Congratulations! Your organization  has agreed to embark on the exploration or build of a Real-Time Interaction Engine powered by AWS. Whether you are actively looking to replace an existing monolith (see Real Time Interaction Engine for AWS: Breaking the Monolith), or build on existing in-house development, you will need to consider what should be in your initial Minimum Viable Product (MVP). 

Below, we will explore what the top 3 features for implementing a real-time interaction or decision engine are.

#1 Whitelist / Blacklist

One of the foundational features of any Real-Time Interaction solution is the availability of a “whitelist” capability. 

Whitelist  is the simplest way of extending existing outbound campaigns to inbound channels as part of an “inbound push”. For example, customers that have been targeted for an email promotion may also be presented the same offer when they login to an online portal. 

When using whitelists, the decision about which offer to serve to which customer is calculated with data that is at least 24 hours old. Although the decision is made on this relatively stale data, the delivery of the message on the inbound channel such as mobile or web happens in real-time.

Your MVP should be able to accommodate a wide variety of input sources for the whitelist. For example:

  • Your existing outbound marketing tool, such  as Unica Campaign, Adobe Campaign, SAS MA, or Salesforce Marketing Cloud
  • Infrastructure data management tools like AWS Glue
  • Marketing and customer engagement tools like AWS Pinpoint
Whitelists are most often found in conjunction with their blacklist (or “offer suppression”) counterparts. These lists deny specific messages for individual customers, and effectively override the whitelist.
 

#2 Offer Rotation / Contact Fatigue

Often, customers will be eligible for more than one offer in the whitelist at any given time. For example, a high value customer may be targeted by both the wealth management and lending areas of a retail bank. 

To help prioritize offers for each individual, the whitelist should contain a marketing score, which is used to rank offers.  

With the use of marketing scores, it is possible for your MVP to present messages with the highest scores first. Once all offers have been shown, your solution should rotate through the eligible messages to ensure an equal number of impressions. 

In addition, other types of common fatigue rules are often considered, such as: maximum number of impressions within a 15 minute period for a single message, maximum number of impressions within 72 hours for a single message, or a maximum number of targeted messages of any kid over a period of 7 days. 

#3 Readiness to Integrate with Best of Cloud

Your Real-Time Interaction MVP need not replicate features like-for-like with legacy systems. A more forward looking approach is to plan to have your solution easily adaptable to integrate with innovative cloud based services. 

For example:

  • AWS Personalize AI driven product and message recommendations to complement traditional manual marketing segmentation
  • AWS SageMaker to help activate your existing machine learning models on inbound channels
  • AWS Kinesis Analytics to override whitelist messages when real-time customer behavioural patterns are detected (e.g. predicted customer would be interested in lending, but seeking travel credit card instead)

Accelerating your MVP

SpadinaBus has developed an accelerator, which helps enterprise customers to rapidly develop a Real-Time Interaction solution using AWS. This accelerator reduces discovery and design time, as well as project risk, which allows for value realisation within 90 days.

The accelerator includes the Top 3 features described here, as well as others. Contact us to learn more about the Real-Time Interaction accelerator for AWS.