1. Function Overview
You can set monitoring and alarm rules to select various data metrics of different dimensions such as app, ad format, placement, including revenue, impression, eCPM, etc. to monitor data fluctuations. We will provide you with regular data feedback via site letters and emails every day to help you find data problems in time.
2. Operation Guide
2.1 Add monitoring and alarm
You need to add monitoring rules first and select the dimensions and metrics for which data fluctuations need to be monitored.
(1) Enter the monitoring and alarm function
(2) Add rule
Choose different monitor types based on different scenarios:
- Non-AB test: You can select different dimensions, specify application, placement, etc., and monitor specific indicators in the corresponding dimensions.
- AB test: If you are conducting an AB test, you can monitor the fluctuation of the comparison effect of specific indicators for the AB test that is taking effect (by default, the group with the largest and smallest AB test ID will be selected for data comparison). You need to select the application or the corresponding AB test on the page for monitoring.
(3) Related data description
Click [+Rule] to set the dimensions and metrics you need to monitor. For specific metric descriptions, please see the table below:
Merits name | Optional options | Merits meaning | Description and suggestions |
Monitor Type | Non-AB test/AB test |
It is used to distinguish the content of monitoring. |
If you are not using AB testing and only pay attention to daily data fluctuations, please select the [Non-AB Test] |
Monitor Dimension | Account, App, Placement, Ad source, Network Firm, Ad format | Data metrics are aggregated according to the selected dimension. For example, if the estimated revenue metric of the account dimension is selected, the total estimated revenue of the account is counted | Different metric contents can be monitored through different dimensions. One rule only supports a single dimension |
Data Time | Cannot be selected. Only the data monitoring of the most recent day is supported | Background data is impressioned one day later. The nearest day to January 2 is January 1, which means that January 2 will receive an early warning message for abnormal data on January 1 | |
Trigger Conditions | Meet any of the following conditions, meet all of the following conditions | When setting multiple conditions, you can set whether the set metrics need to be met at the same time | If you need to set the monitoring data to rise and fall at the same time, you need to set any condition of rising or falling. You can see the following examples |
Trigger Condition-Data Indicator | DAU, DEU, Estimate Revenue, ARPDAU, eCPM, Imp. / DAU, Imp. / DEU, Engaged Rate, Requests, etc. | The meaning of the corresponding metrics can be viewed:Full report metric description | The provided metrics are consistent with the metrics of the report. The set threshold can be determined by the data corresponding to the comprehensive report |
Trigger Condition-Contrast Type | Previous cycle, The day before yesterday, Specific value |
Compare the most recent day with earlier data or specific values - Previous cycle: the same day of the previous week - The day before yesterday: the day before the data - Specific value: set a specific value to monitor (can be used for eCPM, impression rate, Gap) |
|
Trigger Condition-Value Type | Value, Percentage |
Supports calculation of metrics according to numerical values or percentages | |
Monitoring Cycle | Different time zones, half an hour as dimension selection |
When the set monitoring period time point is reached, the monitoring metrics will be calculated and compared. If the trigger conditions are met, the corresponding notification will be sent after calculation | The set time is the calculation time, not the notification time |
Notification Method | Site notification, email |
You can choose the corresponding notification method to receive monitoring warning content | Add multiple email addresses separated by English commas |
(4) Example
You need to view the revenue fluctuations of each app under the account. You can set the monitoring app dimension and estimate the revenue index to rise or fall by 20% compared with the previous day. The specific configuration is as follows.
(5) Setting the monitoring cycle time for each metric
① TopOn statistics (data without API suffix): Our statistics will be completed before 9:00 Beijing time every day. The monitoring time for this type of data can be set to 9:00 and later
② API data: The update time of API data on each platform is inconsistent. You can click [Suggestions on monitoring time of each platform] to view
After adding the rule, in the set monitoring cycle, we will calculate according to the set rules every day. After the data triggers the alarm, you will receive a notification on the site or email notification, and you can view the details of the hit rule in the monitoring alarm.
2.2 View monitoring alarm
After receiving the alarm notification, you can click the notification to view all the data abnormality reminders that triggered the alarm on that day, and promptly discover data problems in various dimensions.
(1) Site notification, Email
After selecting the notification method, you will see the corresponding monitoring metric data in the site notification/email. Click to jump to the monitoring and early warning center to view the corresponding content.
(2) View monitoring and alarm content in the background
On the monitoring warning page, you can view the number of matched rules for each rule. Click [View] to view the monitoring indicator data of matched rules in the corresponding dimension.
Viewing abnormal report data is not supported. You can view the data in the Full Report.
Note:
Regarding the monitoring of the AB test, the data of the latest day can be viewed in the [Test Results] data of the corresponding indicator in the AB test page and the ad slot data summary report.
3. FAQ
(1) When will notification start after setting a new rule?
If the warning time of the rule is set after the current time, the warning will take effect immediately on that day, otherwise it will take effect on the next day. Example: If the daily warning is set at 10:00 on 2022/1/1 9:00, the first warning notification will be received at 2022/1/1 10:00; if the daily warning is set at 9:00 on 2022/1/1 10:00, the first warning notification will be received the next day, that is, 2022/1/2 9:00.
(2) Set email notification but can't receive email?
You can first check whether there are emails in the trash, because sending emails too frequently may cause this problem. If there is no notification, please contact your operations colleagues.
(3) When monitoring API data, sometimes there is no warning, how to troubleshoot?
Because the update time of the monitored API data is inconsistent on each platform, the warning will not work properly when there is no data. Please set the monitoring time reasonably according to the [Suggestions on monitoring time of each platform] in the background prompt.