Available Attribution Methods

The Marketing Analytics platform supports multiple different attribution methods which work in tandem to provide you with a full picture of your marketing campaigns, no matter what platform you are targeting. Gamesight utilizes a mix of attribution methods, always attempting to select the strongest options based on the information available.

22002200

Example Device Graph composed through multiple attribution methods

Method Details

Session Matching

Supported Flows: Web to Web, Web to Sever Side Event, Web to Self Distributed PC Game
Session Matching is used in marketing flows where the GSID value is available on both sides of the flow. This type of matching is available in several flows:

  • Web Events When you measure Goals in a web browser using the Web Measurement SDK, you will get direct matching of events with touchpoints from the same browser session utilizing simple session cookies.
  • Sever to Server Events It is also possible to pass a GSID value through the Measurement API if you have your own systems to mapping the session ID value to a user. This can enable server-side measurement of web activity
  • Self Distribution If you are self-distributing your game on desktop, it may be possible to append the GSID value into the game package during the download process. In this way you can get deterministic matching to the in-game events send to the Measurement API.

In all of these flows, the session ID value is used to deterministically match an event into the device graph and identify other identities, devices, and touchpoints associated with this user.

Identity Matching

Supported Flows: Web to PC Game, Web to Mobile App, Web to Console, Cross-game Promotion
Identity Matching is applied when you have a shared ID that is known at multiple stages of your marketing flow (such as an account ID or external network ID). This value can be associated with a device using the Web SDK, Measurement API, Touchpoint Parameters, or direct import to build out the device graph for your title. This device graph maps IDs to specific devices allowing touchpoints to be matched deterministically.

There are a few different types of Identity information that you can utilize:

  • Primary User ID - The User ID values provided in Events will automatically be utilized to build Identity matches in the device graph. This commonly is the primary account ID for in-game accounts.
  • External ID - External IDs can be attached to Events as well for additional identities that are available for matching on this user (for example, emails, wallet ids, analytics ids, etc)
  • Org Scoped IDs - In the case that you have IDs that are shared between multiple titles, you can set the scope on your external ID to be organization wide, allowing the ID matches in your device graph to be shared between all games in your org.

Here are a few examples of flows that would directly benefit from Identity matches in the device graph

  • Multiplatform Flows - If you have multiple places where a user can sign in to their account (website, in-game, mobile app, forum, etc) then you will be able to use that shared identity to do deterministic matching for any touchpoints that user has generated on those devices. For example, if a user clicks on an Ad, signs in/up on your website, then uses the same account in-game you will get deterministic matching for that install.
  • Cross Game Promotion - If you would like to measure a campaign that is promoting between titles, it is possible to attach identity directly to the touchpoints measured from that cross promotion through Touchpoint Parameters. This enables Identity matches for cross promotion campaigns.
  • Identity Graph Reuse - If you configured Org Scoped IDs, every time you launch a new game you will be able to re-use portions of your Device Graph, automatically leveraging your first party data to improve your attribution quality for all future titles.

Fingerprint Matching

Supported Flows: Web to PC Game, Web to Console, PC Game to PC Game
The Fingerprint Matching method is actually a combination of many different fingerprinting and probabilistic matching techniques that we have tuned for PC and Console attribution. While not deterministic matching like the other methods, we have measured an accuracy of 90%+ for matches made through our fingerprinting techniques in controlled tests on PC, and over 83% on Console. You can read more about data utilized to generate fingerprints in our Data Handling Details guide.

Referral Codes

Supported Flows: All
If you have a flow such as a Creator Program or Affiliate Program where you want to utilize referral codes to provide attribution, this can be achieved by including the referral code value in your Event Payload. It is often preferable to provide users with a direct referral code field when running these types of programs in addition to automated attribution to allow for measurement on flows that otherwise would not be directly measurable (eg word or mouth). Please get in touch with your Gamesight Rep if you would like to utilize referral code measurement.

External Attribution

Supported Flows: All
In some cases, such as when you are integrating Gamesight with a MMP or some other external attribution source, you may want to provide known attributed details alongside an Event. If you would like to use External Attribution please get in touch with your Gamesight Rep who can provide details for the technical requirements.

Activity Matching

Supported Flows: Web to Web
In some flows, ID data from Ad Platforms (eg Google, FB, etc) is available to be used to stitch together flows in both single and cross device scenarios. For example, Google's parallel tracking implementation has Gamesight utilize the GCLID value to join together multiple GSID values in some cases. This matching is similar to Identity matching in its behavior, except that the ID data is being provided directly by the Ad Platform.

Device ID Matching

Supported Flows: Web to Mobile App, Mobile App to Mobile App
Device ID Matching is used in situations where a mobile device advertising identifier is available on both sides of the flow (such as an IDFA).

For all of these attribution methods, a successful match will lead to an entry being created in the device graph. This ties the device to a specific identity enabling more consistent attribution through deterministic attribution techniques for future campaigns.

Attribution Waterfall

In the case that your marketing flows are supported by multiple different attribution methods, the following waterfall for matching priorities will be applied. Matches will be attempted in priority order down the list until a match is found or no more methods are available.

  1. External Attribution
  2. Referral Codes
  3. Session Matching
  4. Device ID Matching
  5. Identity Matching
  6. Activity Matching
  7. Fingerprint Matching

You can think of the matching priorities as "weights" on the edges in the device graph with the more direct and reliable match sources are prioritized as higher weight than indirect methods.