Performance Diagnosis
Performance diagnosis is the structured process of identifying why SEO metrics changed — especially when they declined. A systematic diagnosis prevents wasted time on incorrect root causes and ensures the right fix is applied.
Learning Focus
After this lesson you can systematically diagnose traffic drops, ranking losses, algorithm updates, cannibalization, indexing issues, conversion loss, and technical regressions.
This lesson covers the seven diagnosis areas (leaves 2.5.1–2.5.7): traffic drop analysis, ranking loss analysis, algorithm update analysis, cannibalization diagnosis, indexing issue diagnosis, conversion loss analysis, and technical regression diagnosis.
Why This Matters
Core Concept
- SEO metrics change constantly. Most changes are noise. Diagnosis separates signal from noise.
- Without a systematic process, teams react emotionally to traffic drops and implement fixes that do not address the root cause.
- A documented diagnosis process makes it easier to communicate findings to stakeholders.
Traffic Drop Analysis
Traffic drops are the most common SEO performance issue. They can have many causes, and the correct response depends on identifying the right one.
Traffic drop diagnosis workflow:
- Confirm the drop: Check GSC (impressions and clicks) and GA4 (organic sessions). Are both dropping, or just one? GSC data changes first (ranking change), GA4 data follows (traffic change).
- Isolate the timing: When did the drop start? Exact date matters for correlating with possible causes.
- Segment by brand/non-brand: Is the drop in brand traffic, non-brand, or both?
- Segment by device: Is the drop isolated to mobile or desktop?
- Segment by country: Is the drop global or market-specific?
- Check for seasonality: Does the drop coincide with previous-year seasonal patterns?
- Check for SERP changes: Did SERP features change for key queries (new snippets, PAAs, AI Overviews)?
- Check for ranking changes: Did specific keywords drop in position?
- Check for technical issues: Server errors, robots.txt changes, noindex tag, canonical changes, site structure changes.
Common causes and signals:
| Cause | Signal | Check |
|---|---|---|
| Algorithm update | Drop coincides with confirmed Google update | Check Google Search Status Dashboard, algorithm trackers |
| Technical issue | Drop is sudden, affects all pages | Check GSC crawl errors, server logs, robots.txt |
| Competitor improvement | Drop specific to certain keywords | Competitive rank tracking comparison |
| Content change | Drop started after a site update | Check content release dates vs traffic drop date |
| Seasonal pattern | Drop matches same period previous year | Year-over-year comparison |
| SERP feature change | Impressions may not drop but clicks drop | Check search appearance report in GSC |
| Indexing loss | Pages removed from index | Check GSC Indexing report |
Ranking Loss Analysis
Ranking loss analysis focuses specifically on position changes for tracked keywords.
Ranking loss diagnosis workflow:
- Identify affected keywords: Which keywords lost position? Not all — which ones specifically?
- Segment by intent: Is the loss concentrated in informational, commercial, or transactional queries?
- Segment by page type: Are specific page types losing rankings?
- Check affected pages: Did the specific ranking pages change (content updates, metadata, URL changes)?
- Check competition: Did competitors improve content, gain backlinks, or launch competing pages?
- Check for cannibalization: Are multiple pages from your site competing for the same query?
Ranking loss types and responses:
| Loss Type | Pattern | Response |
|---|---|---|
| Gradual decline over weeks | Content decay, competitive loss | Content refresh, internal link update |
| Sudden drop across many keywords | Algorithm update or penalty | Review Google guidelines, check manual actions |
| Drop for specific keywords only | Competitor improvement or page-level issue | Improve specific page, review competitor content |
| Mobile-only drop | Mobile usability or rendering issue | Test page on mobile, check mobile rendering |
| Local drop | GBP issue or local competitor improvement | Check GBP listing, review local competitor content |
Algorithm Update Analysis
Algorithm updates can cause ranking and traffic volatility. Knowing how to analyze an update separates signal from noise.
Algorithm update analysis workflow:
- Confirm the update: Check Google Search Status Dashboard, industry trackers (MozCast, Semrush Sensor, Algoroo, Accuranker Volatility).
