Audiences guide

At a glance: AppsFlyer Audiences product provides a seamless and secure way to scale your remarketing programs. Use Audiences to build and connect audience segments to your partner networks, which update automatically daily.

Related articles: Audience partners | User Attributes | Import audience | Incrementality guide

Seamlessly segment users using attribution data points for precise audience creation
Enable cross-network re-engagement using a network of integrated partners
Securely manage your sensitive audience data from one central dashboard
Split the segment, keeping a portion in a control group held back from remarketing, and divide the remainder amongst the partners
Run incrementality testing to measure the true impact of your remarketing campaigns

Concepts

Audiences onboarding steps
	Step	Information	Procedures
1	Define the audience segment using rules	Audience segmentation rulesets Ruleset examples	Add an audience
2	Connect the audience to one or more partners. You will need partner credentials to do so.	Credentials required per partner	Add an audience
3	[Optional] Split the audience amongst partners/control group	Audiences split and control groups	Set split and control groups
Optional	Import an audience

Audience segmentation rulesets

An audience segment ruleset consists of a mandatory primary rule and additional secondary rules to enable precision filtering.
Combine rulesets using operators: and, or, exclude.
As rules are defined and layered, the estimated audience size displays in real-time.
Where an audience consists of multiple rulesets, use the overlap visualization display to better understand your audience segments.
The user identifier used to evaluate rules is either the advertising ID [default] or the customer user ID (CUID). The audience list, sent to networks, always contains advertising IDs but can optionally include CUID for cross-device segmentation.
The examples that follow illustrate ruleset implementation in various scenarios.

Ruleset examples

Example with revenue

Users who installed the app during the last seven days, and spent $1

Example: Location-based rule

Users from New York who uninstalled the app in the last month

Example: User-attributes based rule

Users who launched the app in the last 7 days at least once and whose favorite sport is cricket.

Remarketing example

Using clicks and views for effective remarketing. To convert these users, a little extra convincing or a gentle reminder is needed. These users have a high probability of converting, so remarketing to them, can produce significantly better results, than advertising to new segments. Create precise audiences for remarketing based on clicks and impressions data collected.

Build an audience of potential users that clicked on an ad but did not install in the last two days.

These are users who, in the recent past, were sufficiently interested in your mobile app to click on your ad, but for some reason did not install the app, or installed the app but did not launch it. Some users just need a reminder (push) to install it. Others may need to see more screens or features of your app to finally convert. Therefore, if you have a primary creative you use on new audiences, it might be better to use a different creative for audiences that clicked but did not install.

Primary rule

You must define one primary rule per ruleset.

Rule attribute	Options

App	Users of the App App is any app housed in the account. The team member defining the audience needs permission for all apps referenced.
Engagement type	Who have: Launched the app Viewed an ad Clicked an ad Installed the app Launched the app
Time frame	In the last [n] days (n=1-9999) Between [specified data range] At any time
Frequency	Used the app: At least [n] times Exactly [n] times At most [n] times [n]=1-100

Secondary rules

Secondary rules are optional
You can have multiple secondary rules
Each rule has a unique type, for example, engagement, in-app event, revenue
Rules have different parameters and settings, as listed in the table

Engagement

Engagement type:

Viewed an ad
Clicked an ad
Installed the app
Did not install the app
Launched the app
Did not launch the app
Uninstalled the app

Time frame and Frequency

In-app event

Did event [event name]
Did not do event: [event name]

Time frame and Frequency

Event attribute:

Is
Greater than
Less than

Notes relating to in-app events

The in-app event list updates Daily.
In-app events, having attributes of more than 100 unique values aren't available by default. For example, the in-app event product can have many more than 100 unique values. To enable such an in-app event, reach out to your CSM or send an email to hello@AppsFlyer.com
If the Did event is selected, and the in-app event has attributes associated with it, you can filter by in-app-event attribute value.

Revenue

Total revenue generated

Is
Greater than
Less than

Notes about revenue:

The value is in USD.
The value is the total value of all events reported from a device using the af_revenue parameter.

Location and metro code

Location
Metro code

Is
Is not

Location options

Country
State
City

Metro code options

Metro codes

Notes relating to location

If you select multiple countries, you can't select states.
If you select multiple states, you can select cities.

