Most Amazon reports show a day as one number. That number hides a lot. Inside a single day, some hours buy clicks that turn into orders and some buy clicks that do not, at the same bid. Hourly data is where that difference shows up, and it is the starting point for any scheduling decision.
This playbook covers how to pull hourly data, clean it, read it in the right order, and turn what you see into a schedule. It assumes no special tools beyond a spreadsheet.
Step 1: pull the right data
You need spend, clicks, orders and sales by hour, for at least four weeks. Eight is better for accounts with modest volume. Break it out by day of week if you can, and by campaign if you want to schedule campaigns differently.
Hourly data comes from Amazon's advertising data and reporting tools or from third-party software that pulls it. The standard daily reports in the console do not include it. Before anything else, confirm the time zone. Hours should be in the marketplace's time zone. A pattern read in the wrong zone will shift every conclusion by several hours. Time zones and ad scheduling covers this in detail.
Step 2: clean it
Remove event days. Prime Day, Black Friday, Cyber Monday and similar events have their own hourly shape, with higher conversion and higher costs. Leave them in and they distort every hour they touch.
Remove days with obvious problems: a listing that was suppressed, a stockout, a day the budget ran out at 9am. Those days tell you about the problem, not about the hour.
Note late attribution. Orders can be attributed to a click days later. Leave the most recent few days out of the analysis, or wait until they settle.
Step 3: compute conversion rate by hour and check sample size
Add up clicks and orders for each hour across all the days in your range. Divide orders by clicks to get each hour's conversion rate. Then compute the account's overall conversion rate across all hours.
Put the two side by side. Every hour is now either above, near or below average. This comparison is the core of the whole exercise. The hourly heatmap guide shows a visual way to lay it out.
An hour with a handful of clicks and no orders has not proven anything. Before you trust an hour's conversion rate, check how many clicks it is based on. A rough guide: if your conversion rate is around 10 percent, an hour needs a few dozen clicks before zero orders means much. Lower conversion rates need more clicks.
Hours with too few clicks go in an "unclear" pile. Do not schedule around them yet. Collect more data.
Step 4: add cost per click
Some hours are weak and cheap. Some are weak and expensive. The second kind costs more. Compute cost per click by hour and look for hours where conversion is below average and CPC is at or above average. Those are the most expensive hours in your account per order.
A simple combined view is cost per order by hour: spend divided by orders. It folds conversion and CPC into one number you can compare to your target.
Step 5: split weekdays from weekends
Many accounts have different patterns on weekends. Mornings may start later; evenings may run longer; some B2B categories go quiet entirely. If you have enough data, repeat the analysis for weekdays and weekends separately. Weekend vs weekday PPC covers what usually changes.
Step 6: read the shape
Now look at the whole day. Most accounts fall into a few common shapes.
Evening peak. Conversion climbs through the afternoon, peaks in the evening, and drops sharply overnight. Common in consumer and gift categories.
Business hours. Conversion is strong from mid-morning to late afternoon on weekdays and weak at night and on weekends. Common in B2B and office categories.
Flat. Conversion varies little by hour. Common in some consumables and everyday essentials. Scheduling helps least here.
Illustrative example: Parkway Home's data shows conversion near or above average from 10am to 10pm, a steady fall after that, and very few orders from around 1am to 6am despite steady spend. That is an evening-peak shape with a clear overnight trough.
Step 7: turn it into a schedule
Hours far below average, with enough clicks to trust: pause, or cut bids sharply.
Hours somewhat below average: lower bids.
Hours at or above average: leave alone, or raise bids if budget allows.
Start conservatively. Pause only the clearest dead hours first, run it for two to four weeks, and compare. Building a schedule from data covers this step end to end.
Then re-read it each season. Hourly patterns shift. Summer evenings run later than winter ones. Q4 lifts conversion across more of the day. Re-pull the data at least once a quarter and adjust.
One last habit: keep the spreadsheet. Each time you re-pull the data, add a tab rather than overwriting the last one. After a year you have four snapshots of the same account, and the changes between them tell you more about your shoppers than any single pull. They also show whether a schedule you set months ago still matches how the account behaves today.
Frequently asked questions
Where do I find hourly data for Amazon ads?
Hourly performance is available through Amazon's advertising reporting and data tools, and many third-party tools pull it for you. The console's standard daily reports do not show it. Whichever source you use, make sure the hours are in the time zone of the marketplace, not your own.
How much hourly data do I need before making changes?
At least four weeks, so each weekday appears several times. Accounts with low click volume may need eight weeks before the overnight hours have enough clicks to judge. Exclude major event days, which distort the normal pattern.
What is the most important metric in hourly data?
Conversion rate by hour, compared to the account's daily average. Spend and clicks show where money goes. Conversion rate shows whether those clicks become orders. Cost per click by hour matters second, because some hours are expensive as well as weak.
Off Hours turns the schedule you read from your hourly data into dayparting rules that run on a 15-minute cadence, with every change logged. Start a free 14-day trial.