How to find low-CTR keywords in Search Console

Why a low CTR report is useful

A URL can appear thousands of times in Google results and still attract few visitors. That may indicate an opportunity to improve the title, search intent match or position—but it is not necessarily a broken page.

CTR means clicks divided by impressions. If a page received 200 clicks from 10,000 impressions, its CTR is 2%. The hard part is deciding whether 2% is unusually low for that page’s queries and rankings.

Before making a change, separate three questions:

• Does the page rank well enough to earn clicks?
• Does its snippet communicate the answer searchers want?
• Is Google showing it for queries that belong to a different page?

Find candidates in Google Search Console

1. Open Performance → Search results.
2. Choose a complete 28-day or 90-day date range.
3. Enable Total clicks, Total impressions, Average CTR and Average position.
4. Open the Pages tab and look for URLs with substantial impressions but relatively few clicks.
5. Select a candidate URL, then switch to Queries to see which searches generated its impressions.
6. Look at Devices and Countries if one segment behaves differently.

You can export the table, or use the Search Analytics API to repeat the analysis at scale.

Prioritize the right opportunities

Compare pages within broadly similar ranking and intent groups. For example, a transactional page at position three and an informational guide at position ten should not be measured against the same universal CTR target.

An illustrative shortlist:

These are fictional numbers to demonstrate prioritization.

The first query might be an interesting candidate for snippet analysis. The second may need better ranking and content coverage before a title edit can move the needle. The branded query is not a comparable benchmark.

Diagnose the snippet and search intent

Before rewriting the title, read the actual page and evaluate the results page for its important queries. Ask:

• Does the page answer the dominant query in its first few paragraphs?
• Does the title promise something the content actually delivers?
• Does the title distinguish the page from competing results?
• Did average position move materially between periods?
• Are impressions concentrated on mobile, where titles may truncate differently?

A good title is specific, useful and truthful. Do not automatically stuff it with years, numbers or ‘ultimate guide’ claims unless they match the content. Google may also rewrite the displayed title.

A simple title-testing workflow

Pick one high-value page, record 28 days of baseline data, and change a single element—often the title or intro. Annotate the change date.

After publishing, wait for Google to recrawl and for enough search data to accumulate. Compare equivalent windows, watching clicks, impressions, position and query distribution.

If CTR improves but rankings and impressions changed heavily, the title edit is not a clean explanation by itself. Treat the result as evidence to investigate rather than a laboratory-grade A/B test.

Automate the shortlist with an AI assistant

These prompts need an authorized Search Console connection to answer with real numbers. A language model without data can explain the method but cannot identify the true top opportunities.

A better way to rank CTR opportunities

Start with potential impact, not an arbitrary “CTR under 3%” rule. A query with 500 impressions and 1% CTR might represent fewer realistic additional clicks than a query with 30,000 impressions and a slightly better CTR.

An initial worksheet can include:

You can estimate scenario clicks, but label them explicitly as scenarios, not forecasts.

For example, if an article receives 10,000 impressions and 200 clicks, CTR is 2%. At a hypothetical 3% CTR with the same impression volume, it would receive 300 clicks, or 100 additional clicks. That is only arithmetic: actual impressions, rankings and query mix can change, and the proposed title may have no impact.

Walk through a real title-and-content review

Suppose a page titled “API Guide” is receiving impressions for “how to connect search console to claude.” If the page actually explains API authentication but not Claude integration, changing the title to promise a Claude tutorial would be misleading.

A better path:

1. Identify the query and read the existing landing page.
2. Decide whether the page should serve that intent at all.
3. If not, create or strengthen a dedicated, useful tutorial and link to it contextually.
4. If the page does satisfy the intent, make the title and introductory answer more specific.
5. Check that the page includes steps, prerequisites and a troubleshooting section.
6. Measure performance across comparable complete windows after the change.

This is the difference between improving click appeal and writing an inaccurate title purely to attract impressions.

How to report results without fooling yourself

Keep a change log with URL, primary query cluster, original title, revised title, publish date, baseline period and follow-up period.

When reviewing performance, ask whether:

• Impressions grew or fell because of demand or rankings.
• Average position changed materially.
• The distribution of queries shifted.
• The same countries and devices are being compared.
• More clicks translated into meaningful actions on the site, if GA4 is connected.

Do not present a simple before-and-after comparison as a controlled SEO experiment. The purpose is to make a better decision with imperfect data.

Frequently asked questions

What is a good CTR in Google Search Console?

There is no universal number. CTR depends on position, query intent, brand recognition, country, device and how Google displays results. Compare similar queries and focus on pages with enough impressions to be informative.

Should I optimize pages ranking outside the top ten?

They may need content quality, search-intent alignment or link improvements before snippet changes can have much effect. A low CTR at position 30 is not surprising.

Does Google use meta descriptions exactly as written?

Not always. Google can generate or select snippets based on the query and page content. Write accurate descriptions for users, not as a guarantee of what appears.

How much time should I wait after updating a title?

There is no fixed number of days. Confirm that Google has recrawled the page where possible, then collect enough complete performance data to compare meaningful periods.

Common mistakes to avoid

Optimizing by page-level CTR only. It combines many queries that may have different intents.

Ignoring brand queries. Branded searches often have distinct CTR behavior and should usually be analyzed separately.

Confusing position with visibility. Average position is an aggregate and may disguise a mix of very different rankings.

Changing dozens of pages at once. That makes it hard to determine what happened and creates editorial risk.

If more than one of your URLs appears for the same queries, see our guide to keyword cannibalization. If clicks dropped sitewide, use the traffic-drop checklist.

Want a conversational version of this report? Explore GSC Magic, built around search data access for AI assistants.