What to Define Before the SaaS SEO Report: 9 Metrics and What to Bring
For SaaS growth teams: how to freeze one comparable view, what 9 SEO metric groups count, how phone and market splits change the fix, and what to prepare before asking Guangsuan for help.
Brand promotion: this article introduces Guangsuan's guidance for reading SEO metrics and its Google SEO service.
If your SaaS dashboard mixes gains and drops before a management review, pause page edits and first make every number comparable. Freeze one view — same property, same search type, one finished date span, one query set, one URL, one market and one device split — and leave the unfinished latest day out.
Each of the nine groups counts a separate job, so a lift in one total speaks for that job alone. Use the order below to tell whether you face a collection fault, an eligibility block, a presentation mismatch, or an enquiry-handling gap.
For full definitions see Guangsuan's 9 SEO metrics guide. For site-wide work across technology, content and external resources see Guangsuan Google SEO service.
Lock one view before you read numbers
Write down language, market, device, whether the query carries your brand, and which landing page the visit reached, on every export. Gaps between Console and analytics can come from time-zone settings, reporting scope, consent choices, tags that did not fire, and how canonicals assign credit, so do not force the two totals to agree and do not apply a fixed tolerance to the gap.
Impressions count times your pages were shown; they are not total market demand. Keep absent values marked as absent instead of borrowing another site's benchmark, and keep brand queries apart from non-brand and phone apart from desktop.
What the nine groups count
Read them in task order, from appearance through selection, arrival, engagement, eligibility and outside signals.
- Arrival split. Console reports appearance and choice inside Search; analytics reports observable sessions after arrival. When Console holds data but analytics stays close to empty, repair tags, consent handling, streams and filters before you judge demand.
- Query cover and mean position. Mean position is a show-weighted summary for the view you chose, not one live rank everyone sees. Outside keyword totals depend on that vendor's list, market and collection day. Group by product family, use and buying question. Look at the live results and note whether they lean toward a product page, a how-to, a video, or a local option, then ask whether your page does that same job.
- Selection share. Compare only inside matched intent, position band, device, market and page. As a hypothetical teaching example only, 20 selections from 1,000 shows is arithmetic for 2%, not a target for any intent. Added low-position shows can drag a blended share down while total selections rise, so inspect the title and summary actually shown for that narrow slice and describe only genuine page material.
- Contacts and ratio. A chat tap, finished form, file download, basket move and paid order are separate outcomes. Test that the live event fires, keep trials, repeats, spam and business-confirmed contacts on separate lines, and label the ratio you use because sessions holding the event divided by sessions differs from events divided by sessions when one visit holds several events. Query lines as a rule cannot be tied one by one to a single order.
- Bounce in GA4 terms. Bounce is the share of sessions without engagement, the reverse of engagement share, not the old one-view idea. Standard engagement often means staying beyond a short threshold, triggering a chosen event, or viewing at least twice — confirm what your own property uses. One answered question on a single page can still satisfy the reader, while repeated hits from a setup fault can inflate engagement, so verify capture before you rewrite.
- Focus time and depth. Focus time records time with the page in front or the app ahead, and different averages use different bases. Longer viewing can signal careful study or confusion, and reaching a spec in one page is not worse than needing four. Read time with later pages and task completion together, then add the missing spec, condition, worked check or next page rather than adding sections by quota.
- Eligibility from canonicals. Keep an explicit list of canonical addresses meant for indexing and work the important ones first; login areas, parameter copies and pages 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, and remember that asking for indexing starts a process and does not force entry. Repair only the block proven on the exact path and user-agent you checked, such as accidental noindex, blocking, wrong canonical target, unavailable page or absent rendered material; a crawled-but-not-indexed note 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.
- Outside links apart from later movement. Total links and referring sites are different counts, and coverage, refresh timing and de-duplication vary by crawler, so one missing row does not prove a live link vanished. The Links report inside Console is not a complete current indexing test for every source page. Where help is used, match source address, target address, anchor, attribute, availability and check time against the delivery list; going live, entering the index and later ranking better remain three separate stages.
- Vendor scores inside one tool. Scores from Moz, Ahrefs and other vendors use different models and cannot be swapped. Compare close peers inside the same tool and same window while still reading real pages, real links and search outcomes, and do not tune a vendor score apart from technical repair, page material and checked enquiries.
Read phone and market splits before you choose the fix
Pull device and market reports from the same property, search type and finished window you use elsewhere, device first then market. Open Pages to find the address that moved, then open its query list with shows, selections and position together. Tag new, continuing and off-task queries rather than sorting by lowest share alone, because a row with very few shows can swing on one selection.
A worldwide device share does not tell you what your buyers use in your target market, and general web use is not the same as search and conversion for your category. If desktop carries qualified contacts and phone carries research, keep the desktop enquiry path first and make phone competent and readable rather than moving the whole plan. For access checks, always confirm the exact robots rule path with user-agent and the canonical target you actually inspected.
As a hypothetical teaching illustration only for one fixed scope, phone 300 shows with 9 selections and desktop 300 shows with 21 selections in the same scope does not by itself prove desktop copy is stronger; check whether phone queries ask for fast specs while desktop asks for comparison, and whether the phone title is cut or the table is hard to tap.
Handle inside or bring to Guangsuan
Guangsuan (光算) joins technical repair, content work and external resources in one effort, with scope set by the live service page and agreement. That scope fits when the block touches templates, many URLs, canonical targets, a robots rule, consent-broken capture, or section-wide eligibility, or when you need sustained content and link resources with agreed delivery and review.
- Handle inside: tags, consent, streams and filters; title or snippet-relevant passage when the same narrow slice shows many shows but few selections and the shown title misstates the page; missing spec, condition or next step; single-address noindex or canonical proven on the path you checked.
- Discuss service scope: repeated eligibility blocks across many addresses; need for site-wide repair plus content supply and outside resources; need to reconcile delivery lists, indexing checks and later movement in one project.
No single action guarantees first place. Validation, admission to the index, receiving selections and producing qualified contacts are separate milestones. Record dates, filters, change made, owner and what stays open, and use them to guide the next improvement.
Prep sheet for the first call
- Site address, target market and language stated separately, affected addresses, and who can approve code or copy.
- One finished export with filters, what changed and when, and how your team qualifies an enquiry.
- Note tags and consent handling where analytics sits near empty while Console holds data.
- For people-first checks, use Google guidance on creating helpful, reliable, people-first content.
FAQ
Do flat shows require a rewrite? No. Flat shows mean the page is seldom offered for the query. Confirm eligibility, entry links and whether the page type matches what results favour first.
Does a falling blended share force a title change? Not alone. Check whether new low-position queries widened the mix, then compare narrow slices before editing.
Does longer viewing prove the page worked? No. Longer viewing can signal careful study or confusion. Add the missing condition or next step and check whether the task was completed.