice. A successful live test does not establish that a changed page has been reindexed or will rank. [2]
If the intended page is not indexed, leave this ranking workflow and investigate that reason first. If Google selected another canonical, inspect whether the pages are genuinely equivalent and whether redirects, canonical declarations and internal links express the intended destination. Do not remove canonical tags indiscriminately. Google's canonicalization documentation explains how signals work together. [3]
A useful implementation record contains two separate dates: when you changed the page and when Google's indexed view was observed to reflect it. Keeping them separate prevents a deployment from being reported as a search result.
3. Build a page-specific search baseline
In Performance, choose the relevant search type, apply an exact page filter, and examine Queries, Countries and Devices. Use a completed reporting period and record every filter. Compare like-for-like views rather than a whole-site average against one service page.
Clicks, impressions and average position describe different aspects of visibility. Average position is an aggregate across recorded impressions, not a fixed rank for every searcher. A manual search can differ with time, location, device and search history. [4]
Our working recommendation is to keep an initial baseline and then review comparable completed periods. For a new, low-volume page, widen the window when evidence is sparse; do not turn a calendar deadline into a claim that Google should have delivered results.
There are two important limits. Query tables omit some anonymized queries and can omit rows because of reporting limits. Also, separately exported page and query totals do not establish which query led to which page. Google documents these reporting differences. [5]
Consequently, "I cannot see the query in this export" is not equivalent to "nobody searches for it." Zero recorded impressions in a selected view do not reveal the cause by themselves. Record the observation, check the filters, then investigate possible explanations.
4. Compare the customer's task with the page you published
Select one intended search and examine three relevant organic results in the market you serve. Record the date, market, query and result URLs. This is a small qualitative sample, not a permanent ranking survey.
Use the following comparison:
| Question | What to record |
|---|---|
| What is the searcher trying to do? | Learn, compare suppliers, buy, troubleshoot or reach a named business |
| Which page types appear? | Service pages, guides, directories, tools or another format |
| Which decision does each result help make? | Scope, suitability, price structure, process or next action |
| What evidence supports it? | Demonstration, methodology, named expertise, useful examples or relevant references |
| What can our page contribute? | A specific improvement the intended customer can use |
On smaller screens, scroll the table horizontally.
Do not copy the ranking pages or assume their shared characteristics caused their position. Use them to test whether you are answering the intended task.
For example, a hypothetical agency targets "white-label link building for agencies" with a short definition of backlinks. The page may explain a term without helping an agency evaluate delivery. A more useful page could explain prospect approval, responsibility for writing, reporting, paid-placement treatment and what happens when an opportunity is declined.
Those are proposed improvements to the buying decision, not a formula for outranking another agency. Google's helpful-content guidance asks whether content contributes useful, original value rather than merely repeating existing material. [6]
5. Choose a change that tests a specific explanation
Use an evidence record instead of a list of guessed causes:
| Observation | Working explanation | Bounded action | Verification |
|---|---|---|---|
| Page has no relevant contextual internal link | Customers and crawlers have a weak route to it | Add a useful link from a related service or guide | Confirm the destination and crawlable link in rendered HTML |
| Search results mostly answer a supplier-selection task; our page defines a term | The page may not serve the intended decision | Add scope, process and approval information supported by our actual service | Review whether a buyer can answer the selection questions |
| Google-selected canonical differs from the intended equivalent URL | Conflicting signals may be consolidating another version | Review equivalent pages and align the intended signals | Reinspect indexed canonical after processing |
| Page gets relevant impressions but few clicks | Presentation or result competition may be involved | Review title accuracy and the actual result context | Compare similar query, device and market segments later |
| Visitors reach the form but requests are missing | Contact-path failure may obstruct enquiries | Trace an authorized TEST request through storage and notification | Confirm its record and receipt, separately from search metrics |
On smaller screens, scroll the table horizontally.
