Appearance

Local Engagement Engine (LEEN) ​

We have succeeded to a degree in building apps that generally have good user experience. However, we face a new challenge, the issue of engagement. The average user spends more than 15 minutes to 1 hour in apps built by Hubtel. Only a maximum of 15% of total active users within a month come back within a day. This poses a serious concern since apps which are not frequently are marked as in-active by the operating systems and are usually suggested for deletion. To combat this, we need to make sure our app can engage the user based on past behaviors in a manner that is extremely helpful but not disturbing. There seems to be a way to achieve this using our analytics events complementing a new engine. This engine, the Local Engagement Engine is simple but has the potential of incredible outcomes.

Overview ​

The local engagement engine is designed to enhance user interaction by delivering targeted in-app messages and notifications. This is achieved by monitoring user events, matching these events to predefined engagement behaviors, and managing an engagement artifact queue to optimize the timing and relevance of messages sent to users.

Workflow Description ​

  1. Event Capture: Each user interaction with the app triggers an event. This can range from opening the app to engaging with specific features.

  2. Behavior Matching: Captured events are evaluated by the Event to Engagement Behaviour Matcher to determine if they correlate with predefined engagement behaviors. This decision point routes the process either towards creating an engagement artifact if a match is found or loops back to event capture if no relevant behavior is identified.

  3. Engagement Artifact Creation: When a matching engagement behavior is identified, an Engagement Artifact is created. This artifact is tailored to the user’s recent activity and designed to enhance engagement, such as through a personalized notification or an in-app message.

  4. Engagement Queue Management: The newly created engagement artifact is added to the Engagement Delivery Queue. This queue is then reordered based on the Click-Through Rate (CTR) of each type of engagement, prioritizing those with higher user interaction rates.

  5. Queue Threshold Evaluation: The length of the Engagement Queue is continually assessed against a predefined threshold. If the queue length exceeds this threshold, it is trimmed down to the set limit by removing the least effective engagements as determined by their CTR. If below threshold, the process proceeds without trimming.

  6. Scheduling Engagement Artifacts: Engagement artifacts within the optimal queue length are scheduled for delivery as notifications or in-app messages. This scheduling considers the user's likely receptiveness, based on past interaction data and the urgency of the message.

  7. User Interaction and Feedback Loop: Once an engagement artifact is displayed, the system monitors for user interaction. If the user engages with the notification or message, the CTR for that type of engagement is increased, enhancing its priority in future queue reorderings. Conversely, if the user does not engage, the CTR is decreased, potentially lowering the priority of similar future engagements.

  8. Continuous Improvement: The engagement engine continuously refines its behavior matching and artifact creation strategies based on ongoing user interaction data. This adaptive approach ensures that users receive the most relevant and timely engagements, increasing overall app usage and satisfaction.

Conclusion ​

This local engagement engine strategically integrates event monitoring, behavior analysis, and user feedback to cultivate a personalized user experience. By dynamically adjusting engagement strategies based on user behavior, the engine not only maintains user interest but also fosters deeper interaction with the app.