Ai Overviews and Organic Traffic: How to Read Stable Impressions Alongside Falling Clicks

AI Overviews are changing how people interact with search results, and one of the clearest warning signs is a familiar one: organic impressions remain stable while clicks decline. At first glance, this can look like a ranking problem. In reality, the cause may sit in the space between visibility and action, where an AI-generated answer satisfies part of the search before a user ever reaches a traditional result.

That does not mean every traffic loss is caused by AI Overviews. Keyword cannibalisation, ranking dilution, weaker snippets, changing search intent and declining demand can produce a similar pattern. You need to separate those causes before changing content, merging pages or redirecting URLs.

This guide explains how to read stable impressions alongside falling clicks, how to identify AI Overview exposure, and how to run a practical keyword cannibalisation audit. It also shows how a structured writing and publishing platform such as SEO Letters can support the research, content overlap analysis and internal linking work required to respond properly.

What stable impressions and falling clicks actually suggest

An impression means that your page was recorded as appearing in a search environment. A click means that the searcher selected your result and visited your site. Those two metrics are related, but they are not interchangeable.

When impressions stay fairly consistent and clicks fall, your pages may still be visible while becoming less attractive, less prominent or less necessary to the searcher. That distinction matters because the correct response depends on which of those explanations is strongest.

The main possibilities include:

AI Overviews are answering the query before the traditional listings receive attention
Your result is appearing lower on the page despite being counted as an impression
Searchers are choosing different result formats, such as videos, shopping results, forums or local packs
Your title and meta description are no longer competitive
Search intent has shifted away from what your page offers
Multiple pages from your site are competing for the same keyword
Another page has taken the stronger position while your wider site still collects impressions
Search demand has changed, even though Search Console continues to show impressions
Brand or navigational searches are being answered directly within the search interface

This is why a falling click-through rate should be treated as a diagnostic signal, not a complete diagnosis.

The basic CTR calculation

Your click-through rate is calculated as:

CTR = clicks ÷ impressions × 100

For example:

Period	Impressions	Clicks	CTR
Previous period	100,000	8,000	8.0%
Current period	102,000	6,200	6.1%

In this example, visibility has remained broadly stable, yet the site has lost 1,800 clicks. The 1,800-click gap is commercially meaningful, but it does not immediately prove that rankings have collapsed.

A page could be ranking in a similar position while receiving fewer clicks because an AI Overview, featured snippet, video carousel or expanded set of commercial results is taking more visual space. The whole thing is increasingly dependent on what appears above and around your organic listing.

Why AI Overviews can reduce clicks without reducing impressions

AI Overviews can change the search journey before a user reaches the blue links. A searcher may ask a broad informational question and receive a generated summary that combines information from several sources. They may then stop searching, refine the query, or click one of the cited sources rather than the standard organic result they would previously have selected.

This creates a measurement problem. Your page may still be present. It may still be cited. It may still be relevant to the query. Yet the traditional result can receive fewer clicks because the page is no longer the first useful destination.

The impact is likely to vary by query type:

Query type	Likely AI Overview effect	Organic traffic risk
Basic informational queries	Summary may satisfy the user	High
Definition searches	Direct answer may remove the need to click	High
Comparison searches	Overview may provide an initial shortlist	Medium to high
Complex how-to searches	User may still need detailed guidance	Medium
Product evaluation searches	Commercial pages may retain value	Medium
Transactional searches	Product and service pages can remain important	Lower, though not negligible
Branded searches	Users may still click official pages	Low to medium
Local searches	Local packs and map features often dominate	Medium

A short question such as “what is keyword cannibalisation?” can be answered quickly. A query such as “how do I fix keyword cannibalisation across 50 product pages?” still requires a detailed process, examples and implementation guidance.

That distinction gives you a useful starting point. If stable impressions and falling clicks are concentrated among broad, answerable queries, AI Overviews may be part of the explanation. If the losses are concentrated on commercial pages, you should investigate ranking position, competitors, SERP features, page quality and internal linking before blaming AI Overviews.

