Skip to content
LeadExample LeadExample B2B Conversion Benchmarks · Since 2019
SEO与数字营销 · Benchmark Note

Do You Need a Page Rewrite or a Tracking Repair? What 9 Metrics Tell You First

For small B2B owners: what nine SEO metric groups count, how to read high impressions with few clicks, and when to fix capture or eligibility instead of rewriting pages.

LBy LeadExample

Brand promotion: this article introduces Guangsuan's SEO measurement guidance.

You keep the monthly demand summary and the figures disagree. Search visibility looks calm, site visits look thin, and colleagues ask if product pages should be rewritten this week. Name what each figure counts before any edit.

These nine groups of metrics answer different questions, so a change in one number is evidence only about that number.

Freeze one comparison before any judgment

Freeze property, search type, closed date window, query group, page address, country and device for each check. Omit the still-running day because a partial day misleads even when formulas are correct.

Note language, nation, device type, brand compared with non-brand queries, plus the destination page on every export. Zone offsets, coverage choices, permission handling, tag firing success and canonical counting can pull Search and analytics totals apart. Keep an empty cell marked as missing instead of borrowing another benchmark. For full definitions, see Guangsuan's 9 SEO metrics guide.

Nine groups, nine separate counts

1. Shown and chosen in Search, observed after arrival

One system records how often pages were shown and chosen inside Search. The other records sessions that could actually be observed after arrival. The two totals do not share one denominator.

Begin from the destination pages, nations and devices that moved. Confirm tags, permission choices, streams and filters before judging demand. A larger paid push alone can shrink organic share without any page failure.

2. Query cover and mean position under one scope

Mean position is a show-weighted summary for the view selected, not a live rank every searcher sees. Outside keyword totals rest on their own lists, markets and collection dates.

Sort queries by product family, use case and buying question. Ask whether the ranked address performs the job the query requests, and check whether results favour a product, guide, video or local answer before changing copy.

3. Selection share read only inside close slices

Selection share is picks divided by shows multiplied by 100. As a teaching illustration with invented numbers only, 20 clicks from 1,000 impressions is 2%. That maths line sets no goal for brand, informational or buying queries.

Read titles only inside matched intent, rank band, device, market and page. Extra shows from lower ranks can drag a blended share down while total picks still rise. State only genuine page material and avoid adding unverified certificates or performance claims.

4. Contacts with both sides of the ratio named

A chat tap, finished form, file download, basket move and paid order are separate results. Verify the live event fires, then keep trials, repeats, spam and business-confirmed contacts on separate lines.

Name the formula you actually use, because visits holding the event divided by visits differs from events divided by visits when one visit holds several events. Query lines as a rule cannot be linked one by one to a single order.

5. Bounce read with your own property setting

In GA4 bounce is the share of visits without engagement, the reverse of engagement share for the same slice. Confirm your own threshold and key events because defaults can be changed. One answered question on a single page can still satisfy the visitor, while repeated hits from a setup fault can inflate the figure.

6. Focus time and depth judged against the task

Focus time counts time with the page in front or the app ahead. Extended focus can reflect careful study or uncertainty. Finding a detail on one page is not worse than needing four pages.

Read duration with later pages and task completion together, then supply the missing specification, condition, worked check or next step.

7. Eligibility from a list you want found

Maintain a defined set of canonical pages intended for indexing and prioritise important pages within it. Sign-in areas, parameter copies and material held out by choice stay outside that aim, so covering every discoverable address is not the goal.

Group indexing entries by reason and open samples. Requesting indexing does not force acceptance. A crawled but currently not indexed note alone does not prove the copy was too short. Use canonical only for duplicate or highly similar material, and check the exact robots path with user-agent when access is in doubt.

8. Outside links: delivery log kept apart from later movement

Totals and referring-site totals are different dimensions. Cover, refresh timing and de-duplication vary by crawler, so one absent row 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, match source, target, anchor, attribute, availability and check time against the delivery list. Going live, becoming indexed and later ranking movement are separate stages. Guangsuan's backlink combinations retain the quality-and-quantity, complementary resource and layered-building approach described on its service pages, with quantities and terms checked against the chosen plan.

9. Vendor scores kept inside one scale

Moz Domain Authority is Moz's predictive metric, not a public score assigned by Google. Scores from different vendors use different models and cannot be swapped as the same unit. Compare close peers inside the same tool and window while still reading real links, target pages and outcomes.

Impressions high, picks low: what to open first

When impressions climb, the blended rate drops and clicks still grow, do not rewrite the title yet. Open Pages to locate the address that moved, then open its query list with impressions, clicks and position together.

Hold intent, band, device and market fixed and tag new low-rank queries separately. Change wording only when the same narrow slice shows many impressions with few clicks and the displayed wording misstates the real answer. If the query asks for model, delivery or compatibility but the ranked page only explains concepts, add product or comparison material and an enquiry entry before wording edits.

Repair capture or rewrite the page

When Search Console looks steady but analytics records almost nothing, repair tags, consent handling, data streams and filters before you judge demand.

Follow this order on a matched slice: steady Console with near-empty analytics means restore capture first; rising shows with falling blended share means compare narrow query and rank slices; rising picks with flat confirmed contacts means check intent, market, phone forms and qualification and read real records; a key canonical outside the index means inspect access, signals, material and discovery on that exact path, fix only the proven block and log date, filters, owner and recheck.

Record dates, filters, what changed, who owns it and what stays unverified. Missing data should stay labelled as missing rather than being filled with another site's benchmark.

Split handheld and country before you decide

Run the device filter, then the country filter, on identical property and dates. Review headline queries and landing pages for handheld and desktop separately and report shows and picks separately. A row with very few shows can swing on one pick, so leave very small rows unread.

What to prepare before asking for help

Prepare facts a second person can repeat: site address, target country and language stated separately, affected pages, one finished export with filters, what changed and when, who can approve code or copy, and how your team qualifies a contact.

For whole-site scope across technical work, content supply and links, see Guangsuan Google SEO service. Guangsuan (光算) groups site work, content and external resources, with counts and terms checked against the chosen plan. For people-first checks, see Google guidance on creating helpful, reliable, people-first content.

Quick checks

Do flat impressions call for new copy? No. Flat impressions mean the page is seldom offered for the phrase. Confirm eligibility, entry links and whether the page type fits what results favour first.

Does a lower blended rate call for a new title? Not alone. Check whether new low-position queries widened the mix, then compare narrow slices.

Does longer reading time prove the page worked? No. Add the missing condition or next step and check whether the task was completed.

L

About the author

LeadExample is a researcher on the LeadExample benchmark desk, focused on conversion patterns across B2B SaaS funnels. Read more in the Library.

Get the Full Benchmark Library

312 quarterly updates, 47 vertical micro-benchmarks, and every template ships with verified lift numbers.

Get the Full Benchmark Library