Every service, joined on the page
Insights merges the latest run of all ten checks on a shared page key, scores every URL and ranks what to fix by the traffic it carries. Narrative reports explain the rest in plain language.
Insights is where the latest results of every SiteRefiner service meet. Each page on a site gets a health score and a priority rank, you get one list of what to fix first, and narrative reports explain SEO opportunities and ad spend in plain language a client can read. Keywords and Spend & Return sit alongside it.
One score per page. One list per site. Technical findings, Lighthouse, Search Console and GA4 land on the same row, so a broken page with real traffic rises to the top on its own.
A score you can explain to a client
Every page scores 0 to 100. Critical findings cost far more than warnings, and the page's mobile performance counts too, so one number says how healthy a page is and the list beneath it says why.
- Critical findings COST THE MOST
- Noindex in the sitemap, broken internal links, missing titles, canonical pointing elsewhere.
- Warnings
- Missing alt text, duplicate descriptions, long redirect chains, content invisible without JavaScript.
- Mobile performance PSI
- Lighthouse on mobile, folded into the health score alongside the technical findings.
Findings come from the latest Site Audit and Link Crawl runs on the page. Fix one, re-run, and the score moves.
One page key
Every finding is normalised to the same path, so /probate from Site Audit, Search Console and GA4 is one row, not three.
Fix the page that carries the traffic
Priority weighs how broken a page is against how much search visibility and traffic it holds. The top ten sit on the Insights page with their issues and a recommendation each, recomputed whenever a new run lands.
What moves a page up
- How broken it is SITE AUDIT · LINK CRAWL · PSI
- A noindex or a 404 on a page with traffic jumps the queue. Slow mobile performance drags it up too.
- Search visibility GSC
- How many clicks and impressions the page holds today.
- Traffic GA4
- Sessions from every channel, not only organic search.
Recomputed whenever a new run lands.
Reports that answer back
SEO Opportunity and Google Ads runs end in a written report: a headline, three to eight prioritised recommendations with impact and effort, quick wins and notes per category. Ask follow-up questions in the same thread.
Striking-distance queries are the fastest win this month.
14 queries sit in positions 8 to 20 with 3,900 monthly impressions between them. Two service pages lose clicks to weak snippets, and one high-intent topic has no page at all. Lifting the top five queries onto page one is worth roughly 310 extra clicks a month.
- Shorten the /pricing title, cut off at 68 characters in results
- Add "open Saturday" to the /contact meta description
- Point the old /teeth-whitening URL at its replacement
Mostly service terms one or two places off page one.
Ranking holds; the snippet is losing the click.
Cost questions with no page that answers them.
Old posts with rising impressions and flat clicks.
Spend with zero conversions. Add as negatives or tighten match types.
Paid terms where the site ranks top 3. Test pausing them.
Lowest cost per conversion with room to grow.
Organic wins that would replace paid clicks over time.
Never invents numbers
Every figure in a report comes from the run's own data. A sizing is labelled as a sizing.
Ask in the thread
Follow-ups stay in the run's own thread, so the context never has to be re-explained.
Two written reports
One for SEO opportunities, one for keyword and budget strategy. Every other service reports through Insights.
Sized in clicks, not adjectives
SEO Opportunity cross-references Search Console and GA4 into four buckets and sizes every opportunity in potential clicks, which is what feeds the narrative.
The four buckets
- Striking distance
- Queries ranking in positions 8 to 20, close enough that a fix moves them onto page one.
- Low CTR
- Click-through well below what the position usually earns. The snippet is the problem.
- Content gaps
- Queries with impressions and no page that matches them.
- Underperforming
- Pages whose visibility is rising or steady while clicks are not.
One keyword universe per site
Keywords arrive from Search Console queries, Google Ads, content targets and manual entry, deduplicated automatically. Each carries a position trend and, new in September, an estimated search intent. Track up to 50 per site.
| Keyword | Sources | Intent | Position | Trend, 12 weeks | Volume | Tracked |
|---|---|---|---|---|---|---|
| electric kiln uk | GSCAds | Commercial | 6.4 | 2,900 | ||
| pottery kiln for beginners | GSCContent | Informational | 11.8 | 1,300 | ||
| how hot does a kiln get | Content | Informational | 17.3 | 2,400 | ||
| kiln wash | GSC | Informational | 4.2 | 880 | ||
| second hand kiln | GSCAds | Transactional | 9.1 | 1,600 | ||
| buy kiln shelves | Ads | Transactional | 23.5 | 590 | ||
| kiln and co | GSC | Navigational | 1.0 | 320 | ||
| glaze firing schedule | Manual | Informational | Not ranking | 720 |
Buy, build or own
Spend & Return joins organic position and clicks with paid cost and conversions per keyword, then derives organic value, overlap, wasted spend, CPA, ROAS and blended CAC. The planner turns that into a split you can act on.
Northpoint Cycles. Illustrative data.
Six buckets
- Buy
- Paid clicks pay back and organic is out of reach.
- Build
- Close enough organically that content beats spend.
- Both
- Fund it now while content closes the gap.
- Own it
- Ranks top 3 with no spend. Leave it alone.
- Overlap
- Ranks top 3 and still paid for. Test pausing.
- Skip
- Neither pays back at this value per conversion.
Plan a month of spend
Return funds only the keywords that pay back most. Balanced goes further down the list. Conversions keeps buying as long as a keyword still converts.
Keyword split
Funded keywords
| Keyword | Bucket | Spend | Clicks |
|---|
Illustrative data and a simplified model over 20 sample keywords.
Vitals, channels and funnels
Page Speed Insights reads Core Web Vitals for every page on mobile and desktop. GA4 brings channels and engagement over a rolling 28 days, and the funnel check follows the GA4 events you define.
Questions
What is the Insights view?
One page per client site where the latest results of every service meet. Each URL carries a health score and a priority rank, and the top pages to fix sit at the top with their issues and a recommendation each.
How is a page's health score worked out?
From the technical findings of the latest crawls and the page's mobile performance. Critical findings weigh far more than warnings, so one number says how healthy a page is and the list beneath it says why.
How does SiteRefiner decide what to fix first?
By weighing how broken a page is against how much search visibility and traffic it holds. A broken page with real traffic rises to the top on its own, and the ranking is recomputed whenever a new run lands.
What is a narrative report?
A written report at the end of an SEO Opportunity or Google Ads run: a headline, prioritised recommendations with impact and effort, quick wins and notes per category. Every figure comes from the run's own data, and you can ask follow-up questions in the same thread.
Where do keywords come from?
Search Console queries, Google Ads, content targets and manual entry, deduplicated into one list per site. Each keyword carries a position trend and an estimated search intent, and up to 50 per site can be tracked.
What does Spend & Return show?
Organic position and clicks next to paid cost and conversions for each keyword, so you can see what the site already owns, where it pays for clicks it could earn and where spend never converts. A planner turns that into a split you can act on.