Skip to the indexed answers
Digital Marketing Index

usdigitalmarketingindex.com SEO

How do you investigate falling organic CTR when search impressions are rising?

By finding out where the new impressions came from before concluding that anything got worse. Click-through rate is clicks divided by impressions. When impressions rise faster than clicks, the rate falls, and that can happen while every existing page performs exactly as before. The site may simply have started appearing for more queries, in lower positions, where few people click. That is growth, reported as a decline. The opposite reading is also possible: the same queries, the same positions, and fewer clicks, because something above the result now takes them. The investigation separates these by breaking the totals apart, by query, by position, by device, and by page, and asking in each segment whether CTR really fell or whether the mix of segments changed.

Changes in query mix

Start here, because it explains the pattern without anything having deteriorated.

In Search Console’s performance report, compare the two periods with the query dimension.

Segment Question
Queries present in both periods Did CTR fall for the same queries?
Queries new in the later period How many impressions did they add, and at what position and CTR?
Queries that disappeared Did high-CTR queries drop out?
Branded and non-branded Did the share of branded impressions, which carry high CTR, shrink?
Informational and transactional Did the balance change?

A worked illustration with invented numbers:

Segment Period 1 impressions Period 1 clicks Period 1 CTR Period 2 impressions Period 2 clicks Period 2 CTR
Existing queries 100,000 5,000 5.0% 100,000 5,000 5.0%
New queries 0 0 60,000 600 1.0%
Total 100,000 5,000 5.0% 160,000 5,600 3.5%

Total CTR fell from 5.0% to 3.5%. Clicks rose by 600. Nothing declined: CTR on existing queries is unchanged, and the new queries added impressions at a low rate. The headline number moved because the denominator grew with low-CTR impressions.

A limit to keep in mind: Search Console omits anonymized queries from the query table, so query-level segments will not sum to the totals (Search Console help on data discrepancies). Use the page dimension as a cross-check, since the omission described there applies to the query table.

The same mix effect distorts average position. The average can improve while the queries that matter slip, as shown in the guide to why average search position can improve while rankings for business-critical queries fall.

SERP feature effects

If CTR fell on the same queries at the same positions, look at what the results page contains.

Elements that can sit above or beside an organic result include advertisements, answer boxes, AI-generated summaries, local packs, shopping units, image and video carousels, and question panels. When one appears or expands for a query, the organic result is lower on the screen at an unchanged position number, or the question is answered before a click.

How to examine it:

  1. Take the queries with the largest CTR drop and steady position.
  2. Search each, on the device where the drop occurred, from the relevant country.
  3. Record what appears above the site’s result, with the date and a screenshot.
  4. Compare with any earlier record the team holds.
  5. In Search Console, check the search appearance dimension for changes in how the site’s own results are displayed, such as a rich result gained or lost.

Without an earlier record, a current observation shows what the page looks like now, not that it changed. Say so in the findings. Keeping dated observations for priority queries makes the next investigation possible.

The site’s own listing matters too. A title link or snippet that Google now generates differently, or a rich result that stopped appearing, changes how the result looks at the same position.

Position and device shifts

Position can change the rate without any change in appeal.

Check Method
Average position by query Compare periods for the matched queries
Distribution, not just average Count queries in positions 1 to 3, 4 to 10, and beyond, in each period
Device mix Compare each device's share of impressions; CTR at a position can differ by device
Country mix New impressions from countries the site does not serve carry low CTR
Search type Check that web, image, and video are not mixed
Page Has a different URL started ranking for the query?

A page moving from position 11 to position 9 can gain impressions, because it is now shown to searchers who do not look past the first set of results, and its CTR at 9 is low. Impressions up and CTR down can be the signature of ranking improvement on the edge of page one.

Search Console’s help explains what position is: the topmost position of a link to the site, averaged across searches. An average moves when the mix of searches moves.

Page-level next tests

An investigation worksheet:

# Hypothesis Evidence to gather Finding Confidence
1 New low-CTR queries diluted the total Matched-query CTR, both periods; impressions and CTR of new queries High, medium, low
2 Branded share of impressions fell Branded and non-branded split, both periods
3 Position slipped on existing queries Per-query position for the matched set
4 Pages entered page one at low positions Position distribution
5 Device mix shifted Impressions share by device; CTR by device
6 Country mix shifted Impressions by country
7 Results page features changed for key queries Dated observations of the results pages
8 The site's listing changed: title, snippet, rich result Search appearance data; current listing against a previous record
9 A different URL now ranks Page dimension for affected queries
10 Search type or data change Filters checked; any reporting change in the period

Fill in the finding column with numbers, and the confidence column honestly. Several hypotheses can be true together. Hypotheses 1, 2, 4, 5, and 6 are mix effects, in which nothing got worse for any given search. Hypotheses 3, 7, 8, and 9 are real changes in how the site appears.

Confidence guide:

Level Basis
High Shown directly in the data, such as matched-query CTR unchanged and new queries accounting for the added impressions
Medium Consistent with the data, with another explanation not excluded
Low Plausible; no before-and-after evidence, as with results page features lacking an earlier record

What follows from the findings:

Finding Next step
Dilution by new queries No fix needed for CTR; assess whether the new queries are worth pursuing, and report clicks alongside CTR
Lower positions on existing queries Investigate rankings for those queries
New first-page entries at low positions Continue the work that moved them
Results features taking clicks Decide whether the query remains worth targeting; consider what the listing can offer that the feature does not
Listing changed Review titles and structured data on the affected pages
Title or snippet weak for the query Test a revised title on a few pages, with comparable pages unchanged as a reference, and compare after several weeks

A title test changes one thing on a small set of pages. Read its result in clicks and CTR for the queries those pages already ranked for, at comparable positions.

Reporting advice: do not present total CTR alone. Show clicks, impressions, and CTR for a fixed set of queries next to the totals. The fixed set shows whether performance changed. The totals show how much the footprint grew.

One feature deserves separate treatment because of its effect on informational queries, and it is examined in the guide to how AI Overviews affect organic click opportunities for informational searches.

The order of work: split by query first, check whether matched queries really lost CTR, then examine position, device, and the results page, and state for each explanation how sure the evidence makes you.

Link trail

Reader comments on this answer

Add your comment to this answer

Your email address will not be published. Required fields are marked *