So what actually happened?
- The exit: Between July 21 and 23, Amazon’s Shopping impression share dropped to 0% across major markets, following weeks of pullback in the U.S. This wasn’t a gentle fade; it was a near-overnight disappearance from Shopping auctions.
- The return: By August 23–26, Amazon resumed shopping ads across many international domains, but notably not in the U.S. Several industry trackers observed impression share rebounding to pre-exit levels outside the U.S. within days.
Why the vanishing act? No official statement, however, the dominant theories are: a market-by-market incrementality test, shifting budgets away from Google, or some hardball rate negotiations. All are plausible, but none are confirmed.
What Happened While They Were Gone?
Let’s just say the market didn’t wait. Competitors swooped in quickly, soaking up impressions and pushing CPCs into temporary relief. Several sources – including Incubeta’s Zuzana Hughes – reported short-term efficiency gains. These reverted toward baseline within weeks as auctions re-equilibrated. Notably, some saw increased volume but mixed ROAS, a classic sign of “filling the gap” without Amazon’s conversion gravity, but larger advertisers like Temu and Shein appear to have capitalized on the outage:
Two big takeaways:
- Auction vacuums fill quickly – Once a dominant bidder exits, competitors quickly respond, rapidly compressing the opportunity window.
- Category structure matters more than headlines – Amazon does not compete evenly across every vertical, so the distribution of gains/losses is highly uneven
What we observed across our clients
Client A: After Amazon’s return, their impression share came back softer, down 28% from the pre-exit baseline. But our CTR improved by 5%, and CPC dropped 7%. That hints at Amazon playing more selectively post-return. Temu followed a similar trend – massive surge while Amazon was out (~40% impression share increase), then a moderate pullback, but still more present than before.
Client B: ROAS post-return dipped 20% compared to the Amazon-free window, but still landed 29% higher than before the exit. CTR jumped 30%, thanks to stronger AOVs and conversion rates that gave us room to bid more aggressively while staying profitable. Temu and Shein? Their post-return impression shares remained consistent at ~20% and ~12% respectively, up significantly from pre-exit levels.
In short, when Amazon doesn’t come back swinging in a category, aggressive challengers step in and stay in.
Was This a Test? A Negotiation? Or Both?
We can’t know for sure. But we can observe the footprint: a rapid exit, a selective international return, and category-level asymmetries that look like budget reallocation to the highest-yield verticals/regions first, with the U.S. conspicuously slower to re-engage. Read it primarily as a capital-allocation exercise with an eye on incrementality and platform economics.
Navigating What’s Next: Three Principles to Stick To
- Vacuums are short-lived. Don’t just react, prepare. Build systems that can flex fast when the giants flinch.
- Vertical reality > viral narrative. Your category’s behaviour matters more than the headline; always run your own analysis.
- Plan for agility. In shocks, budget rigidity is value-destructive. Build systems where bidding can flex automatically to mop up changes in demand.
