Retailers often require their message to be tailored specifically to groups of their customers. This may include customization by language, age bracket, location, loyalty scheme, etc. Gametation is engineered to make tailoring the message precise and easy. The mechanism by which this is achieved is targeting.
Targeting is a decision-making process that affects nearly every aspect of the Gametation experience. Everything – from game play through ad views and prize redemption, to the graphics displayed to the user, is driven by or affected by targeting. It is a powerful tool, deeply integrated into the Gametation platform and throughout the application, and is one of the features that sets Gametation apart from its competition. It is the foundation that enables customization, which it does at many levels and granularities.
In essence, targeting gives us the opportunity to create campaigns that are tailored specifically to meet client objectives. It is what allows us to customize an experience according to a specification for a given situation, for example:
When a retailer needs to segregate user audiences e.g. by spend bands, visitation frequencies or interest categories, while offering a unique, individual set of experiences to each group, it is targeting at work; when ads are best tailored to a user based on his or her preferences, stated or implied, it is again targeting at work; when winning probabilities1 require adjustment based on loyalty scheme or membership level, yet again it is targeting driving the decision; The list goes on and on. In fact, targeting is used in many other situations, driving behaviors that include but are not limited to:
- Matching the language of the campaign to a user (including text and graphics)
- Tailoring an experience to current, local weather conditions (for example, offering a distinct set of prizes, ads, graphics, etc. because the weather is sunny)
- Locating stores near a user
- Determining the daily sponsor
- Selecting a leaderboard for a multi-player competition
From a more detailed technology-oriented perspective, targeting is the process of selecting a set of objects and properties to best match a combination of user attributes and environmental conditions according to configured, predetermined selection criteria:
The set of specific user attributes and environmental conditions available is extensive. It is campaign and scenario specific, so different criteria and data will typically be considered in different scenarios within a single campaign. For example, the criteria and data used for selecting prizes that should be available to a user in a game are typically different from the criteria and data used to determine the ads displayed to the same user, even though both selection processes use targeting. At the same time, some selection criteria and data will be identical for all aspects of a campaign, such as the language used for graphical and textual assets.
User attributes considered for targeting depend on the selection criteria configured for the scenario. They may include user preferences, experience, loyalty points collected, loyalty program membership, age, location, tags, request parameters, and other user specific information shared voluntarily or implicitly with the platform. Environmental conditions considered by targeting are also determined by the configured selection criteria, and include concepts such as current or forecast weather, date (including time, day of week, day of month, etc.), browser settings, and other information that is not user specific.
Selection criteria are pre-coded algorithms for matching patterns configured by the administrator on objects and customizations (also known as property groups or targeting groups) with actual values of user attributes and environmental conditions. The actual user attribute and environmental condition values are determined every time a user interacts with the system, whether to play a game, redeem a prize or other forms of engagement, except for cases where the system deliberately retains the game-time values to maintain consistency throughout the user’s experience. Whenever actual values are reevaluated, targeting is performed again to align the experience with the current values.
Each selection criterion has its own unique matching logic and related settings. Most criteria can be configured to either filter or prioritize the objects they evaluate. Some can produce a more detailed evaluation result, such as the location targeting criterion which can assign a result based on distance, or the language targeting criterion which can determine a precedence based on browser settings. Some criteria can evaluate complex conditions, such as the loyalty point targeting criterion which can evaluate and aggregate point totals in many distinct categories, while other such as the administrator targeting criterion can evaluate only simple conditions.
Selection criteria are combined and configured by the campaign administrator to achieve complex, conjunctive semantics. By carefully configuring the selection criteria for a scenario, the administrator can achieve semantics such as “a 1.5 liter bottle of a soft drink is available as a prize only on the 3rd Tuesday of a month when the temperature outside is above 25 degrees Celsius”. The semantics could further be expanded to offer the prize more frequently to users in the age bracket of 20 to 30, and to offer a higher winning probability1 to users who are members of a loyalty program.
A few additional facts about the use of targeting within Gametation. Targeting is:
- Highly customizable and widely integrated into application scenarios;
- Deeply integrated into the platform;
- Optimized for data retrieval – reading objects from remote storage optimizes queries based on targeting information;
- Available for real time filtering and ordering of random collections of objects;
- A tool readily available to developers for integration into new scenarios;
- Extensible – defining new selection criteria is easy through implementation of an interface;
- Designed to assist in campaign planning, management, and support by:
- Answering what-if type questions;
- Providing a mechanism for ad-hoc queries;
- Providing an experimentation framework for A/B testing, which is in essence just a specific case of tailoring experiences to groups of users defined by the characteristics necessary to meet the experiment objectives.
5 years of ongoing development has delivered a platform with a rich and deep set of opportunities for tailoring campaigns to meet clients’ business objectives and for building long-term value with more developments in the pipeline.
1. Subject to local legal regulations, campaign specifications, terms and conditions, etc.