The difference between visibility and qualified visibility

A raw impression count can hide important changes in the quality of your visibility. You might receive the same number of impressions from:

Position 1 searches on high-intent terms
Position 5 searches on broader terms
Mobile impressions where the result is pushed below several features
Search queries that are only loosely related to your service
Branded searches with a lower propensity to click
Long-tail queries where AI-generated answers dominate the first screen

Those impressions are not equal. A practical analysis should look beyond totals and review the following dimensions:

Average position
Query type
Device
Country or market
Landing page
Search appearance
Brand versus non-brand
Informational versus commercial intent
Presence of AI Overview or other SERP features
Conversion rate after the click

A stable impression count with a stable average position but lower CTR points towards a SERP behaviour change, a weaker snippet or greater answer satisfaction. A stable impression count with a worse average position suggests ranking dilution or increased competition.

This is where a page-level view becomes essential. Site-wide reporting can make unrelated trends appear connected. One directory may be growing while another loses traffic, leaving the total impression count almost unchanged.

AI Overviews, Search Console and the limits of available data

Search Console is useful, but it does not provide a perfect AI Overview report. Depending on how Google records and exposes search features, AI-generated result areas may not appear as a clean, separate line in your standard performance data.

You should avoid assuming that every lost click came from an AI Overview simply because the timing appears to match. Search interfaces change frequently, and Google does not always expose every feature with the granularity SEOs would prefer.

Use a layered measurement process:

Layer one: Search Console trend analysis

Compare at least 28 days with the previous period, and where possible compare the same period year on year. Review:

Clicks
Impressions
CTR
Average position
Queries
Pages
Countries
Devices
Search appearance categories

Look for clusters rather than isolated examples. If hundreds of informational queries show similar CTR decline while average position remains stable, the pattern deserves investigation.

Layer two: Manual SERP sampling

Choose a representative set of affected queries and inspect them manually. Record:

Whether an AI Overview appears
Whether the overview appears above the organic results
Whether your page is cited
Whether your page ranks in the standard listings
The number and type of competing features
The wording of the current title and description
The apparent search intent

Do not rely on one search from one location. Results can differ by country, device, account history and query wording.

Layer three: Analytics and revenue data

A click decline matters most when it affects meaningful outcomes. Compare:

Organic sessions
Engaged sessions
Lead submissions
Ecommerce transactions
Assisted conversions
Revenue per landing page
Conversion rate by query group
New versus returning users

A page may lose low-value informational clicks while maintaining commercial conversions. That is a different situation from losing product-page visits and qualified leads.

Keyword cannibalisation can create the same pattern

Keyword cannibalisation occurs when several pages on your site target the same or very similar search intent, causing search engines to switch between them or struggle to identify the best page.

The result is not always a dramatic ranking collapse. Often, the site continues to receive impressions across multiple URLs, but no single page consistently captures the strongest position or the highest click share.

This can look like:

Stable impressions at the site level
Falling clicks for the target topic
Several URLs ranking for overlapping queries
A rotating landing page in Search Console
Lower average position for the preferred URL
Inconsistent snippets and titles
Internal links pointing to multiple competing pages
Backlinks divided across related URLs
Content that answers the same question with slightly different wording

That is ranking dilution SEO in practical terms. Your authority and relevance signals are spread across pages instead of concentrated behind the URL most capable of satisfying the searcher.

A simple cannibalisation example

Imagine an SEO software company has these pages:

URL	Main target	Secondary overlap
`/keyword-cannibalisation-guide/`	What is keyword cannibalisation?	How to fix keyword cannibalisation
`/fix-keyword-cannibalisation/`	How to fix keyword cannibalisation	Keyword cannibalisation audit
`/keyword-cannibalisation-audit/`	Keyword cannibalisation audit	SEO content overlap
`/seo-content-strategy/`	SEO content strategy	Content cannibalisation

Each page may be useful. The problem begins when all four pages use the same headings, target the same phrases and attract the same internal links.

Search Console could show stable total impressions for “keyword cannibalisation”, but clicks may fall because Google alternates between URLs or ranks the less suitable page. The user sees a result that does not quite match the expected intent, and your site misses the opportunity to present one authoritative resource.

AI Overviews can make this harder to spot. If the overview answers the broad query, your pages may receive fewer clicks across the cluster, while cannibalisation still weakens your position for the deeper queries where a click remains valuable.

