Key takeaways: Dayparting works when three things are true: your hourly data shows a repeatable difference in conversion rate or cost per order across the day, the account has enough clicks per hour block to trust the pattern, and you have a way to change bids by schedule. When it works, the usual result is the same sales at 10 to 20% lower spend. When the pattern is weak, dayparting adds complexity and nothing else.
Dayparting on Amazon works for some accounts and does nothing for others, and the hourly data tells you which you are before you start. If conversion rate by hour is flat, dayparting cannot help; you would be moving spend between hours that perform the same. If the data shows a repeatable pattern, for example late-evening clicks that convert at half the daytime rate, reducing bids in the weak hours saves spend without losing many sales.
What dayparting is and what Amazon gives you
Dayparting means changing bids or budgets by time of day and day of week. Amazon's native options are limited. Budget rules can raise budgets on a schedule; there is no native bid schedule in Campaign Manager. Hourly performance data is available through Amazon Marketing Stream, which most third-party tools use to offer scheduled bid changes.
So a dayparting test has two prerequisites: hourly data you can read, and a tool or a disciplined manual routine that can change bids on a schedule and change them back.
Reading the hourly data
Pull at least four weeks of hourly performance for the campaigns you want to test. Group by hour of day, then look at clicks, conversion rate and cost per order in each block. Weekday and weekend often differ, so split them if the volume allows.
You are looking for a pattern that repeats week over week, not a single bad Tuesday. The usual patterns we see fall into a few types.
| Pattern | What it looks like | Dayparting response |
|---|---|---|
| Flat | Conversion within a narrow band all day | None. Leave bids alone |
| Late-night trough | Clicks after a certain hour convert far below average | Lower bids 20–40% in those hours |
| Commuter peaks | Conversion strongest in early morning and evening | Raise bids modestly in peaks, hold elsewhere |
| Budget exhaustion | Budget runs out mid-afternoon, evening conversion lost | Fix budget or reduce morning bids before dayparting |
| Weekend shift | Weekend converts differently from weekdays | Separate weekday and weekend bid sets |
The budget-exhaustion pattern is the one to catch first. If a campaign runs out of budget at 3pm, the evening hours show no data at all, and any dayparting analysis is built on a day that ended early. Fix the budget, collect another two weeks, then look again. That is covered in Amazon PPC Budgets: Out-of-Budget Campaigns Without Overspending.
How much data is enough
A conversion rate on twenty clicks is noise. An hour block needs enough clicks over the test window to make its conversion rate meaningful; a working minimum is a few hundred clicks per block over the analysis period. A campaign spending $20 a day will not have that in any single hour, and grouping into four or six-hour blocks is the only honest way to read it. A campaign spending $500 a day can be read hour by hour.
This is why dayparting is a tool for higher-spend accounts and campaigns. Below a certain volume, the hourly pattern you think you see is variation, and acting on it moves spend at random.
Running the test
Read this flowchart as text
- 4+ weeks of hourly data for one campaign group
- Repeatable pattern in cost per order by hour?
- No dayparting. Revisit in a quarter
- Enough clicks per block to trust it?
- Group into 4–6 hour blocks or wait for volume
- Set schedule: lower bids in weak blocks, hold or raise in strong
- Run 3–4 weeks. Change nothing else
- Same sales at lower spend, or more sales at same ACoS?
- Keep. Re-check pattern quarterly
- Revert. Pattern was not real
"Change nothing else" is the hard part. If you daypart and also add negatives, adjust placement modifiers and change the main image in the same month, you will not know what the dayparting did. Run it on one campaign group with everything else frozen, and compare against a similar campaign group that was left alone.
The measure of success is one of two things: the same sales at lower spend, or more sales at the same ACoS. If total sales fall, the reduced hours were carrying more than the data suggested. Full detail on how we run and judge the test is in E8 · Adjusting bids by time of day.
Where dayparting goes wrong
Pausing instead of reducing. Setting bids to zero or pausing campaigns overnight resets the auction and can cost impressions in the first hours after they resume. Reduce bids in weak hours; do not switch off.
Confusing time zones. Amazon reports in the marketplace's time zone, not yours. An account run from Houston for the UK marketplace has a six-hour offset. Check which zone the hourly data uses before you build a schedule.
Over-fitting. A schedule with twenty-four different bid levels per day, each based on a few dozen clicks, is fitting noise. Three or four blocks with clear differences is usually all the data supports.
Never revisiting. Shopping patterns move with seasons, daylight and the competitor set. A schedule built in February may be wrong by August. Re-run the hourly analysis quarterly and before any peak event.
Dayparting a budget-starved campaign. If a campaign is capped by budget, the first fix is the budget or the bids, not the schedule.
B18How we choose placementsmore of your budget on the placement that converts best for your product.
Frequently asked questions
Does dayparting work on Amazon?
For accounts with enough volume and a real hourly pattern in cost per order, yes; the typical outcome is the same sales at 10 to 20% lower spend. For accounts with flat hourly conversion or low volume, it does nothing useful and adds a layer of complexity.
Can I daypart without a third-party tool?
Partly. Amazon's budget rules can raise budgets on a schedule, and you can change bids manually at set times, but there is no native bid schedule. Most accounts that daypart seriously use a tool built on Amazon Marketing Stream hourly data.
Should I pause campaigns overnight to save money?
No. Reduce bids in the weak hours instead. Pausing resets the campaign's activity and can cost impressions when it resumes, and some overnight clicks do convert. A 30% bid reduction keeps a presence at a lower cost.
How long should a dayparting test run?
Three to four weeks with no other changes to the test campaigns, compared against a similar group left alone. Shorter tests are swamped by weekly variation; longer tests start to overlap with seasonal shifts.