admin 
On July 21, 2025, Amazon’s U.S. Google Shopping impression share against the median competing retailer sat at 60%. Two days later, its impression share reached zero.
We witnessed the same withdrawal across Amazon’s international markets as well. Smarter Ecommerce tracked Amazon disappearing from Shopping auctions in the United Kingdom, Germany, and more than a dozen other markets in real-time. Within one month, Amazon had returned across every international market Smarter Ecommerce tracked. The United States remained the sole exception.
Now, one year later, Amazon’s U.S. Google Shopping impression share is still 0%, according to Tinuiti’s Q2 2026 Digital Ads Benchmark Report.
Neither Amazon nor Google has publicly explained the decision. However, one year of data allows us to evaluate and hypothesize what happened after Amazon left. Initial coverage predicted lower cost-per-click and more available impressions for competing retailers.
One year of data now gives us a clearer view of what happened following Amazon’s exit. The auction refilled with other retailers quickly, Amazon returned internationally, and its U.S. absence became increasingly difficult to dismiss as a temporary pause.
Let’s start by recapping what happened just over a year ago and the industry’s reaction.
Amazon’s Exit Was Framed As An Immediate Opportunity
Mike Ryan and the Smarter Ecommerce team drew early attention to Amazon’s withdrawal as it unfolded. Ryan credited Josh Duggan of Vervaunt and Andy Taylor of Tinuiti with identifying the trend earlier.
In his post, Ryan stated:
I have not seen an ad stop this dramatic from Amazon since the acute phase of the pandemic lockdowns, which coincided with a 20% drop in cost-per-click (bear in mind there were many confounding variables, that’s why I say ‘coincided’)
Ryan focused first on what Amazon’s exit could mean for Google. He described Amazon as a major source of revenue for Alphabet and “a tide that lifts all boats in terms of auction prices.”
Smarter Ecommerce’s analysis used even stronger language. It called the withdrawal a potential “gold rush for cheaper clicks and new customers” and described the missing demand as a “colossal void.”
Those descriptions reflected the size of Amazon’s Shopping presence. They also shaped the immediate assumption that competing retailers could gain traffic at a lower cost.
Why Lower CPCs Seemed Likely
Amazon’s impression share against the median retailer remained near 60% throughout the first quarter of 2025. Removing bids with that reach should reduce auction pressure across many product categories. Amazon’s 2020 withdrawal gave advertisers another reason to expect lower costs. Its temporary pause during the pandemic coincided with a 20% decline in CPCs.
Ryan carefully described the two events as coinciding. Pandemic-era changes in demand, supply, budgets, and consumer behavior also influenced advertising costs.
The comparison still gave the industry a recent example of CPCs falling while Amazon stayed out of Shopping auctions. Ryan predicted that competing retailers would welcome the 2025 withdrawal. Brands and marketplace sellers faced a less favorable possibility. Many receive additional Shopping exposure when Amazon advertises products sold through its marketplace. Amazon’s exit removed some of that visibility.
Optmyzr took a more cautious view of the opportunity in an August 1 response. It advised advertisers to monitor CPCs, impression share, spending, and conversions before adjusting their budgets. In their response, Optmyzr noted that some categories could see lower CPCs. Others could face more aggressive bidding as competitors moved into the available auctions.
Duggan had already seen little evidence of an immediate pricing shift across Vervaunt’s clients.
“Amazon appears in ~30% of Shopping auctions across our client base,” Duggan said. “CPCs haven’t dropped…but it’s definitely one to review.”
Other advertisers saw an opportunity to replace some of the exposure Amazon previously provided. Brandon Yann of VML recommended that brands consider running their own Search-to-Amazon campaigns with attribution. That approach could help brands continue sending Google traffic to their Amazon listings without relying on Amazon’s Shopping campaigns.
The early reactions pointed to several possible outcomes. Direct retailers could gain impressions, marketplace sellers could lose exposure, and competing advertisers could quickly replace Amazon’s bids.
How Did The Google Shopping Auction Actually Respond?
Tinuiti’s 2025 Q3 benchmark report provides the clearest view of how Google Shopping performed after Amazon left.
