Content Experiments
Content experiments test how changes to content quality, format, and structure affect search performance and user engagement.
Learning Focus
After this lesson you can test content refresh, expansion, format changes, expert quotes, FAQ blocks, media enrichment, and pruning through controlled experiments.
This lesson covers the seven content experiment types (leaves 10.4.1–10.4.7): content refresh testing, content expansion testing, format testing, expert quote testing, FAQ block testing, media enrichment testing, and pruning tests.
Content Refresh Testing
Core Concept
Test whether updating existing content improves performance.
Refresh test elements:
| Element | Test |
|---|---|
| Full content rewrite | Complete rewrite vs minor update |
| Statistic updates | Update all statistics vs keep existing |
| Date change only | Update "last updated" date vs no change |
| Outdated example replacement | Replace old examples with current ones |
| Internal link updates | Add new internal links vs keep existing |
Content refresh methodology:
- Identify pages with traffic decline (20%+ over 3 months).
- Split into test (refresh) and control (no change) groups.
- Apply refresh to test group.
- Compare: traffic, rankings, engagement over 4-8 weeks post-refresh.
- If positive, refresh remaining pages.
Content Expansion Testing
Test whether adding new sections improves performance.
Expansion test elements:
| Element | Test |
|---|---|
| FAQ section | Add 3-5 FAQ items vs no FAQ |
| "How to" section | Add step-by-step instructions vs current format |
| Case studies | Add case example vs no example |
| Data/statistics | Add statistics section vs no data section |
| Related resources | Add "further reading" section vs no links |
Format Testing
Test content format changes.
Format experiments:
| Test | Control | Expected Impact |
|---|---|---|
| List format | Paragraph format | Featured snippet eligibility |
| Step-by-step | Narrative | How-to rich result |
| FAQ format | Standard article | PAA capture |
| Short form (800 words) | Long form (2,000 words) | User engagement |
| Visual-heavy | Text-heavy | Time on page |
Expert Quote Testing
Test the impact of adding expert quotes.
Expert quote elements:
| Element | Test |
|---|---|
| Quote presence | Expert quote vs no quote |
| Quote source | Industry expert vs internal expert |
| Quote placement | Near top vs near bottom of article |
| Quote quantity | 1 quote vs 3 quotes |
FAQ Block Testing
Test FAQ block implementation.
Note: As of 2024, FAQ rich results are restricted to authoritative government/health sites. Tests are only applicable for qualifying sites.
FAQ test elements:
| Element | Test |
|---|---|
| FAQ presence | FAQ block with FAQPage schema vs no FAQ (on qualifying sites only) |
| Question count | 3 FAQ questions vs 6 FAQ questions |
| Question format | Question in H3 vs in FAQ schema only |
| Answer length | Short answers (1-2 sentences) vs detailed answers (3-5 sentences) |
| Schema implementation | FAQPage schema vs no schema (still visible) — on qualifying sites |
Measurement:
- GSC Enhancements for FAQ rich result impressions (on qualifying sites).
- Engagement on FAQ section (scroll depth, interaction).
- Overall page engagement and ranking.
Media Enrichment Testing
Test adding visual media to content.
Media experiments:
| Media | Test |
|---|---|
| Video | Embedded video vs text only |
| Infographic | Visual summary vs text summary |
| Images | 1 image per section vs text only |
| Interactive element | Interactive chart vs static chart |
| Audio | Podcast clip vs text excerpt |
Pruning Tests
Test whether removing or consolidating content improves overall performance.
Pruning experiments:
| Test | Approach |
|---|---|
| Consolidation | Merge 2+ thin pages into one comprehensive page vs keep separate |
| Redirect | Redirect low-traffic page to parent topic vs keep as-is |
| Noindex | Add noindex to thin page vs keep indexed |
| Remove (410) | 410 low-value page vs keep with noindex |
Pruning measurement:
- Change in index quality for remaining pages.
- Change in crawl budget allocation.
- Traffic redistribution (does traffic move to the consolidated page?).
- Overall site indexation health.
Workflow
- Select a content variable to test (refresh, expansion, format, expert quotes, FAQ blocks, media, or pruning). Write a hypothesis with expected metric impact.
- Identify target pages using GSC and GA4: pages with traffic decline for refresh, pages ranking just below featured snippets for format tests, thin pages for pruning tests.
- Apply the content change to the test group. Keep the control group unchanged.
- Run the test for 4-8 weeks for content changes (content tests take longer to show results than metadata tests). Compare traffic, rankings, and engagement.
- Apply winning content changes to control group and other relevant pages. For pruning tests, monitor whether traffic redistributes to the consolidated page.
Common Mistakes
- Changing content and metadata simultaneously: If you refresh content and update the title tag at the same time, you cannot attribute the result. Change one variable at a time.
- Content refresh without changing the publish date: Updating
lastModifiedwithout meaningful content changes provides no user value and may not trigger recrawling. Always add substantive updates.
warning
- Pruning without a redirect: Deleting or noindexing pages without a 301 redirect to a relevant alternative loses any link equity and can orphan inbound links. Always redirect pruned pages.
- Testing expert quotes from non-experts: A "quote" from an unnamed "industry expert" without credentials or a verified bio may add no value and can reduce trust. Use named, credentialed experts only.
- Waiting too long to measure content experiments: Content changes often take 4-8 weeks for ranking impact. But waiting 6+ months risks confounding from algorithm updates. Set a clear measurement window.
Checklist
- Write a clear hypothesis for the content change being tested
- Identify test and control page groups with equivalent baselines
- Apply the content change to test group only
- For pruning tests, implement 301 redirects from removed pages
- Verify schema changes (FAQ, HowTo) pass Rich Results Test
- Run test for minimum 4-8 weeks
- Compare traffic, rankings, and engagement between groups
- Document findings and apply winning changes to remaining pages