How to run a keyword cannibalisation audit

A keyword cannibalisation audit should combine data extraction, search intent mapping and content overlap analysis. Do not start by deleting pages. That is often the quickest way to create a new problem.

Step 1: Export query and page data

Export at least three to six months of Search Console data if available. Include:

Query
Page
Clicks
Impressions
CTR
Average position
Date range

Then group the data by query. Identify terms where two or more URLs receive meaningful impressions or clicks.

You can flag possible conflicts with a simple rule:

Potential conflict = one query + two or more URLs + overlapping intent

The third part matters. Two pages can appear for the same broad phrase without being true competitors if one satisfies informational intent and the other satisfies a transactional intent.

Step 2: Identify URL rotation

Look at whether the URL receiving the most impressions or clicks changes over time. URL rotation can suggest that Google is uncertain about which page should rank, although it can also reflect normal volatility.

Pay particular attention when:

The preferred page disappears while a weaker page rises
Several pages have similar titles
The pages have overlapping introductions and headings
The internal linking structure does not identify a primary resource
The pages have different backlink profiles but similar topical targeting

Step 3: Complete search intent mapping

Map each affected query to a clear intent category:

Intent category	User expectation	Suitable page type
Informational	Explanation or definition	Guide
Investigational	Comparison, evidence or evaluation	Comparison or research page
Commercial	Product or service options	Landing page
Transactional	Purchase, signup or enquiry	Product or conversion page
Navigational	A known brand or destination	Brand page

If two pages target the same phrase but satisfy different intent, clarify their roles. If they satisfy the same intent, consolidation may be appropriate.

Step 4: Perform content overlap analysis

Compare the competing pages for:

Primary topic
Search intent
Heading structure
Entities and subtopics
Examples
Internal links
Calls to action
Metadata
Word count and depth
Original evidence
Conversion purpose

Content overlap analysis is not just a phrase-matching exercise. Two pages can use different words while delivering the same answer. Conversely, they can use similar words while serving different stages of the buying journey.

A useful scoring framework is:

Factor	Score 0	Score 1	Score 2
Same primary intent	No	Partly	Yes
Same target audience	No	Partly	Yes
Similar page purpose	No	Partly	Yes
Heading overlap	Low	Medium	High
Query overlap	Low	Medium	High
Internal link competition	Low	Medium	High

A total score of 8 or more suggests that the pages need a deliberate decision. The answer may be consolidation, a clearer distinction, stronger internal links or a change in targeting.

Step 5: Choose a resolution

Use one of four main actions:

Consolidate when one page can absorb the useful content from another
Differentiate when both pages are valuable but need separate intent
Canonicalise when duplicate or near-duplicate versions must exist for operational reasons
Redirect when a weaker page has no independent purpose, links or traffic value

There is also a fifth option that is often overlooked: leave both pages live but rebuild the information architecture around a parent topic and clearly defined supporting pages.

Do not merge pages simply because they share one keyword. Merge them when they compete for the same user need.

How to distinguish AI Overview effects from cannibalisation

The two issues can coexist. Your analysis should test for both rather than treating them as competing explanations.

Signal	More consistent with AI Overview impact	More consistent with cannibalisation
Impressions remain stable	Yes	Yes
CTR falls across many sites or competitors	Likely	Less likely
Average position remains stable	Likely	Possible
Multiple URLs rank for the same query	Possible	Strong signal
URL receiving clicks changes often	Less typical	Strong signal
Loss concentrated in informational queries	Strong signal	Possible
Loss concentrated in one topic cluster	Possible	Strong signal
Page titles look less compelling than competitors	Possible	Possible
Internal links point to several similar pages	No	Strong signal
Conversions remain stable	Possible	Possible
Ranking position declines for the preferred URL	Possible	Strong signal

The practical approach is to create two hypotheses:

SERP displacement hypothesis: AI Overviews or other features are reducing clicks despite similar rankings.
Site architecture hypothesis: Multiple pages are dividing relevance and weakening the preferred result.

Then test both against the same query groups. This keeps the investigation disciplined.

Build a query-level click-loss model

A useful way to avoid vague conclusions is to calculate expected clicks based on your historical CTR.