- Identify affected pages: In GSC, compare the period before and after the update date for impression/click changes.
- Identify affected queries: What query types were most affected?
- Check relevant Google guidance: If the update targets specific areas (helpful content, reviews, core), review the corresponding Google guidelines.
- Audit affected pages: Do they violate the relevant guidelines? If yes, fix and wait for recovery. If no, the impact may be temporary or related to other factors.
Response to algorithm updates:
| Update Type | Recommended Response | Not Recommended |
|---|---|---|
| Core update | Improve page quality, E-E-A-T indicators, content usefulness | Making drastic changes based on speculation |
| Helpful content update | Remove or improve content that does not demonstrate first-hand expertise | Adding more low-value content to compensate |
| Reviews update | Ensure reviews include original research, evidence of product use | Removing reviews that are legitimate but not "in-depth" |
| Spam update | Remove content that violates Google guidelines | Ignoring guideline violations |
| Link spam update | Disavow toxic links if manual action; otherwise, continue natural link building | Mass disavow of legitimate links |
warning
Key nuance: Most ranking fluctuations during algorithm updates are not permanent. Wait 1-2 weeks after the update completes before making significant changes, unless a manual action is confirmed.
Cannibalization Diagnosis
Cannibalization occurs when multiple pages from your site target the same or very similar queries, competing against each other for rankings.
How to detect cannibalization:
| Method | How It Works | Tools |
|---|---|---|
| GSC query-to-page mapping | If one query returns multiple URLs from your site | GSC Performance report (click on a query to see pages) |
| Keyword ranking overlap | Track which pages rank for the same keyword | Rank tracker with per-query page tracking |
| Content similarity | Pages covering the same topic with high content overlap | Manual review or content analysis tools |
| Internal link confusion | Multiple pages targeting the same topic without clear hierarchy | Site audit, content inventory |
Cannibalization patterns:
| Pattern | Example | Solution |
|---|---|---|
| Two blog posts on the same topic | "email segmentation guide" and "how to segment email lists" | Consolidate into one page, redirect the other |
| Product vs category page | Product page ranks for a category query | Strengthen category page, deoptimize product page for that query |
| Blog post vs service page | Blog post ranks for a service query | Add service page targeting, remove intent from blog |
| Pagination cannibalization | Page 2 ranks for same query as page 1 | Use canonical or view-all page |
Cannibalization fix priority:
- Determine which page should rank for the query (the one that best matches intent and has the highest conversion potential).
- Consolidate content onto that page.
- Implement a 301 redirect from the other page(s).
- Update internal links to point to the consolidated page.
- Monitor to confirm the remaining page maintains or improves ranking.
Indexing Issue Diagnosis
Indexing issues prevent pages from appearing in search results even though they are crawlable and high-quality.
How to diagnose indexing issues:
- Check index coverage in GSC: Review the Indexing report for errors, warnings, and "crawled - not indexed" entries.
- Check sitemap submission: Are the pages that should be indexed included in your sitemap?
- Check canonical tags: Are pages self-canonical or pointing to a correct canonical?
- Check noindex tags: Are valuable pages incorrectly marked noindex?
- Check robots.txt: Is valuable content blocked from crawling?
- Check for crawl budget issues: Are too many low-value URLs consuming crawl budget and preventing important pages from being crawled?
Common indexing issues and fixes:
| Issue | GSC Signal | Fix |
|---|---|---|
| Page not in sitemap | "Indexed, not submitted in sitemap" | Add to sitemap |
| Duplicate page | "Duplicate without user-selected canonical" | Add self-canonical or consolidate pages |
| Crawled but not indexed | "Crawled - currently not indexed" | Improve content quality, reduce similarity to other pages |
| Blocked by robots.txt | "Blocked by robots.txt" | Update robots.txt to allow crawling |
| Noindex tag | "Excluded by noindex tag" | Remove noindex if page should be indexed |
| Server error | "Server error (5xx)" | Fix server configuration |
| Soft 404 | "Soft 404" | Return 404/410 or add substantive content |
Conversion Loss Analysis
Conversion loss analysis identifies when organic traffic stops producing results at the expected rate, even if traffic volume is stable.
