When you're viewing data as detailed as hourly reporting, you may see inconsistencies between the hourly breakdown and daily, weekly, or longer time periods. This is expected due to differences in engine accounts, time zones, general accuracy of hourly data, and other reasons.
Availability of hourly data from different engines
Hourly data is available in all supported search engines except Yahoo! Gemini. The data is available at the advertiser, engine account, and campaign level. All supported engines except Baidu also provide hourly performance data at the ad group level.
If you scope to an advertiser and view hourly data for all ad groups in the advertiser, all metrics for Baidu will be excluded from the report. If you change from hourly to daily or longer, Baidu metrics will be included in the report.
Historical hourly data
The historical window for hourly data depends on the type of engine account:
- All supported engines except Yahoo! Gemini and Baidu: You can use Search Ads 360 to view hourly data as far back as March 01, 2015.
- Baidu accounts that performed their first sync with Search Ads 360 before December 18, 2015:
You can use Search Ads 360 to view hourly data as far back as March 01, 2015.
- Baidu accounts that performed their first sync with Search Ads 360 after December 18, 2015:
You can use Search Ads 360 to view hourly data as far back as 30 days prior to the first sync.
For example, if your first sync with Search Ads 360 is February 1, 2016, you can see hourly data as far back as January 3, 2016.
On December 18, 2015, Baidu changed its policy for the availability of historical hourly data.
If you set a report's time range to include a data earlier than March 01, 2015, Search Ads 360 hides the options for viewing an hourly report.
Accuracy of hourly data
Note that different types of engines provide metrics to Search Ads 360 at different rates. For example, some engines provide fresh metrics every 20 minutes while others may provide updates every few hours. If you're comparing hourly data across engine accounts for the current day, be aware that Search Ads 360 may not have the latest metrics from all engine accounts at the time that you view the data.
In addition, engines can make changes to their data for up to seven days. Most commonly, they may remove spam clicks or spam impressions. So you may see a change in hourly metrics as engines finalize their metrics.
Types of conversions in hourly reports
Hourly data is available only for conversions recorded by the following tracking systems:
Floodlight activities (both online and offline)
In the performance summary graph, you can set the hourly interval and view Transactions, but you can't view GA transactions or Baidu conversions.
On the Dimensions tab, you can click Hour of day and add Actions, Transactions, or Google Ads conversions columns, but you can't add Google Analytics columns or Baidu conversion columns.
No hourly data for...
Hourly data is not available for:
Google Analytics goals or transactions and Google Ads cross-environment conversions. That is, if you add these conversions to the performance summary graph when it shows the hourly interval, you won't see any data for the conversions. Likewise, if you view the Hour of day dimension and add columns with Google Analytics conversions or Google Ads cross-environment conversions, those columns will be empty (both in the UI and in downloads).
Depending on how you've set up time zones in your Search Ads 360 account and in each engine account, Floodlight activities might report conversions in one time zone, while the engine accounts report clicks and other metrics in a different time zone.
When this occurs, Search Ads 360 usually converts hourly Floodlight metrics to the time zone of the engine account so all metrics are reported in the same time zone. There is one exception to this behavior: if the Campaigns or Ad groups tab is not segmented by hour, then the performance summary graph uses the time zone specified by your Search Ads 360 account for Floodlight metrics and the engine account's time zone for engine metrics.
Depending on when conversions occur, and on how far apart the time zones are, you may see differences between hourly reports and reports with a daily or longer scope.
A Floodlight activity records a conversion at 12:30am Eastern time on May 1.
The conversion is attributed to a keyword in a Google Ads account that uses the Central time zone.
When you view an hourly report that includes the Google Ads account in the Central time zone, Search Ads 360 will report the conversion as occurring at 11:30 pm, Central time on April 31.
If you change the hourly dimension to a daily dimension, Search Ads 360 will report the conversion as occurring on May 1.
Different time zones at the advertiser scope
If you scope to the advertiser level, you may include engine accounts that use different time zones. Note that hourly reports will still report engine metrics and conversions in each engine account's time zone.
Engine account "A" uses the Eastern time zone, and reports a conversion occurring at 9pm.
Engine account "B" uses the Central time zone, and reports a conversion occurring at 8pm.
Even though both conversions occurred simultaneously, an hourly report will show the click from Engine account "A" as occurring in hour 21, and the click from Engine account "B" as occurring in hour 20.
Click time versus conversion time
Depending on which conversions you're tracking, you'd normally expect to see some amount of time between a customer clicking an ad and a customer converting. For example, it may take a customer a few minutes to fill out a form on your site, or it may take over an hour for the customer to make a decision and convert.
When you view an hourly report, conversions are attributed to the hour in which the conversion occurred (that is, the hour that the tracking tag fired), which is not necessarily the same hour during which an ad was clicked.To see the average lag time between clicks and conversions, click the Dimensions tab and select the Hour of day dimension. Then add the Actions avg lag (in sec) or Trans avg lag (in sec) column.
Spikes might not be statistically significant
When attempting to detect time-sensitive trends, you'll want to consider at least a few weeks of data. However, be aware that even over a period of a few weeks, competitors' actions, news events, weather, sales, and other types of events can either create a false appearance of a trend or can suppress the visibility of a significant trend.
The best way to detect trends is a period-over-period comparison that takes into account anomalies that may occur during each period.