Amazon PPC dayparting means lowering bids or budget during the hours a campaign converts poorly and keeping full strength during the hours it converts well. It pays off only when a high-spend campaign shows an hour-of-day gap in revenue per click that repeats across months of data. On most campaigns the hourly differences are noise, and cutting hours gives up more orders than it saves in spend.
Dayparting tends to get sold as a quick fix for a high ACOS. After 6+ years on Amazon Ads, my view is that it is a finishing tool for an already-clean campaign, and on the wrong one it cuts orders while looking like discipline. Below: the data test I run first, the three ways it backfires, a decision table, and my setup when the numbers earn it.
What is dayparting in Amazon PPC?
Dayparting is scheduling how hard your ads compete by time of day. In Amazon PPC it means changing bids, budgets, or campaign status on an hourly pattern: less pressure in hours that historically convert poorly, full pressure in hours that convert well. It can be run by hand, with rules, or through third-party PPC software.
Two tactics share the word. Bid dayparting keeps the campaign live all day and changes its price per click by hour; budget or on/off dayparting decides whether it shows at all. One is a dimmer, the other a light switch, and the switch is riskier: a paused campaign does not pay less for the 10pm shopper, it hands that ad slot to a competitor.
One version runs without anyone choosing it: a campaign that runs out of budget every afternoon is already dayparted by its cap, whatever the evening would have converted at. That has its own fix, in my guide to campaigns that run out of budget early.
When does dayparting actually help on Amazon?
Dayparting helps when three things hold at once: the campaign spends enough that an hourly gap is worth real money, each block of hours gathers enough clicks for its numbers to mean something, and the same weak block shows up in every period you check. Miss any one and the schedule is a guess.
Judge hours by revenue per click, not sales. The busiest hours produce the most sales simply because more people shop then; revenue per click (a block's ad sales divided by its clicks) shows whether a 9pm click is worth more than a 9am one. If you daypart, price each block the way I price any bid, from that block's revenue per click times target ACOS; my guide to setting Amazon PPC bids walks through the math.
Is the hourly gap real or noise? A worked example
On most campaigns, the hourly gaps a report shows are largely noise. Splitting a day into hours or blocks divides your clicks into small piles, and small piles produce conversion rates that swing widely on luck alone. Before trusting a weak block, compare its gap with the range chance would produce at that number of clicks.
On my teaching dataset's averages ($0.79 CPC, 9.1% conversion rate), a campaign spending $40 a day buys about 50 clicks. Split the day into six four-hour blocks, assume traffic spreads evenly (it rarely does, and thinner blocks are shakier still), and give every block the same true 9.1% rate, so no hour is genuinely better. The report can still show this:
| Data window | Clicks per block | Expected orders | Orders luck produces, 19 times in 20 | Conversion rate that reads as |
|---|---|---|---|---|
| 30 days | About 250 | About 23 | About 14 to 32 | About 5.5% to 12.7% |
| 90 days | About 750 | About 68 | About 53 to 84 | About 7.0% to 11.2% |
On a month of data, a block converting at exactly the average can report 5.5%, which reads as a disaster, or 12.7%, which reads as your best hours. Three months narrows the range, but a 7% block beside an 11% block can still be two identical blocks with different luck. More clicks help: ten times the clicks shrinks the range to about a third of its width. That is the real reason dayparting belongs to high-spend campaigns.
Guessing wrong has a price. Pause the block that read 5.5%, and if the number was luck, you give up about 8 clicks a day earning the dataset's $2.43 in revenue per click: roughly $20 a day in ad sales, to save under $7 in spend, every day the rule runs.
My practical test is simpler than the math: split the hourly data into separate months and believe a weak block only if it is weak in every one, by more than luck explains at its click count. Then check your source. Hour-level numbers must come from a report or tool that records performance by hour; the bulk file totals the whole date range, and none of the tools on this site read hourly data either.
Three ways dayparting quietly hurts Amazon campaigns
Dayparting backfires in three predictable ways: it treats thin hourly data as a pattern, it pulls the campaign out of hours whose orders still build keyword sales history, and it stacks a manual time rule on top of Amazon's own per-impression bid adjustments. Each can raise ACOS or cost rank while looking like tighter control.
1. Reading noise as a pattern
A block cut for looking weak on too few clicks is the worked example in action. If it was really average, you lose its orders while ACOS barely moves; if it was a better-than-average block in an unlucky month, ACOS rises.
2. Starving the hours that feed ranking
Amazon does not publish how organic rank works, but the working model most operators use, me included, is that rank on a keyword follows the sales you make on it. Under that model an 11pm order builds a keyword's sales history exactly like an 11am one, so cutting hours cuts ranking fuel along with spend.
I have watched the mechanism work in the useful direction. In the Toys & Games account I rebuilt, concentrating spend on the head terms that already converted lifted their organic rank, and monthly sales climbed from $6,605 to $14,105 over 90 days as organic orders stacked on top of ad orders. So I never daypart a launch or a ranking push (more in my guide to how PPC drives organic ranking and my product launch PPC guide).
