Skip to main content

Content Optimization Results: A/B Tested Improvements from FAQ and Category Blurb Generation

Adding unique, helpful content to category pages is one of the most effective ways to improve organic performance. But most e-commerce sites have thousands of pages that need content — far more than any team can write manually.

Similar AI's content generation agents solve this by creating FAQ sections and category blurbs automatically, using product data and ranking keywords. Here's what happened when we tested these content recipes with controlled A/B experiments.

Category Blurbs: 13.3% More Traffic

The first test added ~75 words of category blurb to pages with thin or template-only content. Each blurb was generated using ranking keywords for that specific page, creating content diversity that wouldn't exist from product listings alone.

The Setup

  • Scale: 984 pages across 79,000 words of content
  • Method: Controlled A/B test with test and control groups
  • Measurement period: 5 weeks post-implementation

Results

Metric Test Group Control Group Difference
Organic keywords +8.3% -8.3% +16.6%
Active pages +6.3% -6.3% +12.6%
Impressions +14.9% -14.9% +29.8%
Traffic (clicks) +13.3% -13.3% +26.6%

The test group saw consistent positive impact across all metrics. Pages with blurbs ranked for more keywords, which translated directly to more impressions and traffic.

Interestingly, average rank and CTR decreased slightly in the test group — a sign that pages were picking up additional long-tail keywords where they ranked lower but still captured incremental traffic.

Why It Works

Category blurbs work because they add non-template content that's specific to each page. A "Women's Nubikk Sneakers" page gets content about Nubikk sneakers specifically — materials, styles, price ranges — not generic shoe copy that could apply anywhere.

This specificity helps search engines understand what makes each page unique, which is critical for sites with thousands of similar category pages.

FAQ Additions: 28% CTR Boost

The second test added FAQ sections with 5 Q&A pairs per page. Questions covered important entities like body types, colours, years, mileage, and transmission options — the kinds of questions real buyers ask.

The Setup

  • Scale: 38,000 pages across 2 sites and 2 markets
  • Content generated: 3.6 million words
  • Method: A/B test measuring CTR and keyword rankings

Results

Metric Impact
Click-through rate +28% (when FAQs appeared on SERP)
Ranking keywords +13%
Low-uniqueness pages -89%

Where structured FAQ data appeared in search results, CTR increased by 28%. This was a substantial lift driven by the additional SERP real estate that FAQ rich results provide.

Important Note on FAQ Rich Results

Google has since stopped showing FAQ structured data in search results for most sites. However, the keyword ranking benefits remain:

  • Pages still rank for more keywords because FAQs add non-template content
  • The 13% increase in ranking keywords was independent of SERP appearance
  • Pages with FAQs continue to perform better even without rich results

Example FAQ Content

For a "Used Audi A3" category page, the generated FAQs included:

"What are the most popular body types for Audi A3 cars?"

The most popular body types for an Audi A3 are convertible, saloon and hatchback. If you're also looking for maximum value for money, check out a sports for the most economical options.

"From what year should I buy a used Audi A3?"

Most people who are looking for a second-hand Audi A3 look for one from 2015, 2020 and 2016, but the cheapest years on Gumtree from which you can pick an Audi A3 include 1999, 1997 and 1998.

"What is the average mileage on Audi A3 cars?"

On Gumtree, Audi A3 cars have an average of 76,933 miles on the clock. To give you a guide on pricing, you can expect to pay around £12,103 for an average Audi A3 on our site.

Each answer references actual data from the site's inventory, making the content genuinely useful rather than generic filler.

The Methodology: How We Test

Both tests followed Similar AI's standard A/B testing methodology:

  1. Segment creation: Pages split into test and control groups with similar baseline metrics
  2. Intervention date: Content published to test group only
  3. Measurement window: 5-7 weeks of post-intervention data
  4. Metrics tracked: Keywords, impressions, clicks, active pages, average position, CTR, Googlebot crawls

The experimentation dashboard tracks all metrics automatically, making it easy to see what's working and quantify the impact.

What Makes This Different from Generic AI Content

The key difference between Similar AI's content generation and generic AI tools like ChatGPT:

Generic AI Similar AI
No access to your product data Uses your product catalog and attributes
Can't see ranking keywords Incorporates GSC data on current rankings
Same content could apply to any site Content specific to your inventory and pages
No validation against your blocklist Respects brand voice and excluded terms
Manual copy-paste workflow Server-side API publishes automatically

The result is content that's relevant, accurate, and on-brand — not generic text that could appear on any e-commerce site.

Applying This to Your Site

These results came from classified and e-commerce sites with tens of thousands of category pages. The same approach works for any site with:

  • Large numbers of category or listing pages
  • Thin or template-only content
  • Product data that can inform content generation
  • Enough scale that manual writing isn't practical

If you're running a site with 3,000+ category pages, automated content generation isn't a nice-to-have — it's the only practical way to ensure every page has unique, helpful content.

Summary

Recipe Scale Key Result
Category blurbs 984 pages, 79K words 13.3% more traffic
FAQ sections 38K pages, 3.6M words 28% CTR boost, 13% more keywords

Both recipes delivered measurable improvements in controlled tests. The traffic gains compound over time as pages rank for more keywords and capture more search demand.

"We used to spend weeks writing meta titles and descriptions for new category pages. Now Similar AI generates them automatically using our product data, and they're more relevant than what we were writing by hand because they're based on real ranking keywords."

— Sam Footlik, SEO Specialist, Visual Comfort & Co.


Want to see how content optimization could work for your site? Request a demo and we'll show you examples using your actual product data.

Stay Updated

Get the latest insights on AI-powered SEO, content optimization, and case studies delivered to your inbox.