Open attribution Modeling: Guide to measuring DFINERY performance

What is Open Attribution Modeling?

Open Attribution Modeling is a combination of Ad Touch and Attribution Method, the priority (Tier) of the Attribution Unit and the Ad Touch-Open Lookback Window. This is a powerful function that can increase the transparency of advertising performance measurement according to media operation strategy by adjusting (Ad Touch - Open lookback window).

For example, high priority is assigned or a long lookback window is assigned to attribution units that are judged to have high accuracy or a strong impact on app installations, while attribution units that are judged to have relatively low accuracy or have a weak impact on app installations are assigned high priority. You can assign a low priority to the review unit and set a short lookback window.

How it works and what it means

Measure performance by exploring last ad touches in the order of attribution units assigned to high-priority tiers.

Assume that click-referrer and click-adid are assigned to Tier 1, and click-fingerprint and impression-fingerprint are assigned to Tier 2. At this time, DFINERY first searches to see if there is a last ad touch that matches Tier 1 click-referrer and click-adid attribution units.assigned to Tier 1 If there is no last touch matching the attribution unit, the last touch is searched for attribution units assigned to Tier 2 in the following order:

The implications of attribution modeling are simple yet powerful. With the attribution modeling provided by DFINERY, you no longer need to be dependent on the last touch attribution model.If you want to continue using fingerprinting as an indirect attribution method, but do not value the reliability of the results measured using this method, you can now assign fingerprinting to Tier 2 or Tier 3.

[[Quote:Guide:Small]]A brief definition of Tier 1 and Tier 2.
The units applied to Tier 1 (Click_referrer, Click-Identifier (adid/idfa)) are deterministic mapping modeling.
The unit (Fingerprint) applied in Tier 2 is probabilistic mapping (estimation) modeling.

Understanding attribution modeling

Attribution Unit

DFINERYs attribution modeling adjusts the priority of the attribution unit, which is a combination of Ad-Touch and Attribution Method, and creates a lookback window between Ad-Touch and app execution. It works by specifying (Ad Touch-App Open Lookback Window).

The ad touch and attribution methods provided by DFINERY are as follows.

• Ad touch: click, impression, playable, video
• Attribution method: referrer, fingerprint, ip, adid(idfa), cookie

An example of an attribution unit is as follows.

• click-referrer
  Advertising participants click on the advertising partner's advertising banner, land on the market, install and run the app. At this time, it is determined by the referrer that the app was installed and launched through the advertisement banner.
  
  
• impression-fingerprint
  After being impressed by the advertising partners advertising banner, advertising participants go to the market to install and run the app.At this time, it is determined by fingerprint that the app was installed and executed through the advertisement banner.

Lookback Window

Set the time to acknowledge open attribution contribution after ad touch. For example, if the lookback window of the Click-Referrer unit is set to 7 Days, only open attributions that occurred within 168 hours after the ad touch using the Click-Referrer unit method are recognized as the performance of the advertising medium.

Setting up

Go to the Attribution > Attribution Modeling menu.

Basic setting

The basic settings for attribution modeling are as follows.

[[Quote:Warning:Small]] For certain partners (example: Kakao, etc.), you must move the unit in the 'Not used' tab on the right. [Go to major advertising partner setup guide]

Change settings

• Set the desired attribution method for each tier by dragging and dropping it. Alternatively, you can create a new priority by clicking Add Tier.
• The vertical order of each tier has no effect on attribution modeling.

[[Quote: Information: Small]] Google Play app and One Store app cannot be linked at the same time.

Setting example

If set as below, we will first search to see if the last ad touch exists within the attribution unit included in Tier 1.
after Explore last add touches in Tier 2, Tier 3 order.

• Tier 1: Click-Referrer, Click-Identifier, Google UAC, Twitter (Click), Apple Search Ads (Click)
• Tier 2: Click-Fingerprint, Click-IP
• Tier 3: Impression-Identifier, Impression-Fingerprint, Impression-IP, Twitter(Impression)