3. Fighting Amazon's own bid adjustments
With dynamic bidding on, Amazon already adjusts your bids in real time according to how likely each impression is to convert. A manual hourly cut stacks a second, blunter adjustment on the same bid, and when results change you can no longer tell which system caused it.
Amazon does not say which signals feed that prediction, so I will not claim it reads the clock. But on down only, your schedule can discount impressions Amazon already discounted, pricing you out of auctions that were fine at Amazon's number. On up and down, it can cut the base bid in the block where Amazon wanted to raise it for a likely buyer: exactly the click that campaign exists to win.
Should you daypart your Amazon campaigns? The decision table
For most private-label campaigns the answer is no, and for a few high-spend, proven campaigns it is a careful, bid-only yes. The rule underneath every row: daypart only when the gap is proven and the campaign's job is efficiency, never when its job is discovery or ranking.
| Situation | Daypart? | Why |
|---|---|---|
| High-spend exact match campaign, 90+ days of hourly data, the same weak block every month | Yes, by bid, in blocks | The gap is real and worth money; price each block from its own revenue per click |
| Small daily budget | No | No block gathers enough clicks to separate a real gap from luck |
| Out of budget by mid-afternoon | Not yet | The cap already dayparts it; fix the budget or the leak first |
| Launch or ranking push on a priority keyword | No | Every order builds keyword sales history, whatever the hour |
| Auto or broad match discovery campaign | No | Its job is finding terms; weak hours there usually mean weak terms |
| Weak block traced to irrelevant search terms | No, negate instead | The problem is the term, not the time; a negative keyword fixes every hour at once |
| Up-and-down dynamic bidding on proven terms | Rarely; test one block | Amazon already moves those bids per impression, and a schedule can fight it |
| One schedule applied account-wide | No | Campaigns convert on different clocks; a blanket rule punishes the ones that sell at night |
How I daypart when the data earns it, and what I fix first
Before any schedule, I fix the problems that only resemble an hour problem: zero-order search terms, search term bleed, overpriced bids, a cap that runs dry mid-day. If a clean campaign still shows the same weak block month after month, I daypart by bid, in blocks, with a trim rather than a switch-off.
Scale is why the fixes come first. A weak four-hour block covers a sixth of the clock. In a $45K/mo skincare brand I took on, wasted spend was 41% of all ad spend, and the rebuild brought it to 16% in 90 days. Start with the Wasted Spend Finder and my search term bleed guide, or watch the full read on a real anonymized account in the audit demo.
When a schedule does go on, the campaign stays live every hour and only the price moves. The weak block's bid comes from that block's revenue per click times target ACOS. Afterward I compare the whole campaign's orders and ACOS against an equal period before; if ACOS is no better and orders are down, the schedule comes off. I re-read hour patterns quarterly, never weekly.
Frequently asked questions
Does dayparting work on Amazon PPC?
Sometimes. It works when a high-spend campaign shows the same weak block of hours in several separate months, by more than luck explains, and you trim bids in that block rather than pausing. On small or unproven campaigns the hourly data is too thin to act on, and a schedule mostly removes orders that were fine.
What are the best hours to run Amazon ads?
There is no universal answer; a generic peak-hours chart cannot tell you when your own shoppers buy. Group your hourly data into blocks and compare revenue per click, not sales, because the busiest hours win on sales just by having more shoppers, which says nothing about what each click is worth.
Should I turn off my Amazon ads at night?
Usually not. Pausing hands those ad slots to competitors, and night orders likely build keyword sales history like any others. Shoppers also browse at one hour and buy at another, so depending on how your report assigns a sale, late clicks can look weaker than they are. If night stays weak for months, lower its bids instead.
Can dayparting hurt my organic ranking on Amazon?
It can. Amazon keeps its ranking formula private, but the common operating assumption is that keyword rank tracks the orders you win on that keyword, and a midnight order counts like a noon one. Cutting hours during a launch or ranking push removes orders the keyword needed, so I never daypart in either window.
How long should I collect hourly data before dayparting?
Longer than feels necessary. I read hour-of-day patterns on 90 days of data with promotions, Prime Day, and holiday weeks removed, then confirm the weak block is weak in each month on its own. Even then, on a $40-a-day campaign, a four-hour block's conversion rate is uncertain by about two points either way.
Sources
- Guide to dynamic bidding with Sponsored Products, Amazon Ads
Check the leaks before you touch the clock
A schedule only earns a trial on a clean campaign, so check that first. The free Account Health Snapshot grades wasted spend, conversion rate, and ACOS from a 60-day bulk file against the phase benchmarks. No email, no account, nothing stored. If the grades hold up and you still see an hourly pattern you trust, bring it to the free 30-minute diagnosis call and I will run the noise test on your numbers with you.