Amazon Ads Dayparting: Find the Best Hours From Your Data

Amazon ads dayparting means changing your bids or budgets by hour of day or day of week, so more of your spend lands in the hours when an order costs less. It pays only when your own hourly data shows a real gap in cost per order between hours, and proving that gap takes weeks of data plus an allowance for orders that arrive after the click.

What is Amazon ads dayparting?

Dayparting is a schedule that shifts spend between hours, and it works through one number, cost per order:

Cost per order = spend ÷ orders = CPC ÷ conversion rate

CPC is cost per click (spend ÷ clicks) and conversion rate is orders ÷ clicks. For example, if clicks cost ₹8 all day but convert at 8% in the evening and 2% overnight, an evening order costs ₹100 and an overnight order ₹400. A schedule moves money from the second kind of hour to the first, in one of three ways:

• Bid changes by hour. Lower bids in weak hours and raise them in strong ones. This is the finest control.
• Budget changes. A daily budget is spent whenever the auctions come, so you can't save it for the evening except by bidding less, or pausing, earlier in the day. Raising a budget later in the day only helps a campaign that would otherwise run out.
• Pausing. Switching campaigns off for set hours saves all the spend in those hours and loses all the orders.

The analysis below works the same for Sponsored Products, Sponsored Brands and Sponsored Display. Only the attribution window differs.

What Amazon's console does natively, and what needs the Ads API

The console can raise budgets on a schedule, but lowering bids in your weak hours every day takes the Amazon Ads API or a tool built on it.

In the console:

• Daily budgets, start and end dates, and pausing or enabling a campaign by hand.
• Budget rules, which raise a campaign's budget by a percentage on a schedule you set or when its performance passes a threshold. No budget rule lowers a budget.
• Dynamic bidding: "down only" lets Amazon lower your bid in real time when a click looks less likely to convert, and "up and down" also raises it when a click looks more likely to. Amazon decides auction by auction; you don't set its hourly logic.

Through the Ads API or a tool:

• Lowering bids in weak hours and restoring them afterwards, on a timer, every day.
• Pausing and restarting campaigns at set hours without doing it by hand.

For hour-by-hour numbers, Amazon Marketing Stream, part of the Ads API, sends hourly traffic and conversion data to software you connect; the API's standard reports are daily. Check your campaign settings for newer scheduling options before you pay for a tool for this alone. Our comparison of Ads API automation and SaaS tools covers the build-or-buy choice.

How to find the best time to run Amazon ads from your own data

The best time to run your ads is wherever your cost per order sits clearly below your target, and only your own data shows where. Work per campaign, or per product if a campaign advertises several:

1. Take at least four full weeks of hourly data, so each weekday appears four times, and leave out the most recent 7 days (14 for Sponsored Brands, Sponsored Display and vendors).
2. Confirm the time zone. Data stamped in UTC runs 5 hours 30 minutes behind IST, so the 01:00 UTC hour is 06:30 to 07:30 in India.
3. Group hours into blocks that follow your traffic, such as overnight, morning and evening, so each block has enough clicks to judge.
4. For each block, work out clicks, spend, orders, conversion rate, CPC and cost per order.
5. Compare each block's cost per order with your target: price × target ACoS, where ACoS is ad spend ÷ ad sales. Your break-even is profit per order before ad spend, which the Amazon profit calculator works out after fees and GST.

Judge blocks on cost per order, not clicks. Clicks show where the traffic is, and an hour with many clicks and few orders is where money leaks.

Worked example. A product sells for ₹799 and leaves ₹240 per order before ad spend. You aim to spend ₹160 per order, a 20% target ACoS. Four weeks of one campaign's data, with the last 7 days left out:

In this example, the evening converts best, but its clicks cost ₹9 against ₹8 in the day, so an evening order costs about the same as a morning one. Overnight stands out at ₹350 per order against a ₹160 target. Whether that gap is real depends on how much data sits behind it.

How much hourly data you need before you act

A block needs several hundred clicks before its conversion rate means much. A rate measured on a sample of clicks carries a margin of error of roughly:

Margin ≈ ±2 × √(conversion rate × (1 − conversion rate) ÷ clicks)

At an 8% conversion rate, that is about ±5.4 percentage points on 100 clicks, ±2.7 on 400 and ±1.4 on 1,600, so four times the clicks halves the doubt. Act on a block only when its whole range sits on one side of your target.

