Merchant listings: 3–7x more desktop impressions, hardly a click

Anyone who filtered Search Console for “Merchant listings” in summer 2026 saw a curve that looked too good: desktop impressions at a multiple of the spring level, plus a CTR and an average position that no longer had anything to do with the shop. I have access to the Search Console of several online shops that have nothing to do with each other: different companies, different product ranges, from large to small. The large ones show the pattern in the same week, the smallest got it later, and all of them lost it on the same day. That is why I do not think any of the shops triggered it. I have not found it in any documentation.

Method and limitations

• Search Console: search appearance “Merchant listings”, by device, April 15 to October 17/18, 2026, pulled per shop via the API. Plus the breakdown by country for September 1 to 14, all devices. The last few days are still preliminary.
• Baseline: April 15 to May 14, 2026, index 100. The window is May 25 to October 4, “after” is October 9 to 18. In the charts, daily values are smoothed as a 7-day average so the weekly rhythm does not interfere. The averaging smears the stop on October 9 over a week; in the daily values it is a single step.
• Merchant Center: impressions and clicks of free product listings per day and country for the largest and the smallest shop, as a cross-check. The two tools count different things, more on that below.
• Demand: Keyword Planner volumes for the matching product categories for Germany, Austria, Switzerland and the UK, monthly values. For the countries most of the impressions came from, I had no volumes.
• Google documentation: the Search Console data anomalies page, checked on October 20, 2026.
• Anonymized: no names, no industries, no number of shops, no absolute numbers. Shop A is the largest, B the second large one, C a small shop. One more small property appears at the end of the next section.
• Limitations: a small sample of shops, all European. Country data only for two weeks within the window. I do not know what Google counted. This is an observation with an interpretation, not proof.

What happened

Desktop impressions for merchant listings rose over about five days, starting on May 16, 17 or 18 depending on the shop, and reached a plateau on May 20. Shop A, index per day from May 15: 105, 210, 322, 503, 583, 713. Shop B: 102, 109, 196, 268, 328, 350. Shop C, small and noisier: 247, 96, 138, 434, 313, 464. They stayed there for almost five months: Shop A on average at 6.5x the April level, Shop B at 3.1x, Shop C at 5.8x. On mobile, nothing rose in any shop. In A and C, mobile impressions even fell, to about half and two thirds, and did not recover; in B they stayed flat.

That only desktop was affected is shown by the direct comparison of devices in Shop A: mobile impressions for merchant listings even declined from May onward, while desktop impressions rose to six to eight times the April level.

It ended everywhere on the same day. Shop A went from index 739 on October 7 to 349 on the 8th and 5 on October 9. Shops B and C had already come down from the plateau by then: B lost 40 percent on September 15, C declined from the end of September. Both also fell to index 3 to 5 on October 9 and 10. Since then, desktop impressions in all shops have been at 4 to 6 percent of the April level.

One more small property I have access to did not get the rise in May. It got its own version from September 28: desktop impressions at 5–12x the April level for ten days, then index 83 on October 8 and 4 on October 9. Same end, different start.

Clicks did not follow

Impressions and clicks: May 25 to October 4 vs. April 15 to May 14 (= 100 %). CTR: April → window → October 9 to 18. Since October 9, Shop C has had too few desktop impressions for a CTR.

The drop in clicks itself is not the finding. These are only desktop clicks on merchant listings, a narrow slice of each shop’s search traffic. Shop A shows the same autumn pattern as in 2025, and on September 15 free product listings largely disappeared in the EEA (more on that below), which hit every shop there, regardless of what it does. The finding is the gap between the impression curve and the click curve, and the day it closes.

The part I find most telling: CTR

Since October 9, desktop CTR for merchant listings has been 16.8 percent in Shop A and 5.4 percent in Shop B. That is the range where mobile CTR of these shops was in April (14 and 4 percent), and since October 9 desktop and mobile in both shops have been only a few points apart. Before October 9, desktop CTR was below mobile in every month I have: in Shop A about half in summer 2025, one fifth to one third since late 2025 (3.2 vs. 14 percent in April), and 0.3 percent in the window.

One interpretation: What Google counted as desktop impressions for merchant listings already contained a large share without any chance of a click before May, a much larger one from mid-May, and none since October 9. A real difference in click behavior between desktop and mobile can explain part of the gap before May, but not a desktop CTR that jumps from 0.3 to 16.8 percent overnight. I read October 9 as a change in what is counted, not as traffic that is gone.

Where the impressions came from

Breakdown by country for September 1 to 14, 2026, all devices, merchant listings:

Russia and Bangladesh appear in every shop, each without a single click.

