Skip to content
Signal network online · 14 chains Request Demo
SEO与数字营销

9 SEO Metrics: Definitions, Segments and What to Fix First

PBy PFN DaiPFN Dai

For analytics teams: what nine SEO metric groups count, how to lock country and device scope, and whether to fix measurement, eligibility, presentation, or enquiry path before changing a product page.

Brand promotion: this article introduces Guangsuan's Google SEO services for readers evaluating product-page changes.

Short answer with conditions: these nine groups of metrics answer different questions, so a move in one number is evidence about that number only. Before you change a product page, freeze one comparable view — same property, same search type, one finished date range, same query group, same page, same country and same device — then check recording, then eligibility, then presentation and enquiry path. For method detail, see Guangsuan's 9 SEO metrics guide. Where repair touches templates or many URLs, whole-site scope is described in Guangsuan Google SEO service.

Freeze the comparison before you read anything

Use one property, one search type and one finished date range for every export. Leave the unfinished latest day out. Record language, country, device, branded versus non-branded queries, and the relevant landing pages for each export. Keep brand queries apart from non-brand queries and read mobile apart from desktop. Leave a missing row marked as missing rather than filling it with another benchmark.

GA4 measures observable activity after people reach your site. Search Console reports visibility and selections inside Google Search. Time zones, scope, consent, successful tag execution and canonical attribution can all create differences between the two totals, so do not require them to match. When Console holds data while analytics sits near empty, restore tags, consent handling, data streams and report filters before judging demand.

What each of the nine groups counts

1. Traffic leaving Search versus activity observed on site

Locate the landing pages, countries and devices driving the change. Verify measurement before judging seasonality, availability, ranking or wording. A smaller organic share can also reflect larger paid spend.

2. Query coverage and rankings under fixed conditions

Average position is a presentation-weighted aggregate for the selected scope, not one live rank seen by every searcher. A third-party ranking-keyword count depends on its own database, market and collection dates. Group queries by products, applications and purchasing questions. Ask whether current results favour a product, tutorial, video or local service, then review whether your page answers that same task.

3. Selection rate compared only inside close groups

As a hypothetical teaching illustration only, with invented figures, 20 clicks from 1,000 impressions is 2%. That line shows arithmetic. It is not a target or a client result. Compare only inside matching intent, position band, device, market and page. Added low-position impressions can pull a combined rate down while total clicks rise. Improve it: write a title that accurately communicates the page's answer and useful distinction, and describe only genuine page material.

4. Contacts and ratios with both sides labelled

A chat tap, finished form, download, basket move and paid order are different outcomes. Test the actual event implementation, then distinguish tests, duplicates, spam and genuine business enquiries in your operational records. Label the calculation you actually use, because sessions holding a key event divided by sessions differs from events divided by sessions when one visit holds several events. Review organic landing-page sessions, sessions with the intended event, and business-verified leads together. Queries as a rule cannot be tied one by one to a single order.

5. Bounce, focus time and depth tied to the task

Use the GA4 definition fixed in your own property. Default engagement criteria include lasting more than 10 seconds, having a key event, or at least two page or screen views; check the property's actual configuration. A single-page visit may answer the reader's question. Duplicated page_view hits can make engagement look artificially strong, so check measurement first.

GA4 user engagement time measures time with the web page in focus or the app in the foreground. Needing four pages to find a specification is not necessarily better than finding it on one. Longer time can mean useful study or confusion. Interpret time, later pages and task completion together, then add the missing specification, condition, example or next step.

6. Eligibility starting from pages you want found

Maintain a defined set of canonical pages intended for indexing and prioritise important pages within it. Login areas, parameter copies and material deliberately kept out remain outside that aim, so covering every discoverable URL is not a useful goal. Group Page indexing report entries by their reasons and inspect representative URLs. Improve it: fix accidental noindex directives, blocking, wrong canonical targets, unavailable pages and missing rendered content. Asking for indexing does not force admission. A crawled-but-not-indexed state alone does not prove thin copy, and a correctly merged duplicate is not by itself a removal job. Use canonical only for duplicate or highly similar material.

7. Links and third-party scores kept apart from later movement

Total links and referring sites are different counts. Coverage, refresh timing and de-duplication vary by crawler, so a missing row in one report does not prove a live link vanished. Google's Links report is not a complete current indexing test for every source page. Where outside help is used, reconcile source URL, target URL, anchor, attribute, availability and check time against the delivery list. Published, indexed and improved target-page rankings are separate milestones.

DA, Ahrefs DR and other tools' metrics come from different models and are not interchangeable units of link equity. Compare peers inside the same tool and time window while reading actual links, target pages and search outcomes.

How device and country segments change the decision

Work from your own Search Console device and country reports on identical property and dates. For September 2026 worldwide, its Desktop vs Mobile Market Share report shows Mobile 60.04% and Desktop 39.96%. That is a real, dated, checkable figure, and it is a share of web traffic by device, not a statement about your industry, your country, or your buyer. It shifts by country and year.

Apply the device filter and look at top queries and landing pages separately for desktop and mobile. A split that looks uneven overall can be even for one product line. It is normal for the same page to rank differently on mobile and desktop. Decide which question you are asking: where are we seen with impressions, or where do people act with clicks. Report both, and treat a row with very few impressions as noise.

Worked check with invented teaching numbers only for one fixed scope: mobile 300 impressions with 9 picks and desktop 300 impressions with 21 picks in the same scope does not by itself prove desktop copy is stronger. Check whether mobile queries seek fast specs while desktop queries seek comparison, and whether the mobile title truncates or the spec table is hard to tap on a phone.

Fix measurement, eligibility, presentation, or enquiry path first

Follow capture, then eligibility, then presentation, then contact path, then outside resources. Record dates, filters, the change made, its owner and what remains unverified.

If Console is steady while analytics is near empty, restore reliable capture before judging demand. If impressions are up, blended rate is down and clicks are up, check query and position mix and compare narrow segments before rewriting titles. If clicks are up while confirmed contacts stay flat, review intent, market, forms and qualification and walk the product path with actual enquiry records. If an important canonical page is outside the index, check access, signals, material and discovery path, fix the proven block and log a recheck.

What to bring before asking Guangsuan for technical and content diagnosis

Bring facts another person could repeat: site address, target country and language stated separately, affected addresses, one finished export with filters, what changed and when, who can approve code or copy, and how your team qualifies an enquiry. Note tags, consent handling and report filters where analytics is near empty while Console holds data.

That is the point to ask about technical repair, content supply and external resources together if checks point past one page, such as a template canonical target, a robots rule for an exact path and user-agent, consent breaking capture, or section-wide eligibility. Resource metrics and delivery criteria specified in a Guangsuan product remain governed by that product's actual terms.

FAQ

Do flat impressions require a rewrite? No. Flat impressions mean the page is seldom offered for the query. Confirm eligibility, entry links and whether the page type matches what results favour for that task first.

Does a falling blended rate force a title change? Not alone. Check whether new low-position queries widened the mix, then compare narrow segments.

Does longer time prove the page worked? No. Longer time can mean careful study or confusion. Add the missing condition or next step and check whether the task was completed.

Allocate capital with intent.

PFN Dai fuses transformer-based on-chain analytics with proprietary DeFi signal models — engineered for funds that don't guess yield.

Request a Platform DemoView Audited Performance