Proactive Assistant

Overview

Proactive Assistant (PA) is an AI (Artificial Intelligence) powered assistant that proactively enables marketers with key actionable insights to amplify their marketing efforts. MoEngage gathers detailed data from multiple sources, uses AI and powerful algorithms to transform it into valuable insights, and delivers insights using a simple and single-view window.

  • Proactive intel at fingertips: PA does not need instructions or queries from you to run and produce insights. But proactively combs through all of the data and generates actionable insights.
  • One-click actionable insights: The performance-boosting insights suggested by the assistant are actionable with a few clicks. You can send out a campaign immediately or save them as segments to analyze them further
  • High-impact insights delivered first: The system creates plenty of insights but showcases to customers the ones most impactful to boost performance. The system uses intelligent algorithms to evaluate the impact of every insight and prioritizes them accordingly.
  • Unlock hidden business opportunities: PA processes a wide array of data related to users, events, campaigns, devices, etc. to mine valuable performance-boosting insights, which customers couldn’t have uncovered otherwise.
  • Limited intel on key segments: A marketer-driven segmentation is often elementary and helps uncover only primary segments based on predefined values of attributes. Brands often miss out on discovering more sophisticated, outcome-driven segments, which only algorithms can help recommend. Understanding and targeting these segments could be paramount to business growth.
  • Manual efforts to mine segments: Extracting target customer segments at present requires manual efforts. Understanding the creation of micro-segments requires configuring variables such as user and event attributes, time duration, and so on. This could be time-consuming and overwhelming for some users.

Accessing Proactive Assistant

Use one of the following:

  • On the sidebar click on Assistant >> Proactive Assistant to access the Proactive Assistant.
    Access_sidebar.png
  • Click on the Assistant button in the header to access the assistant.
    Access_PA.png

Segment Insights

The Proactive assistant surfaces insights based on customer segments. There are two types of segment insights the assistant would be suggesting:

  • Segment Based on RFM: RFM (Recency, Frequency, and Monetary) Model provides auto-segmentation and bucket users into categories such as Loyal, Promising, At Risk, and so on based on user behavior. 
  • Auto Segmentation: Sherpa automatically generates user clusters or segments using proprietary AI. Specify any event and attribute used for generating these clusters.

RFM Insights

RFM Insights proactively provide valuable information to Marketers about the user base. Use MoEngage Sherpa AI engine to get the most important users and uncover insights about user behavior on your product. For more about RFM (Recency, Frequency, and Monetary) segments, refer to What is RFM analysis.

Engage with users belonging to the RFM segment based on the defined Recency, Frequency, and Monetary attributes. For example:

  • An audio streaming company would like to know which of its customers are About To Sleep and how to re-engage them.
    (Recency - App Opened, Frequency - Song Played & Monetary - Subscription Price)
  • An online retail store might want to know which of its customers Needs Attention and how can they be nudged.
    (Recency - Website Opened, Frequency - Order placed & Monetary - Price)

RFM_Insight.png

On the other hand, a brand would like to understand which of its customers have transformed from one RFM segment to another and what should be the strategy to engage with them.

  • A credit card company would like to know which of its customers have transitioned from Champions to Needs Attention, to re-engage them before losing forever quickly.
    (Recency - App Opened, Frequency - Transaction Completed & Monetary - Payment amount)
  • An OTT platform might want to know which of its customers have transitioned from Lost to Can’t Lose Them to understand what worked and keep them further engaged. 
    (Recency - App Opened, Frequency - Content Watched & Monetary - Subscription Price)

Setting up RFM insights

Navigate to Settings > Proactive Assistant > RFM Insights. Specify the events and revenue attributes that you would like to use for RFM analysis. Sherpa will provide the relevant insight, and alerts for any sizable movement of users across the RFM segments. 

Set_-_RFM.png

You can set up up to 5 different RFM Metrics that will be used as input to Sherpa. By Default, Recency and Frequency Events will be set up as App/Site Opened and Monetary Event will be picked from Settings (Settings>General>Key Metrics Parameters). These inputs will be analyzed over the duration added in “Analysis Duration” (Analysis Duration is recommended to be above 7 Days for meaningful insights). 

Auto Segmentation

It is a segmentation technique where the algorithm, K-mean clustering in this case, automatically creates clusters based on the values of the primary attributes.
A brand would be interested in discovering algorithm-recommended, hidden key customer segments based on their primary attributes to understand them better or engage with relevant campaigns. Eg: A fashion brand’s customer segments

  • Segment 1 - 2,345 users having 
    • Added to cart with average count: 242
    • Product Viewed with average count: 658
  • Segment 2 - 5,125 users having
    • Songs Played with average count: 242
    • Songs Favorited with average count: 658

AutoSeg.png

Setting up Auto Segmentation

Navigate to Settings > Proactive Assistant > Auto Segmentation. Specify the session event, conversion event, revenue attributes and add “important user actions”. “Important user actions” will be considered whenever Sherpa runs clustering and always includes relevant insights for the most important user actions. 

Set_-_Auto.png

You can set up up to 5 different Auto Segmentation Metrics that will be used as input to Sherpa. By default, Session Events will be set up as App/Site Opened, and Conversion Event will be picked from Settings (Settings > General > Key Metrics Parameters). These inputs will be analyzed over the duration added in “Analysis Duration”. Analysis Duration is recommended to be greater than or equal 7 Days for meaningful insights

Actions on Insights

Users will be able to create a custom segment or a campaign directly from an insight. Navigate to the insight that you want to use for a campaign or segment creation, and click on Actions on the bottom right.

Name the Custom Segment, save it and you will be to create a campaign with it. 

  • Segments created from the RFM Insights will be dynamic, which saves the RFM segment definition, not the exact users. The next time when a campaign with an RFM segment is sent, one time or periodically, that RFM segment gets calculated and the newly generated RFM segment will be used in the campaign.
  • Segments created from the Auto Segmentation cluster will be static, the users in the segments will be saved and will remain the same whenever it is used for analysis or a campaign.

Action.gif

 

Share: To share insight, navigate to the insight card on the top right and click on copy insight. 

Pause: If an insight type is not relevant to you right now, you can pause it for 30 days.

Feedback: You can also share any specific feedback that you have for any insight to MoEngage using the feedback button.

Proactive Assistant insights have a life of 60 days before they expire. 

Similar Use Cases
Was this article helpful?
0 out of 0 found this helpful

How can we improve this article?