For each query group:

Expected current clicks = current impressions × historical CTR
Click gap = expected current clicks − actual current clicks

Suppose a group produced 40,000 impressions this month. Historically, its CTR was 6 percent, so expected clicks would be 2,400. Actual clicks are 1,680.

2,400 − 1,680 = 720 estimated lost clicks

This does not prove the cause, but it helps you prioritise. You can rank query groups by estimated click gap, commercial value and likelihood of AI Overview exposure.

A basic prioritisation score could include:

Dimension	Low priority	Medium priority	High priority
Estimated click gap	Under 100	100 to 500	Over 500
Conversion value	Minimal	Moderate	High
Cannibalisation evidence	None	Possible	Clear
AI Overview exposure	Rare	Mixed	Frequent
Ranking position	Improved	Stable	Declined

Focus first on groups with a high click gap and clear commercial value. An informational page losing 500 low-intent clicks may be less urgent than a service page losing 50 clicks that normally generate qualified enquiries.

Improve the internal linking strategy before rewriting everything

Internal links help search engines and users understand which pages matter. They cannot force a ranking, but a consistent internal linking strategy can reduce ambiguity across a content cluster.

Start by choosing a primary page for each topic. Then create a deliberate hierarchy:

Pillar page for the broad subject
Supporting pages for distinct subtopics
Commercial page for the relevant product or service
Conversion-focused links where the user is ready to act

For example, a pillar page about AI Overviews and organic traffic might link to:

A guide to measuring AI search visibility
A keyword cannibalisation audit guide
A technical SEO measurement framework
A content-refresh service page
A product page for your SEO workflow platform

The links should use contextually accurate anchor text. Do not force the exact same anchor phrase into every paragraph.

Internal linking checks for competing pages

Review:

Which URL receives the most internal links
Whether anchor text identifies the preferred page
Whether related articles link to competing pages equally
Whether orphaned pages still rank for important queries
Whether commercial pages are buried beneath purely informational content
Whether links point to outdated or redirected URLs
Whether navigation and breadcrumbs reinforce the topic hierarchy

A site can publish excellent content and still underperform because its architecture sends mixed signals. This is one reason automated content production without workflow control often creates more pages than a site can properly organise.

Rewrite for answer depth, not just answer length

When AI Overviews answer simple questions, adding another 500 words to a basic definition may not restore traffic. Your page needs to offer value that a short summary cannot easily replace.

That value might include:

Original data
Clear implementation steps
Screenshots or worked examples
Expert interpretation
Decision frameworks
Templates
Industry-specific application
Risks and limitations
Product comparisons
Calculation methods
Evidence from testing
Updated recommendations

For this topic, a useful article should not stop at saying that AI Overviews may reduce clicks. It should show how to segment Search Console data, how to identify competing URLs, how to model click loss and how to decide whether a content change is justified.

This is where SEO Letters can support the production workflow. It can help move from keyword research and topical clustering to structured article creation, internal links, schema and publishing, while leaving the strategic decisions with you.

Use content refresh campaigns to protect existing traffic

A content refresh should not mean changing the publication date and replacing a few sentences. It should respond to evidence.

For a page affected by AI Overview-related click decline, check:

Has the search intent changed?
Does the introduction answer the question clearly?
Does the page provide enough detail for a user who needs to act?
Are the examples current?
Are competitors offering better tools, tables or evidence?
Is the title still accurate and compelling?
Does the page link to the correct supporting and commercial resources?
Are there duplicate pages competing for the same topic?
Are authorship, review and source details clear?
Is the page updated according to real SERP changes rather than assumptions?

A refresh campaign is especially useful when a site has hundreds of established articles. Manual review of every page is slow and inconsistent, while publishing new articles may deepen the overlap problem.

An automated workflow can identify pages with declining CTR, falling conversions or outdated content, then route them into a review process. The final decision should still involve editorial judgement, particularly for regulated, financial, health or high-stakes subjects.

Create content that earns the click after an AI Overview

You cannot control whether Google displays an AI Overview. You can influence whether your result looks like the best next step.

Your title should align with the actual intent and promise a useful outcome. Avoid vague headlines that could be replaced by a generic summary.

