How to Improve CTR in Google Search Console

If a page ranks top-5 but gets fewer clicks than pages usually get at that rank, something about its SERP snippet is losing clicks. Here's how to use GSC to find those pages, and how to rewrite the meta description to recover them.

Why "expected CTR" is the useful metric

Industry CTR benchmarks are useless for your site. A top-3 informational query averages 25-35% CTR for one site and 8-12% for another, depending on the SERP layout (featured snippets, People Also Ask, shopping blocks, AI overviews). What matters is your own curve: what CTR does your site average at position 3? At position 5? If a given page is well below your personal curve for its position, its snippet is the suspect.

Our CTR Benchmark report computes the curve automatically and flags outliers. For the raw definition of CTR in GSC, see GSC CTR explained and What is a good CTR in GSC?

Step-by-step

Open the CTR Benchmark report. Opportunities → CTR Benchmark.
Filter to pages with at least 500 impressions in 90 days and an average position ≤ 10. Below position 10, CTR is noise.
Sort by "CTR vs. expected" ascending. The biggest under-performers rise to the top.
Pick one. Open the page. Run a Google search for its top query in an incognito window. Look at your actual SERP snippet. What's unconvincing about it?
Check the meta description. 80% of the time it's missing, truncated, generic, or doesn't match the query. That's your fix.
Rewrite the meta description. Formula: [what the page is] + [specific benefit or proof] + [soft call to action]. 140-160 characters. Include the target query once, naturally. Examples:
- Before: "Read our blog post about cold brew."
- After: "Our 10-minute guide to the best cold brew coffee beans of 2026, with brewing tips, ratios, and our top 3 picks. Updated monthly."
Deploy and wait 2-4 weeks. Then reload the CTR benchmark report and see whether the page moved above the curve.

What Google shows vs. what you write

Google rewrites roughly 40% of meta descriptions - picking body text instead. That's fine. Your job is still to give Google a good option. When Google does use your description, a well-written one consistently lifts CTR 0.5-2 percentage points. Over thousands of impressions, that adds up.

If Google keeps picking body text and ignoring your meta, fix the body text. Add a strong summary sentence near the top that could stand on its own as a snippet.

Meta description patterns that work

Lead with a specific benefit. "Cut onboarding time by 50%" beats "a helpful guide to onboarding."
Include a number. "3 steps," "7 templates," "updated 2026." Numbers earn attention.
Use the target query naturally. Google often bolds matching terms in the SERP, which visually pulls the eye.
End with implicit CTA. "Get started," "free template inside," "no signup." Avoid "click here."
Differentiate from competitors. Copy the top-3 ranking pages for your query, then write something that promises different value.

A worked example

A page ranks position 4 for "prometheus vs datadog" with 6,200 impressions/90d and CTR 2.1%. Expected CTR at position 4 on this site is ~6.5%. The current meta is literally empty - Google is pulling the first sentence of the page, which reads "Monitoring tools are important for any production system." Useless.

Rewrite: "Prometheus vs Datadog: a direct comparison across pricing, cardinality, alerting, and SRE workflows. Updated 2026, with a decision flowchart." Four weeks later, CTR climbs to 4.8% on the same impressions - roughly 167 extra clicks per 90 days from one edit.

Example outcome

On a typical content site, 15-25% of top-10-ranking pages are 2+ percentage points below their expected CTR. Rewriting the meta descriptions on the three worst offenders typically recovers 10-30% of the gap within 4-6 weeks - all without producing new content or chasing new rankings.

Next playbook

When performance problems are actually crawl or index problems: diagnose technical indexing issues →

How to improve CTR in Google Search Console