Device

Device brand
Device model
App version
OS version
Carrier

Is
Is not

Selected value

Attribution

Organic
Retargeting

False
True

User-attribute

User-attribute name

String type
- Is
- Is not
Number type
- Is
- Greater than
- Lower than

For string: Select a value

For number: Enter a value

Add an audience

To add an audience:

Go to Integration > Audiences.
Click Add Audience.
The Build tab displays.
Enter a unique Audience name.
Select the platform: Android or iOS.
[Optional] Select Advanced settings to select the user identifier:
- [Default] Advertising ID: evaluates rules by advertising ID only.
- Customer user ID: evaluates rules by grouping devices using the CUID. Use this method after you verify that your data has valid CUIDs.
Select New ruleset.
Build the audience guided by the user interface.
- For specific information about rule options, see primary rule, secondary rules, and combining rulesets in this article.
- Combine rulesets using the operators and, or, exclude.
Complete defining the audience, click Save audience and proceed.
The Connect tab opens.
Do one of the following:
- To create a new connection:
  1. Click Create new connection.
  2. Type a connection name.
  3. Select a partner.
  4. Complete the process according to the instructions in the user interface.
  5. Click Add connection.
- To select from an existing connection:
  1. Click Manage connections.
  2. Select one or more existing connections.
  3. Click Save changes.
    The Audience connection window opens.
  4. [Optional] Split an audience between partners and a control group
  5. [Optional] Copy split groups to S3.
Click Save changes.

Combine rulesets

To combine a ruleset with an existing audience

Go to Integration > Audiences.
Click Add Audience.
The Build tab displays.
Enter a unique Audience name.
Select the platform: Android or iOS.
[Optional] Select Advanced settings to select the user identifier:
- [Default] Advertising ID: evaluates rules by advertising ID only.
- Customer user ID: evaluates rules by grouping devices using the CUID. Use this method after you verify that your data has valid CUIDs.
Click Select from my audiences.
Select an operator: and, or, exclude.

To combine a ruleset with an imported audience:

1. Create the audience, for example, A.
2. Create the import audience, for example, Import_B.
3. Create a combined audience using Select from my Audiences. Combine A with Import_B using the operators.

Advanced concepts

User-identifier (advanced setting)

Audience rule evaluation is performed using an app user identifier. Select from:

[Default] Advertising ID: Ruleset compliance is assessed using the advertising ID. When compliant with the ruleset, the Advertising ID is added to the audience list. Note: Advertising ID means GAID, IDFA, or equivalent.
Customer User ID (CUID):
- Ruleset compliance is assessed by grouping advertising IDs (devices) associated with the same CUID together. If the group complies with the ruleset, then the advertising IDs comprising the group are added to the audience. Note: The CUID itself is not part of the audience list.
- CUID implementation considerations:
  - Ensure that your data is populated with valid CUIDs.
  - App users (devices) without a CUID are ignored.
  - When an advertising ID is matched to a CUID, historical audience data for the user is updated retroactively and included in audience builds going forward. Note: No change is made to raw data.
  - If an advertising ID is associated with two different CUIDs, the CUID most recently associated is used (wins.)

Example Advertising ID vs. CUID

Raw data for example
Advertising ID	CUID	Purchase $
A	99	1
A	99	2
B	77	1
B	77	5
C	99	3
D	None	50

Users who purchased at least $5
Rule evaluated by	Advertising IDs included in the audience	Explanation
Advertising ID	B, D, E	A and C are not included - the spend was less than $5
CUID	A, B, C	A and C have the same CUID (99) - the combined spend is $6. D is ignored - there is no CUID

Audience split and control groups

Feature

Description

Audience split

Audience split enables you to randomly split the audience amongst partners.
You can split the audience between 2-10 partners.
Allocate each partner a unique portion, in percent of the audience.
Split audiences for a variety of reasons like A/B testing.
You can modify split audience configuration but take into consideration that allocation modifications can impact the validity of A/B or other test measurements.

Control groups

A portion of the audience can be held back in a control group. The control group is not sent to any partner. Control groups are used in A/B testing.

Send a copy of split lists to an S3 bucket

CSV files containing the split audience lists and control group can be sent to an Amazon S3 bucket. They are sent as follows:

Directory structure:

 \YYYY-MM-DD\audienceID_audienceName_timestamp\

Filename: The name of the partner or control group. Note: If the connection to the partner fails, and the file is not uploaded, the parnter_name has the suffix failed. Example: partner_name_failed.csv

Dashboard

Use the dashboard to edit audience settings, view the audience upload status, manage connections to partners.

Manage audiences

To open the dashboard:

Go to Integration > Audiences.

The options available are listed in the table

Option	Description
Change audience settings	Click an Audience name, do one of the following: In the Build tab: view, and change rulesets. In the Connect tab: set audience split, and control group options.
Add audience	Build a new audience
Actions	View history: Displays the audience upload history including audience size, and split between partners. Upload now: Sends the audience to connected partners. Download as CSV: A file is generated per partner. The file is available after the audience has been calculated. It can take some time for this to happen after you create or modify an audience. Download files are limited to 5 million records. For larger files, send the file to an S3 bucket or use SFTP. Duplicate: Copy the audience definition. [Option] remove one or more of the connected partners during duplication. Note: Audience names do not have to be unique. When you duplicate an audience, give it a unique name. Calculate now: Available for imported audiences. Recalculates the number of users in an audience. Use as needed, to refresh the audience size, if it does not look correct. Delete the audience
Bulk actions	Carry out the same action on multiple audiences. To use bulk actions, select two or more audiences and a bulk action as follows: Upload now Show audience overlap: displays the audience overlap between the selected audiences Delete

Audience tab columns
Column	Description
Audience name
Size	Number of users in the audience
App name	The app referenced in audience selection. This is not necessarily the app being targeted
Upload to	Icons of partners to whom the audience is sent indicates that the audience is split between partners
Upload status	The current audience upload status In case of failure, click View details and take corrective action in accordance with the detailed error message that displays
Incrementality	[Displays if you subscribe to Incrementality] Click on a link to go to the Incrementality dashboard

Manage connections

The connections tab contains the partners and connection status for each connection.