The first check is mechanical. Google recommends crawlable links and descriptive, relevant anchor text. That supports discovery and understanding; counting links alone does not establish a ranking diagnosis. [7]
Some faults deserve immediate correction even without a ranking explanation: broken navigation, unreadable mobile content or a failing contact action. Fix them for the people affected. A laboratory speed score, however, is not a complete explanation for poor visibility. Google's page-experience guidance discusses multiple aspects and warns against treating good scores as a guarantee of top results. [8]
For substantial page revisions, group related changes into one documented release. Avoid changing the title, URL, navigation, offer and content repeatedly while trying to attribute every fluctuation to a single edit. Keeping a change log makes later interpretation more defensible, although it does not create a controlled experiment.
6. A worked example: what the evidence permits
The following scenario is invented to demonstrate the method. It is not a Layerly customer case.
A bookkeeping provider's service page is indexed and Google selected its intended canonical. Its selected 28-day page report shows 120 impressions and no clicks. The visible queries concern bookkeeping templates. The business wants enquiries about outsourced bookkeeping for consultancies.
What can we say? The selected report records some search visibility and no clicks. The visible query rows suggest that at least part of the observed visibility concerns templates.
What can we not say? We cannot conclude that all 120 impressions were template searches unless the data supports that grouping. We cannot identify a Google penalty, prove that the domain lacks authority, or predict how many links would solve the problem.
A useful next investigation compares the service page with supplier-selection results. Does it describe the consultancy audience, work included, handover process and appropriate next step? Are its title and main content consistent? Does related content link to it in context?
The proposed release adds missing, truthful service details and a relevant contextual link. Implementation is accepted when the content and link are verified. A later search review asks whether relevant query visibility changed. An enquiry review separately checks actual requests. These are three milestones, not one success claim.
7. Decide when to review or escalate
Keep a review record with the intended query, page, target market, baseline, proposed explanation, change date and next evidence needed. Use the accompanying worksheet to keep implementation checks separate from search and commercial observations.
If a previously successful page has declined, first investigate a traffic drop rather than treating it as a new-page problem. Google's troubleshooting guide includes technical problems, changes in demand, ranking-system changes, security and spam issues. It also notes that changes can take different amounts of time to show effects. [9]
Our proposed review cadence is an implementation check at release, followed by a search-data review after a suitable completed interval. Choose that interval according to traffic and business risk. Where the sample remains too small, report insufficient evidence and keep monitoring; do not manufacture a pass or fail from the date alone.
Before ordering more content or buying links, ask: which observation supports this purchase, what will the work deliver, and how will we distinguish delivery from the outcome we hope for?
Questions business owners ask
Will submitting the page again make it rank?
An indexing request is not a diagnosis of relevance or competitiveness. Check the indexed version and recorded performance first. If you have changed the page, track that change separately.
Should I put the exact question in the URL?
A short, descriptive URL can make the topic clear. Google recommends readable URL words. Renaming an established URL purely to insert a phrase adds migration work without establishing that the name caused the ranking gap. [10]
Should I buy backlinks immediately?
First establish the problem and the expected deliverable. A link purchase does not resolve a wrong page, an irrelevant offer or a broken enquiry path. Assess any later outreach proposal against its audience, editorial fit and terms, rather than using it as a universal repair.
What can Layerly check?
Our public website check examines up to five public pages for supported technical observations. It does not access your Search Console, measure rankings or diagnose a complete SEO strategy. For help reviewing this decision path, send the page, target market and the question you want to be found for through our human-review form: https://layerly.net/website-check/#request
Related guides
- Indexing and canonical decisions: https://layerly.net/insights/noindex-canonical-decision-guide/
- Tracing a missing website enquiry: https://layerly.net/insights/website-enquiry-diagnostics/
- Business changes on an existing domain: https://layerly.net/insights/search-traffic-business-intent/
- Broader discovery and indexing starting point: https://farshidghaffari.net/insights/website-not-showing-on-google/
Sources
- Google: A guide to Google Search ranking systems
- Google: URL Inspection tool
- Google: Canonicalization methods
- Google: Search results Performance report overview
- Google: Troubleshooting Performance data discrepancies
- Google: Creating helpful, reliable, people-first content
- Google: Link best practices
- Google: Understanding page experience
- Google: Debugging Search traffic drops
- Google: URL structure best practices
Sources reviewed 1 November 2026. The matrices, worksheet and review cadence are Layerly working proposals. The bookkeeping example is hypothetical, not a customer result. Prepared with AI assistance and source checking.