Attribution Window

Attribution Window

Lookback Window settings can be configured in the units of App and Channel at large. In setting up App Lookback Window, default value for the entire App can be set. In setting up Channel Lookback Window, individual settings can be configured for each Paid Channels.

In case the Lookback Window values for App and Channel are different, Channel Lookback Window will be given the priority. For example, when App Lookback Window has been set for 7-day in Install Event, and Channel A has the the Install Event set as 3-day, and Channel B has not additionally changed Lookback Window by channels, the Install Event Lookback Window for Channel A will be set for 3-day. For Channel B’s ads, a 7-day Lookback Window will be applied, equally as its App Lookback Window.

On Airbridge Dashboard, when App is registered upon account creation, App Lookback Window and Channel Lookback Window by channels are configured as same, in default.

Setting Lookback Windowlink

Lookback Window which measures attribution performance by Install, Launch(Open), and In-App events can be configured by types of Apps. The value from ‘Setting App Lookback Window’ is default, which applies values from Channel which does not designate a specific Channel Lookback Window by channels.

Reference: Airbridge Attribution Model

Setting App Lookback Windowlink

  1. Go to ‘Settings’ from the left menu tab, and select ‘App Lookback Window Setting’
  2. Select items for Statistical Report (Install, Launch(Open), In-App Event)
  3. Configure Lookback Window by statistics items

Setting Install Lookback Windowlink

Definitive and Probabilistic Methodlink

Airbridge employs a total sum of six different technology types to identify if the User is the same User before install and after the install. Depending on the level of accuracy, the method is grouped into two: Definitive Method and Probabilistic Method. Definitive Method: These matching methods come under 100% accuracy, and the method examples include Google Play Store Referrer, Query String Parameter, Deep Link, ADID matching, platform matching, and cookie matching. Probabilistic Method: By combining non-unique values (i.e. device type, IP address), this method allows a probabilistic estimation to see if a User on first visit to the app is the same User who viewed and clicked the ads.

MethodTechnology TypeAccuracyClick Lookback WindowImpression Lookback Window
DefinitiveAndroid ReferrerApprox. 100%
(Loss -10%)
0-30 days (Default: 7-day)-
ID Matching (IDFA, GAID)Approx. 100%0-30 days (Default: 7-day)-
Platform MatchingApprox. 100%0-30 days (Default: 7-day)-
Deeplink Deferred Install
Matching
Approx. 100%0-30 days (Default: 7-day)-x
Cookie Matching
(iOS Safari 9.0 Above)
Approx. 100%
(May have Loss)
0-30 days (Default: 7-day)-
Probabilistic Matching
(IP Address + User-Agent)
(Fingerprinting)
Probablisitc Matching
(IP Address + User-Agent)
(Fingerprinting)
Approx. up to 85%0-96 hours (Default: 24-hour)-

Reference: Install Same User Identification Technology and Look Back Window

Launch(Open) and Setting In-App Event Lookback Windowlink

There are two types of touchpoints on Launch(Open) and In-App conversion: ‘Deep Link Touchpoint’ and ‘Winning Touchpoint on Install’. Basically, the conversion performance on Launch(Open) and In-App conversion will be included to Winning Touchpoint within Lookback Window. But if there was a ‘Deep Link Launch’ during Lookback Window, than ‘Deep Link Launch’ will be credited with a priority.

This is because the Deep Link Launch can offer a performance measurement with more accuracy on Re-engagement Ads. Suppose that a User installed an App via ad, and then re-engaged with App via an ad with Deep Link, to purchased a product within Lookback Window. In this case, the Deep Link Touchpoint from a latest re-engagement ad is given more credit for the attribution to purchase, rather than Winning Touchpoint for install.

In-App Event’s Lookback Window can be configured in the units of ‘In-App Event Category’.

