Amazon PPC Dayparting with Claude: Scheduled Bid Increases That Amazon Runs for You

Dayparting on Amazon means bidding harder in the hours when your ads convert and holding back in the hours when they don't. For years that took a third-party tool that changed bids on a timer around the clock. Amazon now runs the schedule itself: a Sponsored Products bid rule raises bids during a window you set, and Amazon applies it every day without anyone sending an update.

This post covers how to find your hours, how Amazon's scheduled bid increases work, what they cannot do, and how to set one up from Claude with DataDoe, starting with a dry run.

Step one: find the hours that convert

Dayparting is only as good as the hourly data behind it. In DataDoe, hourly ads performance comes from Amazon Marketing Stream, in the table amazon_ads_performance_by_campaign_by_hour: one row per campaign per hour. The Marketing Stream tables are off by default in DataDoe, so turn them on first.

A few rules keep the analysis honest:

• Hours are in the Ads profile's time zone, including daylight saving. That is the same clock the bid rule uses, which is what you want.
• Group, then recompute. Filter on the local date, group by weekday and hour, sum spend, clicks, orders and sales, and only then calculate ACoS or conversion rate. An average of hourly ACoS values is not the ACoS of those hours.
• Empty is not zero. A missing value means the data is not available for that hour, not that nothing happened.
• It is the ad hour, not the purchase hour. The hour belongs to the ad activity. It is not when the shopper finally bought, and it is not hourly profit. Profit stays a daily figure.
• Give it time. Collect at least a few weeks so that each weekday and hour has enough clicks to mean something.

What you are looking for is a block of hours, often a few hours on weekday evenings or weekend afternoons, where conversion rate is clearly higher than the rest of the day on enough clicks.

How Amazon's scheduled bid increases work

A scheduled bid rule for Sponsored Products has four parts:

• An increase, as a whole percentage from 1 to 100.
• A recurrence: daily, or weekly on chosen days, such as Saturday and Sunday.
• One time window per rule, in whole hours in the profile's time zone, for example 17:00 to 21:00. The window can run to 24:00.
• A duration: the date the rule starts and, optionally, the date it ends.

A rule does nothing until it is attached to one or more campaigns. From then on Amazon raises the bids in those campaigns during the window and returns them to normal afterwards. DataDoe does not send a bid change when the window opens or closes, and no agent needs to be online.

Rules only raise bids

The rule is an increase, never a decrease. That changes how you set it up:

1. Set the campaign's base bids for your ordinary hours, the level at which quiet hours are still worth paying for.
2. Add a rule that raises bids in your strong window.
3. For two strong windows, such as weekday evenings and weekend afternoons, create two rules, one window each.

If your data says the night hours lose money, lower the base bid and let the rule lift the good hours back up. Trying to dock bids overnight with a rule will not work.

Set it up from Claude

With DataDoe connected to Claude, the whole loop runs in one conversation. A typical flow:

1. "Using the last six weeks of hourly ads data, show conversion rate and ACoS by weekday and hour for my top five Sponsored Products campaigns. Only include hours with at least 30 clicks."
2. "Propose one scheduled bid increase per strong window. Keep the increase at 25% or lower and explain why each window qualifies."
3. "Run it as a dry run first." Claude builds the rule, for example a 20% increase on Saturday and Sunday from 17:00 to 21:00 starting next Monday, and DataDoe validates the payload without touching your account.
4. "Create it for real and attach it to these three campaigns." The rule is created, its ID comes back, and a second step attaches it to the campaigns.

Before any live change, the Action type has to be switched on. Actions stay off until you enable them, most AI clients ask before a change, and every change is logged. The account also needs "Read and write" access on the Accounts page in DataDoe. Both settings live in DataDoe, not in Claude.

What to know before you rely on it

• No unlink and no delete. There is no Action to detach a rule from one campaign or to delete it. You can pause it, but pausing stops it for every campaign that uses it. Plan one rule per group of campaigns that should move together.
• Keep the rule IDs. Rules can be looked up by their IDs or by one campaign at a time. There is no search that lists every rule on the account, so ask Claude to note the IDs it creates. The DataDoe Action history keeps them too.
• Changing a schedule. Send the dates and the recurrence together when you edit a rule. DataDoe loads the current rule and sends Amazon the complete version, including the parts you did not change.
• Sponsored Products only. Scheduled bid increases exist for Sponsored Products. For Sponsored Brands and Sponsored Display, use budget rules instead.
• Amazon decides eligibility. Amazon can reject a rule when the profile or campaign is not eligible. The Action result shows Amazon's message.
• Batches. Up to 10 rules per Action through MCP and 25 through the REST API.

Budget rules: the other half of dayparting

Raising bids in the evening does not help if the budget runs out at 3 pm. Budget rules cover that side, for Sponsored Products, Sponsored Brands and Sponsored Display:

• Schedule rules raise a campaign's daily budget during a date range or an Amazon event, with an optional time window within the day.
• Performance rules raise it when the campaign meets a threshold, such as ROAS or ACoS for Sponsored Products and Sponsored Display.

One limit: Amazon retired the budget rule history endpoints in 2023, so there is no record of when a rule actually fired. Judge budget rules by their results in the hourly data and by budget usage from Marketing Stream.

Measure it after two to four weeks

Compare the window hours before and after the rule, on the same weekdays, and recompute the ratios from summed spend and sales:

• "For the campaigns with the weekend evening rule, compare clicks, conversion rate and ACoS between 17:00 and 21:00 on weekends, the four weeks before the rule against the four weeks after."
• "Did spend outside the window change after the rule started? Did any campaign run out of budget before 17:00?"
• "Show bid and budget changes on these campaigns since the rule started." Campaign changes are kept in amazon_ads_entity_history, so a manual bid edit in the middle of the test is easy to spot.

If the window holds up, widen the increase in small steps. If ACoS in the window drifts above the rest of the day, pause the rule and look at the hours again.

For everything else Claude can read and change in your ad account, see Amazon Ads in Claude and ChatGPT. For the metric to judge dayparting by, see ACoS vs TACoS.

Questions people ask

Does Amazon support dayparting natively?

Yes, for Sponsored Products, through scheduled bid increases: a percentage increase during a daily or weekly time window that Amazon applies itself. Sponsored Brands and Sponsored Display have schedule-based budget rules instead.

Does DataDoe change my bids every hour?

No. DataDoe creates the rule and attaches it to campaigns once. Amazon applies the schedule from then on.

Can I lower bids at night with a rule?

No. Rules only increase bids. Lower the base bid for your quiet hours and use a rule to raise the strong ones.

Which time zone does the schedule use?

The time zone of the Amazon Ads profile, the same one the hourly performance data uses.