Reports

The Reports section of the dashboard provides in-depth analysis of your use of Engage. These standard reports explain aggregate user engagement activity, response rates, and important high-level statistics such as opt-in/opt-out and uninstall levels by platform.

Choose your app from the Urban Airship dashboard, click the Reports dropdown menu, and make your selection. Each report type is detailed in Report Types.

The next section, Overview, describes the features common to all of the aggregate reports. The appendices at the bottom of this page provides detail about our proprietary Push Influence model.

To learn about message-level reporting, see the following sections in the relevant composer documentation:

Overview

The top of a report displays its type, definition, and a Date Range Filter. Some reports also have a question mark icon to the right of the name. Click the icon to display help content specific to that report.

If the report can be exported, a Download CSV button is to the right of the Date Range Filter.

If the report can be exported and printed, the Export icon is in the top right corner of the report’s chart. Click the icon and choose either Print or Export CSV.

Note

Statistics is the only report that does not have a Date Range Filter or export option.

Hover over a chart’s data point to display its value. Click the data point to narrow the displayed date range.

Date Range Filter

Click the dropdown menu to change the date range. The report updates after each change.

Date Range Data Displayed
Last 7 Days Dates for today and the previous 7 days will display on the horizontal axis, with data shown for each day.
Last 2 Weeks Dates for today and the previous 14 days will display on the horizontal axis, with data shown for each day.
Last 30 Days Dates for today and the previous 30 days, in increments of 4 days, will display on the horizontal axis, with data displayed for each day by clicking the data point in the graph.
Last Month Dates for the current month, including today, will display on the horizontal axis, with data shown for today and each day back to the beginning of the current month.
Last 3 Months The current and 3 previous months will display on the horizontal axis, with data shown by month.
Last 12 Months The current and 12 previous months will display on the horizontal axis, with data shown by month.
Today Based on the time you entered the request, hourly increments will display on the horizontal axis, ending at the previous hour and going back to midnight.
Yesterday Based on the time you entered the request, hourly increments will display on the horizontal axis, ending at the previous hour and going back to midnight the previous day.
Custom Range See Custom Range below.

Note

  • To view the narrowest time frame possible, select Today in the Date Range Filter. Depending on when you make the request, you will see data for a period of 24 or fewer hours.
  • Due to time zone differences, the date range displayed at the top of the chart may not correspond exactly with your selected date range.

Custom Range

After selecting Custom Range, dates will appear to the left of the dropdown menu. Click each date and make your selection from the calendar that appears. The report updates after you make each date selection.

Note that the time frame shown in the report chart is based on the time you entered the request.

Custom Range Data Displayed
Same dates in ‘From’ and ‘To’ If you selected a day other than today, hourly increments from midnight to 11 PM will display on the horizontal axis. If you selected today, hourly increments will display on the horizontal axis, ending at the previous hour and going back to midnight.
Yesterday’s date in ‘From’ and today’s date in ‘To’ 4-hour increments will display on the horizontal axis, beginning at midnight of the previous day and ending at the previous hour, with data displayed for each day by clicking the data point in the graph.
A 3-day period with a date previous to today in ‘To’ Beginning at midnight on the date in the ‘From’ field of the selected date range, the time will display on the horizontal axis in 4-hour increments, ending at the previous hour, with data displayed for each day by clicking the data point in the graph.
A 3-day period with a date previous to today in ‘To’ Beginning at midnight on the date in the ‘From’ field of the selected date range, the time will display on the horizontal axis in 4-hour increments, ending at midnight on the date in the ‘To’ field.
A 4-day or longer time period ending with today’s date in ‘To’ In 4-day or longer increments, data will display by the day.

Report Types

Each report provides a different piece of analytics information.

  • Push Sends: How many pushes are you sending?
  • App Opens: How many users are opening your app as a result of your push?
  • Time in App: How much time are people spending in your app?
  • Unique App Opens: How does the app activity of opt-in vs. opt-out users compare?
  • Push Response: What is the level of direct and indirect response to your pushes?
  • Event Tracking: How have specific push messaging campaigns affected user activity and key conversions?
  • Devices: What percentage of devices are eligible to receive push?
  • Statistics: What is your hourly, monthly, and lifetime push activity?

Push Sends

Push Sends assesses the total notification sends during a specified time period. The default view presents the daily total of push sends over the last seven days, with total, average, high, and low values below.

Switches for Sends, App Opens, and Time In App are in the upper left corner of the chart. Additional selected data is superimposed over currently displayed data. Toggle data selections by using the switches, or click the switch names below the chart. Time In App is measured in seconds.

Note

Frequently, increasing push sends will decrease the Time In App metric. Sending pushes increases the number of app opens, but those opening the app via a push may spend a very short amount of time in the app. Because Time In App is calculated as an average, the increased number of users spending short periods of time in the app drives the metric lower.

