How to Improve CTR in Google Search Console

Position 1 in Google earns a 39.8% click-through rate. Position 6 earns 4.4%. Those are the latest figures from First Page Sage, updated December 2025, and they explain why a click-through audit is the cheapest traffic you will ever buy. You already rank. The impressions already happen. Every click you are not getting is sitting in a report, waiting, and the fix is usually a sentence of copy, not a month of link building.

Most guides on this keyword hand you the same seven tips: better titles, schema, long-tail keywords, A/B test. All true, all a decade old, and all skip the part that decides whether the effort pays off: which pages to touch, and which to leave alone. Rewrite the title on a page nobody sees and you have done nothing. This pillar covers the whole loop, from finding the pages worth fixing to proving the fix worked, with the 2026 wrinkle the older posts pretend does not exist.

What a good CTR actually is

Click-through rate (CTR) is the share of people who saw your result and clicked it: clicks divided by impressions, reported per query and per page. A "good" number on its own means nothing, because CTR is chained to position. A 4% CTR is a disaster at position 1 and a small win at position 8. Before you judge any page, anchor it to what its rank should return.

Read the curve, not a single benchmark. The drop from position 1 to 2 is bigger than the entire distance from 4 to 10. That shape is the reason ranking one spot higher beats almost any copy tweak, and also the reason a page at position 6 with a 1% CTR is worth your afternoon: it is earning a quarter of what it should. Google's own retention window keeps 16 months of this data, so you are never short of pages to check, only short of a way to rank them.

Fix the pages that are leaking, not the ones that look bad

Here is where the old advice falls down. It tells you to sort by CTR ascending and fix the worst numbers. Do that and you will spend a Saturday rewriting titles for queries that get 12 impressions a month, because low-volume junk always has the lowest CTR. The pages that matter hide in the middle, and two signals find them.

A click leak is a query where your page ranks well enough to earn clicks but pulls a click-through rate below what its position normally returns. Position 5, 3,000 impressions, 1.4% CTR: the curve says 5.1%, so roughly 110 clicks a month are walking past you. That is a leak, and it is almost always a title-and-intent problem, because the rank proves Google already likes the page.

A striking-distance keyword is a query sitting between positions 5 and 20: not on page one yet, close enough that one round of on-page work can push it there. These are a different job. The leak is a copy fix; the striking-distance query is a relevance fix, usually adding the exact phrase to the title and a heading, tightening the section that answers it, and pointing an internal link at it. Both start from the same report, and both are arithmetic, not taste. Whether a query qualifies is something a formula decides, which is exactly why a checked tool beats eyeballing a spreadsheet.

To build the two lists by hand, open the Performance report, add the Average Position column, filter to queries with more than 500 impressions, then read position against CTR row by row. Anything under the curve for its rank goes on the leak list. Anything between 5 and 20 goes on the striking-distance list. If reading a thousand rows against a CTR curve sounds like the wrong use of a Saturday, that judgment is what the Search Console analysis in Tabkeel does for you: connect the property read-only and it returns the leaks and the striking-distance queries already ranked, with the position and the gap attached to each.

How to improve CTR in Google Search Console, step by step

Once you have the two lists, the work is the same on every page. Six steps, worst leak first.

1. Open the query in the Performance report. Filter to the one query, switch to the Pages tab, and confirm which URL ranks for it. You are fixing the page Google actually serves, which is not always the page you assumed.
2. Read the live SERP for that query. Google it. Look at the titles ranking around you and the intent they signal. If the top results are all how-to guides and yours is a product page, the problem is intent, not wording, and no title tweak saves it.
3. Rewrite the title tag to earn the click. Put the exact query near the front, stay under about 60 characters so it does not truncate, and give the searcher a concrete reason yours is the answer. No other edit on the page moves the number as much.
4. Rewrite the meta description as the pitch, not a summary. Google rewrites it often, but a sharp one wins the display more than half the time and sells the click when it shows. Say what the reader gets, in their words.
5. Add structured data where the SERP rewards it. FAQ, review, and product schema can attach rich results that pull the eye. Match the schema to the query type; do not bolt FAQ markup onto a page with no questions.
6. Mark it as changed and wait. Note the date and the old numbers. CTR moves on a rolling 28-day view, so a fix needs three to four weeks before the verdict is real.

Rewriting the title so the click actually happens

The title tag does most of the work, so it deserves more than "put the keyword in it." A few patterns move the number, backed by the way people scan a results page.