Weak:

AI Overviews and Traffic

Stronger:

AI Overviews and Organic Traffic: How to Diagnose Falling Clicks Without Misreading Stable Impressions

The stronger title signals a problem, a method and a practical result. The content then needs to deliver that promise quickly.

Useful click incentives include:

A downloadable audit template
A calculation model
A detailed comparison
A visual workflow
A specific industry example
A current benchmark
A step-by-step implementation method
A clear explanation of when not to consolidate pages

Do not use misleading curiosity tactics. A click earned through a false promise usually produces poor engagement and weak commercial value.

A worked example: stable visibility, falling clicks and mixed causes

Assume a B2B software company has noticed the following change:

Metric	Quarter one	Quarter two
Organic impressions	1.20 million	1.23 million
Organic clicks	74,000	61,000
Average position	11.2	11.6
CTR	6.2%	5.0%
Organic leads	1,180	1,030

The initial assumption is that AI Overviews caused the entire decline. The data does not support that conclusion yet.

The SEO team segments the queries:

Informational queries: clicks down 25 percent, average position stable
Comparison queries: clicks down 14 percent, average position slightly worse
Product queries: clicks down 9 percent, two URLs competing
Branded queries: clicks stable
Mobile queries: clicks down 21 percent
Desktop queries: clicks down 8 percent

Manual SERP checks show AI Overviews on many informational queries. A keyword cannibalisation audit also finds three competing comparison pages and inconsistent internal links to the main product page.

The likely explanation is mixed:

AI Overviews are reducing clicks for broad informational searches
Mobile layouts are increasing the visual cost of lower positions
Cannibalisation is weakening comparison and product visibility
Some titles are no longer aligned with current search expectations

The response should also be mixed:

Improve informational pages with original examples and stronger depth
Consolidate two overlapping comparison pages
Retarget the third page around a distinct intent
Strengthen the internal linking strategy towards the commercial page
Test titles and descriptions on high-value pages
Track leads and revenue, not clicks alone

That is a more credible diagnosis than assigning every loss to AI Overviews.

How SEOLetters supports the measurement-to-publishing workflow

SEO Letters is designed for teams that publish at scale and need a repeatable system between keyword discovery and live content. The platform supports:

Keyword research with difficulty ratings
Topical authority cluster planning
Competitor and site-gap analysis
Structured article generation
Internal link recommendations
Schema and image support
Multi-language content production
Publishing to WordPress, Shopify and webhooks
Performance monitoring
Autonomous campaign scheduling
Content-refresh campaigns
Product-aware articles for affiliate and ecommerce sites

That combination matters in an AI Overview environment because traffic measurement and content production cannot operate as separate activities. If the data shows that a cluster is losing clicks, you need a controlled way to decide whether to create, consolidate, differentiate or refresh content.

You can set a topic, cadence and destination in the platform’s campaign workflow, then review the output and publishing logic before it reaches your site. For organisations managing multiple markets, multi-language generation across 21 languages can also help keep the intent structure consistent, although every language version still requires local review.

The SEO Letters blog writing platform is most useful when it forms part of a wider operating model:

Research the search landscape
Map topics and intent
Identify content gaps and overlap
Assign one purpose to each page
Generate and refine the article
Add internal links and structured data
Publish to the correct destination
Monitor impressions, clicks and conversions
Refresh or consolidate based on evidence

The software handles much of the repetitive production work. Your team retains control over positioning, review standards and business priorities.

Metrics to monitor after making changes

Do not judge a change on clicks alone, especially when the SERP itself is changing. Establish a baseline before editing the page and review results over an appropriate period.

Track:

Query-level impressions
Query-level clicks
CTR
Average position
URL ownership for target queries
Organic sessions
Engagement quality
Leads or transactions
Revenue
Assisted conversions
AI Overview presence from manual sampling
Number of competing URLs
Internal links to the preferred page

Use a decision window that reflects the site’s traffic volume. A high-volume site may identify early movement within a few weeks, while a smaller site may need several months to produce a reliable signal.