Cross-check with Merchant Center

• Shop A, Brazil: For the same two weeks, Merchant Center reports about 9 percent of the impressions that Search Console reports for merchant listings. The ratio is completely different per country, so the two tools do not count the same thing. Merchant Center does also show a rise in free listings from May 20/21, but shaped differently, only in the EEA, with real clicks behind it, and without a stop on October 9. I take these to be two different things that started in the same week: Google changed a surface for free listings in the EEA, and Search Console started counting many desktop impressions that nobody clicks.
• Shop C, the other way round: Its Merchant Center reports impressions and clicks for free listings that run flat from April to October, with no step in May, no step on October 9, almost exclusively Germany, and ten times or more as many clicks as Search Console attributes to merchant listings. According to Google’s help page, Merchant Center also counts the Shopping tab, Images, Lens and YouTube, while the “Merchant listings” search appearance in Search Console only covers web search. The inflated part sat exclusively on the Search Console side.

For Shop C this can be shown as a chart: the same property, once from Search Console, once from Merchant Center, both indexed to April.

What it is not

• Not the impression issue Google documented in April. According to Google’s data anomalies page, impressions were counted inaccurately from May 13, 2025 to April 27, 2026, clicks were not. After the fix, reported impressions dropped, so they had been too high (Search Engine Land). But that issue ended three weeks before the rise here begins, and its fix does not show in the merchant listing series of the large shops: desktop index 96 before and 104 after April 27 in Shop A, 101 and 100 in Shop B. The rise from May 16 and the stop on October 9 are a separate event. Within the window, there are entries on Discover and on AI answers, none on merchant listings or desktop impressions.
• Not demand. Keyword Planner volumes for the matching product categories in Germany, Austria, Switzerland and the UK moved by up to about 20 percent between April and September, seasonally, not by 3–7x. For the countries most of the impressions came from, I had no volumes, and monthly values cannot reflect daily steps anyway.
• Not the removal of free product listings in the EEA. According to Search Engine Roundtable, Google’s Ads Liaison confirmed in a comment on September 18, 2026 that Google removed free product listings in the EEA because of the Digital Markets Act. In mid-September, the removal mainly hit clicks in the EEA, impressions moved in both directions, and most of the inflated volume here lies outside the EEA. Shop B lost 40 percent of its desktop impressions on September 15, so both effects overlap in the data.
• Not the “poisoned” results pages that rank tracker vendors reported from September 13. Wrong start date, and that concerns fetched results pages, not Search Console counting.

My interpretation, with caveats

Desktop only, countries without clicks, CTR near zero, a hard start, a hard end, the same days in shops that have nothing in common, and a Merchant Center that counts something else: this looks like automated queries that Google counted as impressions for merchant listings, at a lower level before May, at a much higher level from mid-May, and not at all since October 9. I cannot prove that from the outside. The alternative is a change in reporting logic on Google’s side in May and again in October, and I cannot rule that out either.

Strength of evidence at a glance

What does this mean for you?

1. Impressions, CTR and position from May 16 to October 8, 2026 are distorted at property level if your property has merchant listings. In Shop A, average position from mid-September to October 8 was 5.5 and CTR 3.1 percent, right afterwards 8.1 and 4.6 percent, and the step falls on October 8/9. Nothing changed at the shop on that day.
2. Year-over-year comparisons with impressions are not usable for this window. Clicks were not inflated, compare clicks.
3. Filter by search appearance and device before you read a trend. And do not take the April level as “normal” either; it was already far above what Google counts today.
4. Check your countries. If many impressions come from countries where you never get a click, that often explains a falling CTR better than any change to your pages.

Sources

• Google: Data anomalies in Search Console (entry of April 3, 2026 on inaccurate impressions May 13, 2025 to April 27, 2026; 2026 entries on Discover and AI answers). Checked on October 20, 2026.
• Search Engine Land: Google Search Console bug inflated impression counts (report on the issue May 13, 2025 to April 27, 2026: impressions counted too high, clicks not affected).
• Google: Performance report (Search) in the Search Console help: how impressions, clicks and position are counted, “Search appearance” dimension.
• Google: Free product listings in the Merchant Center help: surfaces Google Search, Shopping tab, Google Images, Google Lens, YouTube, Maps.
• Barry Schwartz: Report on the removal of free product listings in the EEA, Search Engine Roundtable, September 18, 2026, with measurements from Productrise and Semrush and the confirmation by Google’s Ads Liaison in a comment. As of October 20, 2026, there is no documentation of its own from Google.
• Own data: Search Console and Merchant Center of several online shops, with the operators’ permission, anonymized; Keyword Planner volumes via the Google Ads API.