• Lead with the exact query. A searcher scanning ten blue links pattern-matches on the words they just typed. If your title buries the phrase behind your brand, it reads as a near-miss and loses the click to a result that mirrors the query.
• Add a number or a year when it fits. "7 fixes" and "(2026)" read as specific and current. They set an expectation of a concrete, dated answer, which is what a how-to searcher is hunting for.
• Match the format the SERP is already rewarding. If page one is listicles, a title that promises a list will out-click one that promises an essay, even at a lower position.
• Cut the power-word padding. "Ultimate," "best," and "complete" are so overused they now read as filler and can pull CTR down. Specific beats grand.

Here is one from a page I ran through the exam. A pricing page ranked position 7 for "team plan pricing" with 2,400 impressions and a 0.8% CTR, against the 3% the curve expects. The title was the bare brand name and a tagline. Nothing about pricing, nothing a person comparing plans would recognize as their answer. Tabkeel flagged the gap and returned the correction the way it always does, as a prompt you paste into your own coding agent rather than code it drops in your repo:

The page at /pricing ranks on page one for "team plan pricing" but its title tag is our brand name and a tagline, so it earns almost no clicks. Rewrite the title tag to lead with the searcher's intent, for example "Team Plan Pricing and Seats" followed by the brand, under 60 characters. Rewrite the meta description to state what a team gets and the starting price. Do not change the page body or the URL.

The edit took two minutes. Four weeks later the same query sat at position 6 with a 3.4% CTR. The rank barely moved. The clicks quadrupled, because the result finally looked like the thing the searcher asked for. That is the whole game with a leak: the page was already good enough to rank, it just was not saying so in the one line people read.

The 2026 complication nobody wants to mention

You can do all of this right and still watch CTR fall, because the results page changed under you. AI Overviews now appear on roughly 31% of searches, and they push the organic links down the page. The honest part is that the studies disagree on how much this hurts, and pretending otherwise is how you end up chasing a number you cannot move.

First Page Sage reports that pages sitting under an AI Overview keep most of their organic CTR. Ahrefs and Seer Interactive measured a 58% to 61% drop in organic CTR for the top result when an AI Overview is present. Both can be true, because it depends on the query. Informational questions ("what is a striking distance keyword") get an Overview that answers them, and the click evaporates. Commercial and navigational queries ("team plan pricing," "notion vs coda") get a thinner Overview or none, and the click survives.

So before you brand a page a failure, check the SERP. If an AI Overview is fully answering the query above your result, the CTR you lost is structural, and the move is to compete for the citation inside the Overview, not to rewrite a title that was never the problem. Getting quoted there is a different discipline, covered in why AI does not know your startup. If there is no Overview and the CTR is still under the curve, it is a real leak and the steps above fix it. Featured snippets play the same trick: they can hold a 42.9% CTR for whoever wins them and starve everyone below, so winning the snippet on your own query is offense, and losing it to a competitor is the reason your clicks dropped without your rank moving.

Five ways a CTR audit gets wasted

The mistakes are consistent, and they all come from optimizing the number instead of the traffic behind it.

• Chasing low-impression queries. A 0% CTR on a query with 15 impressions is noise. Filter to real volume before you sort, or you will fix pages that were never earning anything.
• Stuffing the title with keywords. A title written for the crawler reads as spam to the human doing the clicking. It can lift a ranking a hair and sink the CTR, which is the opposite of the trade you want.
• Ignoring intent mismatch. If your page cannot satisfy the query, a better title just buys a click and a fast bounce. Google notices the bounce. Fix the match or drop the query.
• Resetting and not measuring. Changing ten titles in a week with no record leaves you unable to say which one worked. Change, date, wait, read.
• Reading a single day. CTR is volatile day to day. Compare a 28-day window before against a 28-day window after, or you are reacting to weather, not climate.

Close the loop: prove the rewrite worked

A CTR fix you did not measure is a guess you got attached to. After each edit, hold the before window and compare it to the same length of time after the change went live. If clicks rose and the rank held, you converted impressions you were already paying for into traffic. If nothing moved, the query is either capped by an AI Overview or was never a real leak, and you stop spending on it.

This is the loop the Search Console connection is built to run: it flags a page as in-test when you ship a change, then reads the 28-day verdict so you are not squinting at a chart. The reason to compute leaks and striking-distance queries in code rather than by hand is the same reason you use a scale instead of guessing a weight. The number is either past the curve or it is not, and a formula never talks itself into a result that feels about right. Point the Search Console analysis at your property, read the opportunities for free, and the exact title and meta rewrite ships with the Founder plan when you want it written for you.

One thing to clear first, especially on a freshly launched site: CTR data only exists if Google can read the page. If your site renders only in the browser, most crawlers get a thin body and the Performance report has little to show. That is one front of the pre-launch checklist for AI-built sites, and it is worth clearing before you lean on ranking data. The full method behind every check is written up in the methodology, and you can run the whole thing against a URL at the Tabkeel exam.

Frequently asked questions