A practical reporting framework

Question	Metric	Interpretation
Are we still visible?	Impressions	Demand and exposure
Are rankings improving?	Average position	Relative search performance
Are users choosing us?	CTR	SERP attractiveness
Are the right pages ranking?	URL ownership	Cannibalisation control
Are visits valuable?	Leads, revenue, transactions	Business impact
Is the page useful?	Engagement and assisted conversions	Content quality signal
Is the SERP changing?	Manual feature checks	Context for CTR movement

A page with fewer clicks but a higher conversion rate may be improving in commercial terms. A page with stable clicks but declining assisted conversions may be attracting less qualified users.

Common mistakes when responding to falling organic clicks

Blaming AI Overviews without checking rankings

AI Overviews are visible and easy to discuss, so they can become a convenient explanation. Yet competitors, technical issues, algorithm updates and page overlap may be contributing more to the loss.

Merging every page that shares a keyword

Broad keyword overlap is not enough. A guide, a comparison page and a product page can all mention the same phrase while serving different intent.

Publishing more content immediately

If the issue is ranking dilution SEO, adding five similar articles may spread relevance even further. Run the keyword cannibalisation audit first.

Optimising only for impressions

Impressions are useful for measuring visibility, but they do not show whether the audience is relevant or whether the page creates revenue. Include conversion and query intent in the review.

Rewriting titles without reviewing the page

A more compelling title may improve CTR temporarily, but if the page does not satisfy the current intent, the gain will not last. Snippet optimisation should support a genuine content improvement.

Removing pages without checking links and conversions

A low-click page may still earn backlinks, support another page through internal links or assist conversions. Review the wider site role before redirecting it.

A repeatable 30-day action plan

If you are seeing stable impressions alongside falling clicks, use this sequence.

Days 1 to 5: Establish the baseline

Export Search Console data
Compare current and previous periods
Add a year-on-year comparison
Segment brand and non-brand queries
Segment mobile and desktop data
Identify the largest click gaps
Note pages with high impressions but low CTR

Days 6 to 10: Inspect the SERPs

Sample affected queries manually
Record AI Overview appearances
Check featured snippets and other features
Review competitor titles and page formats
Confirm whether your preferred URL still ranks
Note changes in search intent

Days 11 to 15: Run the cannibalisation audit

Group queries by URL
Identify URL rotation
Compare titles and headings
Complete content overlap analysis
Map each page to a primary intent
Select pages for consolidation or differentiation

Days 16 to 22: Implement priority changes

Improve titles and descriptions
Rewrite weak introductions
Add original evidence and examples
Consolidate clear duplicates
Update canonicals and redirects where appropriate
Strengthen the internal linking strategy
Add relevant links to commercial pages

Days 23 to 30: Launch measurement and refresh controls

Record the new baseline
Add affected URLs to a performance dashboard
Schedule a follow-up review
Create refresh tasks for pages with ongoing decline
Monitor rankings, URL ownership and conversions
Avoid making several unrelated changes without documentation

This sequence gives you a record of what changed and why. That is important when organic performance moves slowly or several SERP changes happen at once.

Key takeaway: falling clicks require a layered diagnosis

Stable impressions alongside falling clicks usually mean that visibility has not disappeared, but the value or prominence of that visibility has changed. AI Overviews may be reducing the need to click, particularly for straightforward informational searches, but they are only one possible cause.

Your investigation should include:

AI Overview and SERP feature checks
Query-level CTR analysis
Average position trends
Search intent mapping
Keyword cannibalisation audit
Content overlap analysis
URL ownership and ranking rotation
Internal linking strategy
Conversion and revenue measurement

The most reliable response is not to publish more pages automatically. It is to create a clearer information architecture, strengthen the pages that deserve to rank, and make each URL responsible for a distinct search need.

If you’re managing a large content programme, use SEO Letters to plan, write, refresh and publish SEO content through one controlled workflow. Its research, topical clustering, competitor-gap analysis, internal linking support, scheduled campaigns and performance features can help turn scattered content activity into a measurable publishing operation.

Start with the data. Then decide whether the next action is a refresh, a rewrite, a consolidation, a stronger internal link or a new page. That discipline is likely to matter even more as AI Overviews continue to reshape the relationship between organic impressions and actual website traffic.