Search measurement guide
Diagnose Low Click-Through Rates in Search Console
Fix the query, page, country, device, and period first, check position and intent, then inspect the actual result before editing a title or description to raise clicks.
On this page
A low click-through rate in Search Console is a symptom with several possible causes, and the description is only one of them. Diagnose it in order: fix the exact page, queries, country, device, and date range; check where the page appears and whether those queries are ones it should answer; look at the actual search result and what surrounds it; and only then decide whether a title, description, or content change is justified. Compare equal periods after the change.
There is no universal “good” click-through rate. The rate depends on the query, the position, the device, and what else appears on the results page, so compare a page against its own history under the same filters rather than against a benchmark.
Define the exact query and page cohort
Start from a page, not the sitewide average. In the Performance report:
- Choose the search type (Web) and a date range with enough data, such as the last three months.
- Open the Pages tab and click the page. The report now filters to it.
- Open the Queries tab to see which searches produced its impressions.
- Add Country and Device filters if the business is local or the page serves one device type differently.
- Write down every setting. That is your cohort.
Two counting rules affect what you see. Click-through rate is clicks divided by impressions, and Search Console counts impressions and positions differently when grouped by property or by page: when several pages from your site appear for one search, the property view reports one impression at the topmost position, while the page view counts each page separately. Google’s help on how clicks and impressions are counted, checked September 27, 2026. Grouping by Pages counts table rows by page while the chart stays counted by property, and Google notes that adding a page filter makes the chart and table aggregate differently, which can raise the chart totals. Google’s help on troubleshooting data discrepancies, checked September 27, 2026. So the same traffic can show a different rate depending on the view.
Filtering by query or URL can also change totals, because of data truncation and the omission of anonymized queries. Google’s help on advanced filtering and comparison, checked September 27, 2026. A page with many rare queries may show fewer impressions in the query table than its page total.
Check position and intent before blaming the snippet
Sort the page’s queries by impressions and look at average position beside each one. Then ask two questions.
Is the page low on the results page for most of its impressions? If so, the rate may say more about where the page appears than about what its result says. An impression can count without the result being scrolled into view for standard results, while carousel and expandable items typically must be scrolled or expanded to count. Google’s help on impressions, position, and clicks, checked September 27, 2026. A low rate at an average position of 30 is usually a visibility question, not a wording one.
Do the queries match what the page offers? A page about raised garden beds showing for “raised bed plans free download” may be seen by people who want something else. Rewriting the description will not make it the right answer. The fix might be nothing, or a different page.
Separate branded from non-branded queries where the filter is available. Google notes that branded traffic often comes from users already familiar with your site, while non-branded traffic can indicate new audience growth. Google’s Performance report common tasks, checked September 27, 2026. A shift in the mix between the two groups can change the page’s overall rate without any change in the result.
Inspect the search context and snippet
For the three to five queries with the most impressions, search them yourself on the relevant device and in the relevant location, and save a dated screenshot of each. Record:
- The title link Google shows for your page, and whether it matches your SEO title.
- The snippet text, and whether it answers the query.
- What sits above your result: ads, a map, product listings, videos, an answer box, other features.
- What competing results promise that yours does not, such as a price, a location, or a date.
Google uses several sources to determine a title link automatically, and it primarily uses page content to generate the snippet, using the meta description when it describes the page better. Google’s title link documentation and snippet documentation, both checked September 27, 2026. So the text you wrote may not be the text searchers saw.
Your own search is one observation on one day. It does not show how often a feature appeared during the reporting period. Treat it as context for the numbers, not proof.
Choose a justified edit
Match the action to what you found:
| What you found | Likely action |
|---|---|
| Most impressions at low average positions | No snippet edit; the question is visibility, not wording |
| Queries the page should not answer | No edit, or a separate page for that need |
| Your SEO title shown, but it is vague or out of date | Rewrite the title to state the offer accurately |
| Google shows different title text | Check that the title, main heading, and page content agree |
| Snippet pulled from page text that does not answer the query | Make the page copy answer it; a description alone may not be used |
| Competitors show a fact yours omits (price, area, availability) | Add the fact to the page and, if accurate, the description |
| A search feature above all results for these queries | Record it; the rate may be limited by layout you do not control |
If the page runs WP Visibility, the title and description are set in the Search title and Search description fields of the post’s WP Visibility settings, and templates control pages without custom values; see setting title and description templates. The plugin controls what your page supplies. Google controls what the result displays.
Change one thing at a time where you can. Save the old wording and the date.
Set a comparison period
After the change, wait for enough data, then compare an equal period after with the period before, using the same cohort filters. If the page is seasonal, also compare the same period last year.
Google notes it can be difficult to show that a specific change caused an improvement, because events such as news, periodic changes in user sentiment, or improvements to a competing site also affect performance, and suggests checking whether the timing maps closely to your change compared with similar past periods. Google’s Performance report common tasks, checked September 27, 2026.
Write the result in three parts: what changed in the numbers, what else changed in the same period, and how confident you are. For a small page, the honest conclusion may be that the data is too thin to tell.
To fit this investigation into regular work, add it as an item in the monthly Search Console work queue.
Diagnosis checklist
- Cohort recorded: page, queries, search type, country, device, dates.
- Position checked for the queries driving impressions.
- Query intent compared with the page’s purpose.
- Dated screenshots of the live results for the top queries.
- Action matched to the finding, one change at a time.
- Old wording and change date saved; comparison period set.