Most organizations are already doing some form of omnichannel marketing using disparate 3rd applications and on-prem data stores. Amazon employs a combination of homegrown tools they’ve developed over the years, and they’ve made some of those tools available via Amazon Web Services for anyone to use. Today I’m going to focus on using your historical customer marketing and purchase history to power a recommendation engine called Amazon Personalize that can auto-populate product recommendations in customized emails using Amazon Pinpoint. You can create dynamic audience segments in Pinpoint based on demographic data, behaviors, and custom attributes. If you already have a solution for managing your customer lists you can import an audience from another tool such as a Customer Data Platform (CDP) like Tealium, Segment, or mParticle.
What is Amazon personalize
Amazon Personalize is a pay-as-you-go service on AWS powered by Machine Learning (ML) models to make customer recommendations. The service applies Auto-ML to your data to pick the best pre-tuned algorithm, so you don’t have to be a data scientist to use the service. Once trained, Amazon Personalize generates real-time or batch recommendations based on your customer data input via an API call.
What is Amazon pinpoint
Amazon Pinpoint is a graphically based omnichannel tool built for marketers that allows you to segment customers, create outbound campaigns across multiple channels (text, email, voice, or custom channels), and create journeys to nurture prospects dynamically based on their behavior.
Omnichannel marketing like Amazon – What you’ll need
For Pinpoint, I think most marketers familiar with email services will quickly acclimate to how to create, execute, and review campaigns and customer journeys.
For Amazon Personalize, you’ll need a set of data on your customer’s interactions with your products. You may have more success working with your Amazon team or a developer at your company.
Upload your customer data into Amazon personalize. Typically, a CSV or excel file uploaded to a folder in AWS Simple Storage Service with the following datasets.
Interactions Dataset
- User ID (Your customer identification key)
- Item ID (Your product)
- Timestamp (when did the interaction happen?)
User Dataset
- User ID (Customer)
Item Dataset
- Item ID (Product)
You can add additional metadata (context) to the above data sets. For example, you could add demographic data for your users (age, location, persona, etc.) or product description (color, product family, material, etc.)
Once you upload the data, you’ll pick a Solution. Now you’ll determine what kind of recommendation you want via a “Recipe” choice. Here are the Recipes and an easy to understand table that helps you select the best recipe for your use case.
- User Personalization: The User-Personalization (AWS-user-personalization) recipe is optimized for all personalized recommendation scenarios.
- Popularity count – Popularity-count recommends the most popular item items based on all of your user behavioral data.
- Personalized ranking: The Personalized-Ranking recipe generates personalized rankings of items. A personalized ranking is a list of recommended items that are re-ranked for a specific user. This is useful if you have a collection of ordered items, such as search results, promotions, or curated lists, and you want to provide a personalized re-ranking for each of your users.
- SIMS: The Item-to-item similarities (SIMS) recipe is based on the concept of collaborative filtering. A SIMS model leverages user-item interaction data to recommend items similar to a given item. In the absence of sufficient user behavior data for an item, this recipe recommends popular items.
Creating your A/B campaign tests in Pinpoint
The first few steps for getting started with Pinpoint are intuitive. You need to upload a customer list(s), which Pinpoint refers to as segments.
Next, you’ll set up your Amazon Personalize model in Pinpoint by selecting it in a drop-down menu.
Now you’re ready to create or import an email template. You can create a campaign in Pinpoint that allows you to pick your segment, the channel (Email, text, voice, custom), A/B split, and timing of execution.
You’ll add dynamic “Recommendation” fields to your email using the template designer tool in Pinpoint.
You’ll track and evaluate the open rates, click thought, and overall all effectiveness of the campaign and your machine learning (ML) powered recommendation engine from the same portal.
How expensive is this?
Both services are pay-for-what-you-use so that it will come down to volume. What’s intriguing to me is that PinPoint is a SaaS customer engagement solution for marketers that’s free to get started. (no charge for the first 5,000 endpoints (email/phone number/application) every month. Below are links to the pricing pages with pricing examples.
- $1 per 10,000 emails ($00.0001 per email)
- $0 for the first 5,000 endpoints (email/phone number/application) in your monthly targeted audience (MTA), and $1.20 per 1,000 endpoints after that.
- Transactions per second (TPS) – $0.20 per TPS-hour for real-time recommendations
But wait, there’s more! (what else you can do with Pinpoint and personalize)
We just walked through one use case for Amazon Pinpoint with Amazon Personalize, but there are several other widespread use cases. You can leverage Amazon personalize with your existing web applications for real-time shopping recommendations like you see on Amazon.com. You can also leverage Pinpoint as your omnichannel hub to orchestrate communication to customers and prospects across text, email, voice, and custom applications via web-hooks. Amazon Pinpoint with Personalize is a powerful combination putting machine learning in the hands of professional marketers. I wish you the best on your journey.
Don’t know who your AWS rep or solution architect is, fill out the contact us form for help here – https://aws.amazon.com/contact-us/.