In the worked example, overnight's 2.0% on 300 clicks could be anywhere from 0.4% to 3.6%. Even at 3.6%, a ₹7 click gives a cost per order of ₹194, above the ₹160 target, so that gap is real. Late night's 5.0% on 300 clicks could be 2.5% to 7.5%, a cost per order anywhere from ₹107 to ₹320, so leave it alone.

Small campaigns can't support hour-by-hour decisions. Say a campaign gets 100 clicks a day: four weeks give 2,800 clicks, about 17 for each of the 168 hours in a week, a margin of about ±13 points at an 8% conversion rate. Pool hours into blocks across all days instead.

The attribution trap: orders arrive after the click hour

Don't judge an hour on the orders that came in the same day. Amazon credits an order to the day of the ad click that led to it, if the purchase happens within the attribution window: 7 days for Sponsored Products bought by sellers, and 14 days for vendors, Sponsored Brands and Sponsored Display. Two things follow:

• Recent hours look worse than they are. For example, a shopper who clicks at 10 PM and buys the next morning adds nothing to 10 PM until that order is credited. The later the hour and the more recent the day, the more orders are still to come.
• Counting orders by purchase hour moves the credit. Some spreadsheets and tools match orders to the hour they were placed. The morning then collects credit for last night's clicks, and the evening looks like waste.

You may have read that Indian parents shop around 7 AM, before the day starts. Treat that as a hypothesis for your account, not a fact. If your 06:00 to 10:00 block beats your average on data that credits orders to the click, you have evidence. If the peak appears only when orders are counted by purchase hour, some of those morning orders are last night's clicks being completed.

Guardrails for hourly bid adjustments on Amazon

Limit how far and how fast a schedule can move bids, because it keeps acting every day whether or not anyone is watching:

1. Fix budgets first. A campaign that runs out of budget every evening is already dayparted by accident. Raise its budget or trim weak bids before adding a schedule.
2. Cut bids before you pause. Pausing loses every order in those hours, including the profitable ones, while a lower bid can keep some of them.
3. Move half the gap at a time. The bid change your data implies is roughly (target cost per order × conversion rate ÷ current CPC) − 1. Make half of that change, then re-measure on fresh data, because a lower bid can also cost you placements and change your conversion rate.
4. Don't switch off whole days without evidence. Four weeks give you four Mondays, and one odd Monday, such as a price change or a stock-out, moves that average a lot.
5. Test on a subset first. Apply the schedule to some campaigns and leave similar ones unchanged over the same weeks. Decide in advance what counts as success, for example lower cost per order with total orders holding steady.
6. Switch schedules off for sale events. A schedule built on normal days doesn't describe sale days. Our sale-event ads plan covers those.
7. Check what stacks. Dynamic bidding and placement adjustments work from your scheduled bid, and another automation rule can overwrite it. If a tool runs the schedule, confirm it restores each original bid when a window ends and warns you if a restore fails.

An example dayparting schedule

Worked example, continued. Applied to the table above, the guardrails give two changes and leave everything else alone. It shows the method, not a pattern to copy.

Re-measure once two more weeks of clicks are older than the attribution window, and take the next step only if the gap is still there.

When Amazon Ads dayparting isn't worth it

Dayparting only moves money between hours, so its gain is capped by what you spend in your weak hours. In the worked example, switching the overnight block off entirely would save at most ₹2,100 of ₹24,500, or 8.6% of spend, and lose 6 of 194 orders. Skip it, or wait, when:

• Your campaigns get too few clicks for any block's range to clear your target.
• Cost per order is about the same in every block.
• Cost per order is too high in every hour. That is a bid, search-term or listing problem, so start with why ACoS runs high and how to fix it.

ScaleSKUs builds this hour-by-day view from Amazon Marketing Stream data and leaves out the most recent 14 days, so late orders are counted before any hour is judged. Nothing changes on Amazon without your approval, and live changes need two-factor verification.

Frequently asked questions

Does Amazon have built-in dayparting?

Partly. Budget rules can raise a campaign's budget on a schedule and dynamic bidding adjusts bids auction by auction, but no budget rule lowers anything. Cutting bids in weak hours on a timer takes the Amazon Ads API or a tool built on it, so check your console for newer options before you buy one.

What is the best time to run Amazon ads in India?

The hours where your own cost per order is lowest against your target. We don't know of a time that holds for every category, so test any rule of thumb, including the 7 AM idea, on your own data before it moves a bid.

Should I pause my Amazon ads overnight?

Only if overnight cost per order is above break-even with enough clicks to be sure, and even then cut bids first. Pausing loses every overnight order, including the profitable ones, and saves at most what you spend in those hours.