Google Shopping spending increased 14% year over year during Q3 2025 across standard Shopping and Performance Max campaigns. Clicks rose 15%, while average CPCs declined 1%.
Contrary to early hopes, Amazon’s withdrawal didn’t produce the sharp CPC decline seen during its 2020 pause.
Tinuiti found that Amazon’s absence supported stronger click growth and reduced some pricing pressure. The effect remained relatively modest because other large retailers were already moving into the available auctions.
Optmyzr’s early analysis showed a similar pattern over a shorter period. It compared July 16 through July 22 with the seven days after Amazon withdrew.
They showed that:
- CPCs fell 8.3%.
- Clicks increased 7.8%.
- Conversion value declined 5.5%, suggesting that cheaper traffic didn’t automatically produce more revenue.
The report acknowledged that its seven-day measurement window may not have captured conversions with longer purchase cycles. Its results still provided an early indication that additional traffic wouldn’t deliver the same value across every category.
Temu, Shein, & Walmart Filled Part Of The Gap
Ironically enough, Temu resumed Google Shopping ads in mid-July after withdrawing in April. Shein returned after sitting out for slightly more than one month.
Walmart also captured a larger share of Shopping impressions during the third quarter, which partially offset Amazon’s withdrawal.
Smarter Ecommerce observed a similar redistribution in Europe. Its data showed the share of advertisers encountering Temu as a competitor increased from 60% to 75%.
The available impressions ended up spreading across several large advertisers rather than shifting to one direct replacement, which likely affected how much CPCs would have fallen for the broader markets.
Tinuiti’s Q4 benchmark report showed the pattern consistency. They reported Google Shopping spend increased 16% YoY, with similar click growth. CPCs only fell about 1%.
No single retailer replaced Amazon across the entire market, which would have been almost impossible anyway. Target became more prominent during peak holiday shopping, while Walmart strengthened its position in December.
Temu maintained a smaller presence during Q4, while Shein reduced its activity as holiday purchasing ramped up.
The U.S. Shopping auction continued growing without Amazon, while competing retailers divided the available impressions. International auctions followed a different path because Amazon itself had already returned.
Amazon Came Back Everywhere Except The United States
Amazon’s July 2025 withdrawal extended well beyond its U.S. Shopping campaigns, as mentioned earlier in the article.
But, just after a month of silence, Smarter Ecommerce reported that Amazon reactivated its Google Shopping presence across every international market it tracked, except the United States.
Mike Ryan reported that Amazon appeared as an account-level competitor for about 75% of European advertisers overnight. The return happened almost as quickly as the original withdrawal.
Smarter Ecommerce’s data shows how coordinated that initial pullback was. Amazon’s UK impression share declined from 55% on July 21 to 43% on July 22. It reached zero the following day.
Amazon’s German impression share followed the same pattern, falling from 37% to 34% before reaching zero. Every domain in Smarter Ecommerce’s 20-market dataset recorded a 0% share on July 23.
Amazon coordinated its withdrawal globally, then treated the United States differently when restoring those campaigns.
Smarter Ecommerce continues tracking Amazon and other large Shopping competitors through its Market Observer. The tracker uses Auction Insights data to follow median impression share and refreshes its market data weekly.
Amazon’s activity elsewhere helps clarify the scope of its U.S. decision.
Tinuiti’s Q4 2025 benchmark found that Amazon remained absent from nearly all U.S. Google Shopping auctions through the holiday season, with a small exception of Amazon Pharmacy presence.
Amazon continued running U.S. Google Search ads during the same period, as well as Microsoft Shopping auctions.
That decision continued through the holiday season and Amazon’s 2026 Prime Day event. Tinuiti’s Q2 2026 Digital Ads Benchmark Report places Amazon’s U.S. Google Shopping impression share against the median retailer at 0%.
One hypothesis fits that behavior: Amazon may have tested whether U.S. Shopping ads produced incremental sales. The available evidence can’t confirm that explanation, but the duration makes it worth looking into further.
See also: Amazon Sellers: Inventory Management Tips For 2025
Incrementality Testing Could Be A Credible Explanation
Tinuiti raised the incrementality possibility before Amazon’s complete withdrawal in July 2025.
