Einstein Send Time Optimization for Mobile Model Card

Person or Organization

Salesforce Einstein for Marketing Cloud Engagement

Model Date and Version

• December 2020
• Minor changes can occur throughout the release
• Major changes can occur and are communicated via release notes

Model Type

• Recommendation/Prediction, Latent factor matrix factorization

General Information

• In Engagement, Einstein Send Time Optimization for Mobile optimizes notification push time to maximize the overall engagement of mobile apps.
• The model chooses the optimal time to send a push notification to the contact to maximize the probability of the contact engaging with the app. Contacts engage either by tapping on the push notification to open the app or launching the app directly. The model also detects individual in-app patterns to avoid sending push notifications when contacts are already in the app. A contact can have different optimal send time for different apps.
• We train each customer’s model only using that customer’s data.
• A contact who doesn’t have enough data is part of the Pending Personalized Send Time group. A contact is categorized as not having enough data when there isn’t at least 1 open in the past 90 days. For these contacts, the model defaults to send one of two ways based on the Insufficient Data Option that the admin has selected in the STO Setup page.
• The two options are Send at Einstein’s optimal default time and Send immediately. When you select Send at Einstein’s optimal default time, Einstein aggregates across all of your contacts that do have sufficient data to create a distribution of scores across every hour. Einstein then follows this distribution of scores to randomly assign scores to contacts who don’t have enough data. A contact without enough data is randomly assigned an optimal time, but it’s more likely that the contact gets a time with a higher score than one with a lower score. This randomization is added so that everyone in the Pending Personalized Send Time group doesn’t receive their email at the same time. Randomization helps avoid spikes and helps Einstein get better data.
• Contacts in the Pending Personalized Send Time group are sent a push notification immediately when they reach the STO activity.

Licenses

Einstein Send Time Optimization for Mobile is available to Engagement customers with these editions.

• Corporate Edition
• Enterprise Edition
• Enterprise+ Edition
• Pro Edition with a Journey Builder for Pro Edition SKU
• Einstein Send Time Optimization SKU

Primary Intended Uses

The Send Time Optimization model powers the Einstein Send Time Optimization activity in Journey Builder. Journey Builder is a tool that marketing professionals use to increase effectiveness in their mobile marketing campaigns.

Out-of-Scope Use Cases

Anything other than the primary use case is out of scope and not recommended. For example, reducing overall send latency for reservation booking mobile messages isn’t an intended use case.

Model Input

Einstein Send Time Optimization for Mobile analyzes up to 90 days of historical engagement patterns per app. The engagement history includes these factors.

• Push events and reactions, such as push notifications and taps on push notifications, and associated timestamps
• Data and metadata about customer sending patterns and how campaigns are executed

The engagement history that Einstein STO analyzes excludes these factors.

• Data purchased or collected from third parties
• Demographic Data, which is typically stored in SFMC as data extensions or subscriber or contact attributes
• Specific content within the push template or rendered push message body

Groups

The model doesn’t include any demographic data or other data purchased from third-party data providers. The model uses mobile engagement data such as Push, TapOpen, AppOpen, or TimeInApp.

Environment

The model is trained and deployed in Marketing Cloud Engagement.

Model Performance Measures

Aggregated model performance metrics are gathered to monitor, ensure, and improve the quality of the model. Model performance metrics included the spread and sparsity of input metrics and the correlation of output scores to historical observations. All metrics are aggregated and anonymized.

In general, to provide recommendations, Salesforce requires at least one engagement event, such as opens or clicks of a commercial email in the past 90 days. When you activate Einstein Send Time Optimization, existing data counts toward the 90 days.

However, biased use of Send Time Optimization (STO) for Mobile can still introduce bias to your marketing process.

Scores and Models

Einstein Engagement Scoring scores and models for email are updated approximately weekly. The refresh cadence varies by one to several days based on a customer's individual business unit.