Warning

The report tracks notifications rather than total pushes sent, so rich messages are included in the report only if:

  1. The users receiving the messages have opted-in to push notifications, and
  2. The messages include a notification, e.g. a badge, sound, alert text, etc.

To view the total number of rich messages delivered, see Message Reports.

App Opens

App Opens assesses the total app opens during a specified time period. The default view presents the daily total of app opens over the last seven days, with total, average, high, and low values below.

Switches for App Opens and Sends are in the upper left corner of the chart. Additional selected data is superimposed over currently displayed data. Toggle data selections by using the switches, or click the switch names below the chart.

Note

The App Opens report includes repeated app opens by individual users. To see the number of unique users that have opened your app, see the Unique App Opens report.

Time In App

Time in App assesses the average time users spend in your app during a specified time period. The default view presents the daily average session time over the last seven days, with average, high, and low values below.

This data is collected from user session events, and it is measured in seconds. Though behavior varies by operating system, we calculate this from events received each time your app is opened, brought to the foreground, or sent to the background.

Switches for Time In App and Sends are in the upper left corner of the chart. Additional selected data is superimposed over currently displayed data. Toggle data selections by using the switches, or click the switch names below the chart.

View Time In App with Sends to see the influence your push may have had with users’ session length. In the chart below we can see that increasing push sends decreases the Time In App metric.

Note

Frequently, increasing push sends will decrease the Time In App metric. Sending pushes increases the number of app opens, but those opening the app via a push may spend a very short amount of time in the app. Because Time In App is calculated as an average, the increased number of users spending short periods of time in the app drives the metric lower.

Unique App Opens

Unique App Opens assesses unique app opens by opt-in status and platform during a specified time period. Unique App Opens is the count of users that opened the app, not the number of times a single user opened the app.

The default view displays Opt-in Opens and Opt-out Opens for all platforms over the last seven days, with total, average, high, and low values below. Percentage and Trend charts follow.

Note

Total is the sum of the displayed bars rather than the total number of unique users who opened the app in the defined time period. For example, if you are displaying seven days of data, a given user will appear only once per day but might be counted up to seven times in the total.

Note

This report does not include information about the number of new users that have opted in or out of push notifications. Instead see the Devices report.

To filter by platform, select iOS Only or Android Only from the dropdown menu in the upper left corner of the chart. If your app only supports one platform, the platforms dropdown menu is not displayed.

Toggle displayed data by clicking Opt-in Opens, Opt-out Opens, and Sends per platform below the chart.

Note

The unit of time on the bottom of the chart determines the frequency with which a unique user’s app opens are counted. In the image directly above, a user who opened the app 10 times in March would contribute exactly one unique open for that month. For the daily active use image, a user who opens the app three times on June 25th and four times on June 28th would contribute one unique open for each of those days.

Percentage and Trend

  • The Percentage pie chart displays the comparative percentage of Opt-in Opens and Opt-out Opens per platform, and updates when the data in the chart above changes. Exact percentages are shown on the right.
  • The Trend line graph displays the trending percentage of unique user app opens per platform for the selected time period.

Both are color-coded, corresponding with the chart above. Toggle displayed data by clicking Opt-in Opens and Opt-out Opens per platform.

The Percentage and Trend panes can be printed or downloaded. Click the respective icons in the upper right corner of the pane. Download options are PNG or JPEG image, PDF document, or SVG vector image.

Warning

Additional care must be taken when interpreting the pie chart, as it only counts users who opened your app in the displayed time range. This may not be an accurate representation of the percentages of your install base as a whole.

Daily and Monthly Active Users

The Unique App Opens report may be used to determined the opt-in and opt-out breakdown of your daily and monthly active users.

Select the desired range from the date range filter. By default, the chart is set to the last 7 days, providing a view of daily active users. To see monthly active users, use the Custom Range option.

The report below uses the dates March 1 to May 31. The three pairs of bars correspond to unique app opens for the months of March, April, and May.

Note

The unit of time at the bottom of the chart determines the frequency with which a unique user’s app opens are counted.

When viewing daily active users (7 days’ activity), a user who opens the app three times on a single day and an additional four times on a different day would contribute one unique open for each of those days.

When viewing a monthly active users, a user who opened the app 10 times would contribute exactly one unique open for that month.

Push Response

Push Response assesses the impact of notification sends on app opens, with data displayed per platform.

See Appendix: Push Influence Primer and Appendix: Push Response Terminology for supporting information.

Event Tracking

Event Tracking is an aggregate view of Custom Events during a specified time period.