To manage/add a connection

Go to Integration > Audiences > Connections.
Do one of the following:
- To add a new connection:
  1. Click Create new connection.
  2. Enter a Connection name.
  3. Select a Partner
  4. Complete the fields/login to partner process. Some partners have detailed instructions about how to connect.
  5. Click Add connection.
- To manage an existing connection:
  1. Select a partner.
  2. Select a connection.
    
    '
Do any of the following actions.

Option

Procedure

Delete connection

To delete a connection:

If there are any audiences connected, remove them as follows:
1. Select the audience.
2. The audience connects tab opens in a new tab.
3. Remove the connection.
4. Close the tab.
Repeat this procedure until no audiences are connected to the connection.
Click Delete connection.

Reconnect

To reconnect:

If the connection status is not active, click reconnect.
Update the connection credentials or login to the partner to reauthenticate according to the instruction in the user interface.

Change the connection name

To change the connection name:

In the Connection name field,
Click save changes.

Set split audience and control group

The following options are available:

Split the audience amongst the connected partners and optionally hold back a portion in a control group. Caution: Modifying an existing split allocation between partners causes redistribution of the entire audience. This can negatively impact the validity of A/B or incrementality test measurements.
Send a copy of the audience list to your S3 bucket Note: This is not the same as defining Amazon S3 as a connected partner. The audience lists can be downloaded to a CSV file. Use the CSV file, for example, to send to partners who are not integrated with AppsFlyer.
Create and add a new connection.
Add/Remove an existing connection.

Set split audience and control group

Go to Integration > Audiences.
In the Audiences tab, select the audience.
Go to the Connect tab.
The current settings display.
Do any of the following actions.

Option	Procedure
Split an audience between partners and control group	To split an audience amongst partners/enable a control group: Enable Split audience. The total of the control group and split to partners combined must be 100%. If you don't want to include the control group in the split: allocate 0%. All connected partners must be allocated a portion of the split (Don't allocate a partner 0%) To view the total allocated: scroll down; the total percent split displays. Enter the percentage allocated to: Control group. [Recommended] 15% or more. Enter a split percentage per connection. Scroll down to display the total allocation in percent, ensure that the total allocation is 100%. Click Save. Click Save changes.
Copy split groups to S3	To send a copy of the audience split list to an S3 bucket: Click Copy split groups to S3. Select a connection. Click Save changes.
Create and add a new connection	To create a new connection: Click Create connection. Type a connection name. Select a partner. Complete the process according to the instructions in the user interface. Click Add connection.
Add/remove an existing connection	To remove or add an existing connection: Click Manage connections. Select or clear select connections. Click Save changes.

Traits and Limitations

Traits

Audiences traits
Trait	Supported	Remarks
Ad networks	x
Agencies	x
Agency transparency	x
App-specific time zone	N/A
App-specific currency	N/A
Size limitations	N/A
Organic data	✓	Included
Non-organic data	✓	Included
Data freshness	Daily UTC	Updates are pushed to partners on a daily basis The in-app event lists in the rule-builder update Daily.
Historical data	✓	See data freshness
Team member access	✓	Some functions, like CSV download, setting S3, SFTP, and Slack partner connections require team member access to raw data. As a security measure, where the account of the team member, who created an audience is deleted, AppsFlyer stops uploading audience data to the advertising partner. To check the upload status of your audiences, go to Integration > Audiences. To reactivate the audience, in the Audiences window, duplicate the audience, and configure the split audience settings.

Limitations

A positive condition is required to create a base audience (primary rule). Then reduce it by negative conditions. For example, installed the app AND did not do af_purchase event.
Windows Phone not supported.
Audience names cannot be changed on the connected partners’ side.
Nested event values aren't supported.
When creating an audience, only the top 100 events display. However, when you begin entering an event name, all events display.
In-app events having more than 100 attributes by default aren't recorded. Contact your CSM if more attributes are required.
Clicks and impressions data only contain devices which have an advertising_id/idfa on their original click URL. The devices of users from SRNs, which don't use external attribution links, are not included in the clicks data for audiences.
When using the Total Revenue filter within the query builder, the estimated audience size is not available. This doesn't affect the audience creation process.
You cannot combine audiences created by import with other audiences or rulesets.
Download file size limited to 5 million records.
Google Ads: UAC campaigns aren't supported. Other campaign types are supported.