Amazon had already sharply reduced its U.S. Google Shopping presence between May 21 and June 8. Its timing didn’t align as closely with tariff developments as the earlier withdrawals by Temu and Shein.
Amazon then returned to full strength during a relatively stable period for tariff policy. Tinuiti’s Q2 2025 benchmark concluded that “other considerations may have been at play.”
Mark Ballard, Tinuiti’s Director of Research, suggested Amazon may have been testing whether its Google Shopping investment generated incremental sales.
Amazon, of all companies, is especially well-positioned to test that difference.
Many shoppers already begin product searches on Amazon, visit the site directly, or purchase through its app. Amazon also receives substantial organic visibility for commercial Google searches.
Turning off Google Shopping would give Amazon a cleaner view of what happened to total sales, direct traffic, app activity, and customer acquisition without those ads.
Smarter Ecommerce also included a large-scale incrementality test among its possible explanations for the July withdrawal. Neither company claimed to know Amazon’s reasoning.
Other explanations remain possible, including profitability targets, budget allocation, or changes in Amazon’s relationship with Google. Amazon has not disclosed its methodology, results, or decision-making process.
But, the duration and geographic split give the incrementality hypothesis more weight.
If Amazon truly did conduct an incrementality test, the continued pause is the only result visible outside the company. Amazon kept the program off through the holiday season and another Prime Day cycle.
That willingness to keep a major channel paused offers a more useful lesson for paid media leaders than speculation about Amazon’s internal results.
What PPC Marketers Can Learn From Amazon’s Absence
Amazon’s decision gives paid media leads two issues to consider:
- How quickly they can respond to auction changes.
- How seriously they measure incrementality.
The auction opportunity was there for many verticals, but it varied by category and wasn’t equally valuable for every advertiser. Other retailers expanded their presence, CPCs moved by different amounts, and additional traffic sometimes produced less conversion value.
Advertisers can’t blindly assume that a major competitor’s exit will create an automatic case for more spending. They need to evaluate the quality and cost of the newly available traffic.
That requires monitoring more than CPCs and impression share. Conversion value, new customer acquisition, contribution margin, and product-level performance help show whether additional volume supports the business.
Teams also need enough flexibility to respond while auction conditions are changing. Predefined budget ranges and approval thresholds can shorten the time between identifying an opportunity and acting on it.
The measurement lesson also requires more organizational preparation.
Pausing a major channel can expose the business to immediate lost revenue. At the same time, it can also challenge years of reported performance if sales remain stable without the channel.
Either outcome affects more than just the paid media team. Leadership, finance, analytics, and other channel owners need to agree on the associated risks of pausing a channel before testing begins.
That agreement should define these items at a minimum:
- The test period dates.
- Acceptable revenue loss.
- Decision criteria.
- Conditions for ending the test early.
- Account for conversion lag and seasonal demand.
Few advertisers could replicate Amazon’s year-long U.S. withdrawal at the same scale, and frankly most don’t need to.
Retailers can test selected markets, product categories, audience groups, or matched geographic regions. The test still needs a meaningful control and enough time to measure effects beyond immediate platform conversions.
Platform-reported ROAS remains useful within that analysis. Establishing incrementality requires additional business measures, including:
- Total revenue.
- Direct traffic.
- Organic demand.
- New customers.
- Repeat purchases.
- Contribution margin.
The appropriate metrics being considered depend on the channel’s role. For example, a prospecting campaign requires different success criteria than branded search or a Shopping program serving existing demand.
Amazon’s decision carries another important limitation. Its brand recognition, direct traffic, app usage, product selection, and customer loyalty differ significantly from those of most retailers.
Other advertisers shouldn’t interpret Amazon’s absence as evidence that Google Shopping lacks incremental value for them. Each company needs to test its own customer behavior and economics.
Amazon may never publish the reasoning or results behind its decision. Paid media leaders can still apply the lesson: define incremental value before testing, secure the authority to act, and prepare to accept the answer.
More Resources:
Featured Image: RikaRahayu/Shutterstock