  • Event Name: Human-readable name assigned to a particular Custom Event
  • Notification Attribution: Displays whether your event was directly or indirectly attributed to the push notification
  • Location: The source from which the event originates, most commonly one of Interactive Notification, Message Center, Landing Page, or Custom.
  • Count: Number of instances of this event
  • Value: The value (monetary or otherwise) generated by the event
  • Avg. Value: Value divided by Count
  • % of Total Count: Count divided by the Total Count (listed on the bottom row)

Note

Only events that have been assigned a value within your app project will generate values for the aggregate view. Event values are assigned by you when creating events, and should align with your campaign strategy when considering the respective weights of different activities you seek to measure. The aggregate view is only meaningful when the unit of measure for these activities is the same. See: Custom Events for details on Custom Events setup and strategy.

Note

Using Notification Buttons in your message automatically creates Custom Events, where one event is assigned to each button. When viewing the associated Message Reports or the Event Tracking report, you will see a custom event listing for any button that has been pressed.

Devices

Note

iOS and Android Support only. Analytics must be enabled in the SDK. See instructions in our iOS and Android documentation.

Devices provides a current view of your audience by the number of devices opted-in and -out of notifications. The report displays the total number of devices on which your app has been opened at least once and have not uninstalled, and a breakdown of opt-in and opt-out status for iOS and Android. If your app only supports one platform, Combined Unique Devices is not displayed.

When an opted-in user uninstalls your app and you send a notification to that user’s device, the device is removed from the total device count in the report and placed in the uninstall count. You can use the report to determine the breakdown of your audience by opt-in or opt-out status. See the Unique App Opens report for information related to the effect notifications have on your users.

  • Unique Devices: Total number of devices that have opted in or opted out of push notifications. The Unique Devices count does not include uninstalled devices. If an opted-in device uninstalls your app, and we receive that information, it is removed from the Unique Devices count and placed in the Uninstalled count. See Detecting Uninstalled Devices.
  • Opted-in: Devices with notification permissions on.
  • Opted-out: Devices with notification permissions off.
  • Uninstalled: Android: Number of app uninstalls observed from previously opted-in devices.
  • Uninstalled: iOS: Number of app uninstalls observed from either opted-in devices or devices with background app refresh enabled.

Upon an app open and when Analytics events are not disabled in the app, our SDK will try to send an app open event to our servers along with the opt-in/out status. Once we receive the event, we record the opt-in/out status and generate the Devices report daily. See Detecting Uninstalled Devices for information about the nuances related to the uninstalled devices metric.

Note

The total devices count in Message Reports may differ from the total devices count listed in the Devices report. This may be due to but is not limited to:

  • iOS 7 not giving device tokens to opted-out devices (result is lowered send count)
  • Additional platforms supported beyond iOS and Android
  • Devices registered via the server
  • Timing of when the Devices report is generated and when a user opens your app for the first time or uninstalls your app

Tip

Use the CSV Download to save daily snapshots and perform custom comparison. For example, if you’d like to see how many people have opted in to your app for the first full week in January, download the January 5th and January 11th reports, open the files in a spreadsheet app, and calculate the difference in device counts.

Detecting Uninstalled Devices

If a user uninstalls your app the next time you send a push to that device, we will receive feedback from the platform push service. There are two ways to ping devices to check for uninstalls:

  1. Via a standard push message (iOS & Android)
  2. Via a background push, using the content-available flag (iOS7+ only)

Note

See iOS Overrides for information about using the content-available flag in the API.

In the UI, enable Background Processing to use the content-available flag. See the iOS option in the Delivery section of the Message, Automation, and A/B Tests composers.

For option one, after sending a push to a device that has uninstalled your app, we receive feedback from GCM or APNs that the device is inactive. When we generate the daily report, we will remove this device from the device total and opt-in/out status breakdown, and the uninstalled device will then appear in the uninstall count. This method of checking for uninstalls requires that the device had push enabled, as it is dependent on sending a standard push message.

With iOS devices, you also have the option of pinging devices using background push. Consequently, the iOS uninstall count can include data on opted-out devices with background push enabled as well as devices that were opted-in to push, while the Android uninstall count will only include data on devices that were opted-in to push. To successfully achieve a background push that pings for uninstalls, you must first enable background push in your application. Next, you simply send an empty message with the Content Available flag enabled:

Technically, users can prevent you from sending them background pushes by disabling the Background App Refresh option in their phone settings. However, because this option is enabled by default, we predict very few users will choose to disable it.

Regardless of platform, if a user re-installs the app and opens it, they are moved out of the uninstall count and placed back in to the opted-in/out count.

Note

Uninstall information is only updated after a push notification is sent to an opted-in device (or an iOS device with Background App Refresh enabled) that has uninstalled. A Rich Message without a push notification would not receive uninstall feedback.

