Two figures about Amazon click costs have been in circulation this year, and read side by side they appear to flatly contradict each other.
Amazon has stated that average winning bids on Sponsored Products search ads fell roughly 50% between 2019 and 2025. The FTC, in its August 2026 complaint, alleges that pay-per-click costs rose roughly 50% on major shopping days over a broadly comparable period.
Down 50%. Up 50%. Both figures come from parties with the data and strong reasons to be careful with it. The interesting thing is that they are not necessarily in conflict, and understanding why is more useful to your account than knowing which side to believe.
Why both numbers can be true
They measure different things.
Average winning bid across all auctions is a blended figure spanning every auction Amazon runs, at every hour, on every day, across a marketplace that has grown enormously in the number of participating sellers. As the long tail of advertisers expands, an increasing share of total auctions are low-competition auctions on obscure terms. Those clear cheaply. Add enough of them and the blended average falls, even if nothing got cheaper for anyone in a contested category.
Cost on major shopping days is the opposite kind of measure. It isolates the periods where the largest number of advertisers converge on the same inventory at the same time. Concentration raises clearing prices. If everyone shows up on the same four days, those four days get expensive regardless of what the annual blended average does.
None of that resolves the underlying legal dispute, and it is not meant to. It does mean that "are CPCs going up or down" is not a well-formed question, and that an answer to it is not actionable no matter which direction it points.
What this means for how you measure your own account
The blended account CPC that sits at the top of most reporting is subject to exactly the same dilution problem. It is a weighted average across your competitive terms and your cheap long-tail terms, across your expensive hours and your quiet ones. It can hold perfectly steady for a year while the distribution underneath it shifts substantially.
Which is why the useful move is to stop treating it as a headline metric and start segmenting.
By day of week. Most accounts have a two to three times spread between their cheapest and most expensive day. That spread is invisible in the weekly average and it is directly actionable.
By hour. The bigger spread, and the one most sellers never look at. Auction competition is not uniform across a 24-hour cycle, and neither is conversion rate. The two do not move together, which is the entire opportunity. Our hourly heatmap walkthrough covers how to build this view.
By event versus baseline. Peak periods should be analyzed as a separate population, not folded into a monthly average where they distort everything around them.
By match type. Broad match campaigns accumulate expensive irrelevant auctions that pull the average around. Segmenting here often reveals that a rising blended CPC is really one campaign's search term hygiene problem.
The practical consequence
If peak-period concentration is where cost genuinely compounds, then peak-period policy is where the money is, and most accounts set that policy reactively.
The typical pattern on a major shopping day: budgets are raised the morning of, spend runs hot through the high-traffic hours, several campaigns exhaust their daily budget before the evening, and the account goes dark during a window that was still converting. Then someone raises budgets again the next morning and nobody restores them for three weeks.
Parkway Home ran that exact sequence through two consecutive peak events before pulling the hourly data and finding that roughly a third of their event-day spend landed in a four-hour morning block with a conversion rate well below their evening window. They had been buying the most expensive hours of the most expensive days, and then running out of money before the hours that worked.
The fix was not a bid change. It was deciding in advance how much budget was allowed to be live in which hours, and letting it restore to baseline automatically when the event ended. Event rules exist for precisely this shape of problem, and the restore step is the part people forget when they do it by hand.
What to take from all of this
You cannot control the clearing price of an auction. You have never been able to, and no reporting change or legal outcome will hand you that lever.
What you control is exposure. Which campaigns are live, with how much budget, during which hours, on which days. In a market where cost concentrates into narrow high-competition windows, exposure management is the variable that actually moves your blended outcome, and it is the one most accounts leave on default.
Frequently asked questions
Are Amazon CPCs going up or down? It depends entirely on what you measure. Amazon has stated that average winning bids on Sponsored Products search ads fell roughly 50% between 2019 and 2025. The FTC has alleged that pay-per-click costs rose roughly 50% on major shopping days over a comparable period. These measure different things: a blended average across all auctions versus cost concentrated in peak demand periods.
Why is my Amazon CPC higher on peak days? Peak shopping days concentrate a large number of advertisers into the same auctions at the same time. More simultaneous bidders on the same inventory raises clearing prices. The effect is strongest during high-traffic hours on high-traffic days, which is also when budgets are most likely to exhaust early.
Should I track blended account CPC? As a headline number, no. A blended average across all campaigns, hours, and days hides the concentration that actually drives cost. Segment by day of week, by hour, and by event versus non-event periods. The averages tend to move very little while the segments move a lot.
How do I control CPC on peak shopping days? You cannot control the clearing price of an auction. You can control exposure: how much budget is live, on which campaigns, during which hours. Setting peak-period budget and schedule policy in advance rather than reacting mid-event is the practical lever.
Off Hours is dayparting and budget automation for Amazon Ads. Set your rules once in plain English, and they run every day across every account you manage. Start a free 14-day trial.