Conversion loss diagnosis workflow:
- Confirm the loss: Is the conversion rate declining, or is traffic declining (which would reduce absolute conversions)?
- Segment by landing page: Which specific pages lost conversion rate?
- Segment by traffic source: Is the decline in organic specifically, or across all channels?
- Check for technical issues: Page speed regression, checkout errors, form failures.
- Check for content drift: Did page content change without updating the conversion path?
- Check for competitive pressure: Are new competitors capturing the same conversion intent?
Conversion loss patterns:
| Pattern | Likely Cause | Fix |
|---|---|---|
| All channels declining | Site-wide issue (checkout, form, pricing) | Technical audit of conversion path |
| Organic only declining | Traffic quality change — new organic visitors may be less qualified | Review ranking keyword changes, content targeting |
| Specific page declining | Page-level issue — content, CTA, or experience changed | Review page changes, competitive landscape |
| Seasonal conversion loss | Normal pattern | No action needed; document seasonality |
Technical Regression Diagnosis
Technical regression diagnosis identifies when previous SEO work breaks or new deployments cause technical issues.
Technical regression triggers:
| Trigger | What to Check |
|---|---|
| CMS or platform update | Meta tag handling, sitemap generation, schema output, canonical logic |
| Theme or template update | Heading structure, schema markup, internal linking, meta data fields |
| Plugin or extension update | Schema conflicts, sitemap changes, redirect management |
| Content management system change | URL changes, redirect mapping, content migration |
| CDN or hosting change | Server response times, caching, SSL certificate |
| JavaScript framework change | Content rendering, link discoverability, meta tag generation |
Technical regression monitoring:
- After any deployment affecting templates, CMS, or infrastructure, monitor:
- GSC crawl errors (spike in 4xx or 5xx)
- Index coverage (drop in indexed pages)
- Core Web Vitals (LCP, INP, CLS changes)
- Sitemap validity
- Schema validation (GSC enhancements reports)
- Before deployment, validate in staging:
- Meta tags render correctly
- Canonical tags are correct
- Schema markup validates
- No unintended noindex additions
Regression recovery workflow:
- Identify the deployment that caused the regression (check deployment log vs regression date).
- Roll back if the regression is critical and deployment is not essential.
- Fix the issue in development and re-deploy with validation.
- Monitor post-fix to confirm recovery.
Workflow
- Detect: Metric change is detected (automated alert or manual review).
- Isolate: Determine the scope — which pages, queries, segments, and timelines are affected.
- Diagnose: Work through the relevant diagnosis workflow above.
- Decide: Determine the fix, rollback, or monitor response.
- Act: Implement the fix.
- Verify: Monitor to confirm recovery.
- Document: Record the incident, root cause, fix, and recovery timeline for future reference.
Common Mistakes
- Reacting to normal fluctuations: Daily and weekly ranking fluctuations are normal. Only investigate when the change exceeds normal volatility for 2+ weeks.
- Attributing drops to the first cause you find: Correlation is not causation. A traffic drop that coincides with a core update may be caused by competitive changes, not the update itself.
warning
Fixing symptoms, not root causes: Improving CTR on a page that cannot convert does not solve the conversion problem. Trace the full path.
- Not documenting diagnosis steps: Without documentation, you cannot learn from previous incidents.
- Failing to check for technical issues first: A simple server error or noindex tag can explain a traffic drop. Check technical issues before assuming competitive or algorithm causes.
Checklist
- Traffic drop workflow is documented and followed for any significant decline.
- Ranking loss analysis segments by keyword, page, and intent.
- Algorithm updates are correlated with traffic changes for analysis.
- Cannibalization audit is run quarterly.
- Index coverage is checked weekly.
- Conversion loss analysis distinguishes traffic decline from conversion rate decline.
- Technical regression is verified after every deployment.
- All incidents are documented with root cause, fix, and recovery timeline.