Statistics

Statistics lists notification sends counts.

  • Current Audience: View current audience takes you to the Devices report. View billing audience takes you to your Account Info page.
  • Push Sends (month): The total number of notifications sent during the current calendar month.
  • Push Sends (all time): The total number of notifications sent over the app’s lifetime.

Sends By Platform

The Sends by platform chart displays a breakdown of push activity for each platform over a defined period. The default view is of the past 24 hours.

To define a different period, enter dates and times in the Select Date Range fields, then click the Filter Dates button. Note that times are in UTC.

For the above chart, the yellow line represents pushes sent to iOS, and the blue line represents pushes sent to Android. The table lists the number of pushes per platform and in total for the given period.

Appendix: Push Influence Primer

A number of technical and user-behavior factors make it difficult to attribute an app open to a push notification with 100% accuracy. As the industry’s first and largest commercial push notification provider, Urban Airship has both the experience and the data to recognize and model attribution across all app types with a high degree of confidence. The result is our Push Influence algorithm which informs the Push Response report.

As in the graph below, we see that App Opens rise dramatically after a push notification is sent.

App Opens vs. Pushes Sent

Direct Opens help to indicate the effectiveness of a push notification, but they do not tell the whole story. In the image below, Direct Opens represent only a portion of the spike in App Opens centered around the push notification.

As such, Direct Opens understate the true impact of a push notification. A significant number of additional opens will occur in the wake of a push, but in the absence of an interaction with the notification itself, attribution is complicated. Understanding that users are often occupied when the notification arrives and will return to the app later, we have a higher degree of confidence that a push influenced an open the sooner the open occurs relative to the push.

Indirect Open: when a user is sent a push, but DOES NOT tap it. Instead, it reminds them to open the app later, at a more convenient time.

To complicate matters, we must deal with the fact that some of the observed App Opens would likely have occurred anyway, even in the absence of a push notification. Every app has as natural amount of organic opens which will vary depending on app category, popularity, and other factors. We refer to this natural rate of App Opens as the Baseline Open Rate and calculate this rate for each app when determining Push Influence figures.

With the context of Direct Opens, Indirect Opens, and Baseline Opens, we have a framework for understanding the true impact of push notifications on App Opens. We call this derived metric Influenced Opens.

Influenced Opens represent the number of opens that occur both directly and indirectly as a result of a push notification, less the baseline organic opens that would be expected anyway. The following four bullets describe the basic formula for arriving at the Influenced Opens figure:

  • Influenced Opens include Direct Opens and Indirect Opens by opt-in users.

  • Only opens within a 12-hour window* after the push is sent are included in the calculation.

  • Baseline Opens are excluded from the Push Influence calculation.

  • The Push Influence algorithm is more likely to attribute an open as “Influenced” the closer it occurs to the push.

    * The 12-hour window for App Open attribution has proven to be the most meaningful to Push Influence, when consistently applied across tens of thousands of apps in our database. Beyond 12 hours, we cannot maintain a high degree of confidence in the statistical significance of Indirect Opens vs. Baseline Opens.

Influenced Opens

Indirect Opens occur when a user is sent a push but does not tap the alert. When the user opens the app indirectly, not from tapping the notification, then attribution must be derived.

Following a push notification, we typically observe a flurry of new opens above the Baseline Open Rate for a period of time. These “Influenced Opens” include both Direct Opens and Indirect Opens by opt-in users.

The Push Influence algorithm attributes App Opens that occur within a 12-hour window of the push to this calculated metric, accounting for the Baseline Open Rate.

The further an indirect opt-in open occurs from the original send time, the less likely it is to be considered an Influenced Open.

Appendix: Push Response Terminology

Opt-in Open
Any app open observed by a user who has opted in to receive push notifications. The Urban Airship SDK observes open events for all users and reports both total opens and opt-in opens.
Direct Open
An app open that occurs when a user interacts with a notification to open an app, e.g., taps a push notification in the notification center, lock screen, etc.
Indirect Open
An app open that is attributed to the presence of a push notification, but is not measured directly, i.e., the user does not tap the notification directly. Indirect opens are a derived metric which is explained more fully in Appendix: Push Influence Primer.
Influenced Open
The calculated number of App Opens attributed to a push notification, having occurred within a 12-hour window of the push being sent. This includes both Direct and Indirect Opens, and the algorithm subtracts expected hourly opens that are likely to have happened without the push, and also takes less credit for opens the further they occur from the push.
Baseline Open
App Opens that occur regularly, irrespective of push notifications being sent. Baseline Opens might also be considered a “natural” or “organic” open in that they are expected to occur even in the absence of push notifications. Baseline Opens are derived from historical open trends on an app-by-app basis, and are excluded from Push Influence attribution.