A weekly review is an exception-finding routine, not a full SEO audit. Its job is to catch material incidents, identify meaningful gains or losses, verify important recent releases, and assign a small number of next actions. Most healthy sites do not need every Search Console report investigated every week.
Budget 30–45 minutes for one property. Large or multi-surface sites can automate Performance extraction, but manual health and rollout checks remain necessary.
Choose a stable reporting window
For the routine comparison, use the last seven complete, finalized days against the preceding seven complete days. Matching weekdays controls for normal weekday/weekend behavior.
Also check the same period last year when the business is seasonal and enough history exists. Search Console's interface retains up to 16 months, so preserve external history if longer comparisons matter.
Use the 24-hour view only for an active incident, high-risk release, or time-sensitive publisher workflow. It contains hourly preliminary data that can change, uses the browser's local time, and is not suitable for a finalized weekly scorecard. Normal report dates use Pacific Time.
Step 1: Confirm context before reading charts
Record:
- exact property;
- search type or separate Performance report;
- two date ranges;
- active filters and aggregation;
- releases, migrations, outages, campaigns, and seasonality;
- whether Google lists a reporting anomaly or Search incident.
Open the unfiltered Performance chart before comparison mode to read system and custom annotations. Annotations do not display in comparison mode or the 24-hour view.
Step 2: Check urgent health signals
Open Overview and messages, then confirm:
- Manual Actions: any action is urgent and needs its own documented remediation.
- Security Issues: escalate immediately to the security and engineering owners.
- Page Indexing: look for a new large change in indexed URLs or a spike in a website-caused reason.
- Sitemaps: check newly failed fetches or unexpected submitted/discovered counts.
- Enhancements or Shopping: check abrupt critical-error growth in a business-important feature.
- Core Web Vitals/HTTPS: look for a new template-level regression, not small routine variation.
“Not indexed” is not automatically a problem. Duplicate, redirected, removed, and deliberately excluded URLs can be correct. Investigate changes against the site's intended canonical inventory.
Step 3: Review Performance in a fixed order
Keep Search results, Discover, and Google News separate. They have different eligibility, dimensions, and volatility.
Establish the change
Compare clicks and impressions first. Use CTR and average position as supporting diagnostics.
| Weekly pattern | First questions |
|---|---|
| Clicks and impressions down | Which pages/query groups lost visibility? Demand, ranking, indexing, or reporting? |
| Impressions stable, clicks down | Did CTR change within comparable position, device, country, and appearance? |
| Impressions up, clicks flat | Did visibility expand into lower positions or broader intent? |
| Clicks up, impressions stable | Did CTR, position, brand mix, or result presentation improve? |
| Position changed but clicks did not | Is the movement material enough to matter? Did query mix change? |
Locate concentration
Inspect dimensions in this order:
- Pages: sort the comparison by absolute click difference.
- Queries: filter to material pages and inspect exposed demand changes.
- Countries and devices: identify whether the change is concentrated.
- Search appearance: check business-relevant result features with aggregation caveats.
- Search types: compare Web, Image, Video, and News separately where relevant.
Prioritize absolute impact and business importance. A 500% change from one to six clicks is usually less important than a 10% decline on a high-value landing page.
Query filters remove anonymized queries, and visible rows are top data rather than a complete keyword list. Do not force the filtered rows to equal the unfiltered chart.
Step 4: Verify recent changes
For every important page or template released recently:
- confirm a
200response and intended indexability; - check the declared and Google-selected canonical on representative URLs;
- verify sitemap inclusion and internal discovery;
- inspect whether expected queries and appearances have begun to show;
- compare affected pages with a similar unchanged group where possible;
- add or confirm an annotation for the deployment.
URL Inspection's indexed evidence can lag the live page. Use its live test to validate current access, but do not treat a successful live test as proof of indexing.
Step 5: Triage findings
Use four dispositions:
| Disposition | Meaning |
|---|---|
| Escalate now | Security issue, manual action, sitewide serving failure, accidental blocking, broken migration |
| Investigate this week | Material page group or business segment changed without an adequate explanation |
| Observe | Early, small, preliminary, or plausibly seasonal movement |
| Expected/no action | Deliberate exclusions, completed removals, planned redirects, known campaigns |
Every action should have an owner, evidence, next diagnostic step, and review date. Avoid vague tasks such as “improve SEO.”
What can be automated
The Search Analytics API can automate finalized daily metrics, page/query gains and losses, country/device segmentation, and threshold-based alerts. Cache historical data and use one-day requests for stable extraction.
The public API does not expose most of Page Indexing, Core Web Vitals, HTTPS, Crawl Stats, Manual Actions, Security Issues, Links, annotations, or experimental UI features. Keep a short manual checklist for those reports.
Automated alerts should require both an absolute and relative threshold, enough baseline volume, and more than one data point. This reduces alerts caused by low-volume percentages or preliminary data.
Weekly output template
Property and scope:
Finalized dates compared:
Confirmed incidents/anomalies:
Performance:
- Total material change:
- Concentrated pages/query groups:
- Country/device/appearance concentration:
Health:
- Manual/security status:
- New indexing/sitemap/enhancement issues:
Recent releases checked:
Actions:
1. Finding — evidence — owner — next check
2. Finding — evidence — owner — next check
Observed but not actionable:
Common weekly-review mistakes
- Comparing an incomplete week with a complete week.
- Treating preliminary 24-hour data as final.
- Sorting by percentage change without a minimum-volume threshold.
- Investigating every non-indexed URL instead of intended canonicals.
- Combining Search, Discover, and Google News into one unexplained total.
- Calling a timing correlation a confirmed cause.
- Ignoring the property, filters, or canonical attribution behind a row.
Related content
- The Performance report: complete guide
- Find performance winners & losers
- Reporting anomalies & known issues
- Report completeness & row limits