This article describes how the schema enables audience targeting and the optimization of value metrics.
Table of Contents
- What is the Schema?
- Schema Fields: Sources & Targets
- Default schema fields
- Adding schema fields
- Using data defined in the schema
What is the Schema?
The Schema is a set of definitions that describe data in a visitor’s session. Each definition describes how the system should process context data. The schema applies to all environments in your Evolv AI account organization.
The data defined in the schema can be used in two ways.
- Personalization or targeting ideas to visitors based on their context attributes, e.g., cart_age, return_status
- Optimizing numeric attributes, e.g., order_value, video_minutes_watched
Schema Fields: Sources & Targets
A definition in the schema is made up of a Source Field (the expected context attribute) and Target Fields (how to store and process the data)
Source fields
A source field defines the context attribute key and the type of data it should expect.
Target fields
A target field defines how the value of the source field should be stored and processed.
Each source field can have one or more target fields. See example below.
Default schema fields
The schema includes several browser-based fields by default. This data is automatically collected from implementations that use the Experience Accelerator integration.
Source Field | Type | Options |
web.url | string | Regex pattern match |
device | enum | Mobile, Desktop, Tablet |
platform | enum | Windows, iOS, Android, macOS, Linux |
location | enum | Countries of the world |
web.client.browser | enum | Chrome, Firefox, Safari, IE, Edge, Other, Unspecified |
Adding schema fields
New fields can be added to the schema to map custom context attributes you add from your digital experience.
To learn how to define fields in the Schema, read Adding context attributes to the Schema.
Using data defined in the schema
Targeting audiences with context attributes
A key aspect of any optimization program is identifying segments of your audience that are experiencing some friction and coming up with unique solutions to address those issues.
In browser-based experiences, it’s easy to target segments like device or location using user data captured by the browser. Evolv AI adds these user attributes to the context by default. But what about user data that only your system knows about?
Using Evolv AI’s SDKs, you can set custom context attributes about the users in your digital experience.
Targeting users based on their context attributes is a core capability of the Evolv AI system and can be used in two ways:
- Project Audience –The project audience is the first filter to all test ideas in your project
- Variable Targeting – Variable targeting allows you to test a set of ideas on a narrower segment of your project audience. This is useful when you want to include personalization ideas in a broader audience. This targeting can also be applied to Variable Groups.
- Variant Targeting – Variant targeting is only available through Evolv AI's auto personalization capability, which analyzes context attributes to discover hidden personalization opportunities.
To learn more, read Targeting audiences with context attributes.
Optimizing the value of context attributes
Optimizing a value in the digital experience can be an effective way to test ideas that affect revenue and engagement.
When setting up a project, the conversion value is a scalar quantity you want the system to optimize. In order to optimize a value, you must also identify a conversion event. This value is calculated for each session when the specified conversion event occurs at least once.
Here are examples of Optimization Metrics that include a conversion value:
- Order Value (value) when a visitor reaches the thank you page (event)
- Video Minutes Watched (value) when a video is played (event)
- Lead Score when a lead form submission is confirmed
- Funnel progression value when a visitor makes progress down the funnel (a valuation of the distance a user goes in the funnel)
To learn more, read Optimizing the value of context attributes.