Question-5: You have been hired by a travel company as a Machine Learning expert and you are assigned with the task for customer loyalty program, where you need to inspect the website visitor clickstream like the flight he is looking for etc and you need to use this information to deliver the email, voice, text messages for various offers based on his personal interest and previous data stored (if an existing customer). Which of the following you would be using to solve this problem.
- You would be using AWS Kinesis Stream to receive data in real-time and then using the Kinesis Analytics you can run the query in near-real time and send the messages using SNS.
- You would be using AWS Kinesis Fire hose and deliver all the Clickstream data in the AWS RDS where this real-time data can be joined with the existing data stored and offer would be sent using the SNS over various possible channels like Mobile, eMail, Text messages etc.
- You would be using Kinesis stream and create an EMR cluster using Spark to process data using the Spark streaming and generate various offers by writing SQL queries and send a message with various channel using AWS SNS.
- You would be using AWS Pinpoint to identify the customer interaction patterns and Amazon Personalize over the AWS SageMaker for applying Machine Learning to generate next offer.
Exp: In the given question Airline Wanted to create something like Loyalty program for the customer. Where Amazon Pinpoint to identify customer interaction patterns, group them into segments, and deliver the messages across channels. Amazon Personalize (or Amazon SageMaker) is the machine learning service used to generate the “next-best-offer” predictions that power the personalized recommendations.
Remember earlier where we talked about sending the “right message to right customer over the right channel at the right time?” Amazon Pinpoint takes care of the customer, channel, time, and message transport while Amazon Personalize or Amazon SageMaker takes care of the personalized or recommended offer content.
Channel of delivery: Are you delivering your message across the right channel for each customer? For instance, your younger customers may prefer mobile channels like SMS while your older customers may prefer voice. You also probably won’t want to send a time-sensitive, proximity-based message over email. Amazon Pinpoint can deliver engagements across multiple channels, determine which channel each customer engages with the most, and even redeliver messages between channels if the customer doesn’t engage over the first.
Segmentation: Are you able to segment your customers into groups based on their unique profiles or behavioral characteristics, such as “frequent visitors,” “new customer,” “corporate customer,” or “likely to churn?” Amazon Pinpoint enables you to programmatically segment your audience into groups like these, and machine learning tools like Amazon Personalize and Amazon SageMaker can turbocharge your segmentation capabilities. (I’ll talk more about this in a later post.)