Reference: Airbridge In-App Event

Comparison TargetLookback Window
The latest Deep Link Touchpoint
(Excludes Deferred Deep Link Install Matching)
0-30 days (Default: 3-day)
Winning Touchpoint on Last Install
(Includes re-installs)
0-30 days (Default: 30-day)

If you have any further inquiries relating to this matter, please reach us through ‘1:1 Inquiry’ from the bottom right corner or email us to Customer Support. We will be in touch shortly.

Channel Lookback Window Settinglink

The Lookback Window by matching methods that measures performance of Install, Launch(Open), and In-App Event can be configured according to channel types.

Reference: Airbridge Attribution Model

How to Set Channel Lookback Window Settinglink

1.Go to ‘Setting’ tab from the left, and select ‘Paid Channel Control’ tab 2. Select Paid Channel 3. Select ‘Channel Lookback Window’ tab 4. Configure Lookback Window by Stats items

Configuring Install Lookback Windowlink

Definitive Method and Probabilistic Methodlink

Airbridge employs a total sum of six different technology types to identify if the User is the same User before install and after the install. Depending on the level of accuracy, the method is grouped into two: Definitive Method and Probabilistic Method.

  • Definitive Method : These matching methods come under 100% accuracy, and the method examples include Google Play Store Referrer, Query String Parameter, Deep Link, ADID matching, platform matching, and cookie matching.

  • Probabilistic Method : By combining non-unique values (i.e. device type, IP address), this method allows a probabilistic estimation to see if a User on first visit to the app is the same User who viewed and clicked the ads.

The Definitive Method for a click can be configured up to 90-day in Lookback Window (default set as 7-day), and the Probabilistic Method can be configured up to 96-hour (default set as 24-hour) in Lookback Window. The setting for Lookback Window with a Definitive Method for impressions will soon be offered.

MethodTechnology TypeAccuracyClick Lookback WindowImpression Lookback Window
DefinitiveAndroid ReferrerApprox. 100%
(Loss up to 10%)
0-30 days (Default: 7-day)-
ID Matching (IDFA, GAID)Approx. 100%0-30 days (Default: 7-day)-
Platform MatchingApprox. 100%0-30 days (Default: 7-day)-
Deeplink Deferred Install MatchingApprox. 100%0-30 days (Default: 7-day)-
Cookie Matching
(iOS Safari 9.0 Above)
Approx. 100%
(May have Loss)
0-30 days (Default: 7-day)-
ProbabilisticProbablisitc Matching
(IP Address + User-Agent)
(Fingerprinting)
Approx. up to 85%0-96 hours (Default: 24-hour)-

Reference: Install Same User Identification Technology and Lookback Window

Launch and Setting In-Event Lookback Windowlink

There are two types of touchpoints on Launch(Open) and In-App conversion: ‘Deep Link Touchpoint’ and ‘Winning Touchpoint on Install’. Basically, the conversion performance on Launch(Open) and In-App conversion will be included to Winning Touchpoint within Lookback Window. But if there was a ‘Deep Link Launch’ during Lookback Window, than ‘Deep Link Launch’ will be credited with a priority.

This is because the Deep Link Launch can offer a performance measurement with more accuracy on Re-engagement Ads. Suppose that a User installed an App via ad, and then re-engaged with App via an ad with Deep Link, to purchased a product within Lookback Window. In this case, the Deep Link Touchpoint from a latest re-engagement ad is given more credit for the attribution to purchase, rather than Winning Touchpoint for install.

We are currently preparing for In-App Event Lookback Window by Channel. As of now, the setting value from ‘App Lookback Window' applies to all Channel, at the Setting App Lookback Window from the sidebar**.

Reference: Airbridge In-App Event

Comparison TargetLookback Window
The latest Deep Link Touchpoint
(Excludes Deferred Deep Link Install Matching)
0-30 days (Default: 3-day)
Winning Touchpoint for the latest install
(Includes re-install)
0-30 days (Default: 30-day)

If you have any further inquiries relating to this matter, please reach us through ‘1:1 Inquiry’ from the bottom right corner or email us to Customer Support. We will be in touch shortly.