AWS Messaging & Targeting Blog

How quirion created nested email templates using Amazon Simple Email Service (SES)

This is part two of the two-part guest series on extending Simple Email Services with advanced functionality. Find part one here.

quirion, founded in 2013, is an award-winning German robo-advisor with more than 1 billion Euro under management. At quirion, we send out five thousand emails a day to more than 60,000 customers.

Managing many email templates can be challenging

We chose Amazon Simple Email Service (SES) because it is an easy-to-use and cost-effective email platform. In particular, we benefit from email templates in SES, which ensure a consistent look and feel of our communication. These templates come with a styled and personalized HTML email body, perfect for transactional emails. However, managing many email templates can be challenging. Several templates share common elements, such as the company’s logo, name or imprint. Over time, some of these elements may change. If they are not updated across all templates, the result is an inconsistent set of templates. To overcome this problem, we created an application to extend the SES template functionality with an interface for creating and managing nested templates.

This post shows how you can implement this solution using Amazon Simple Storage Service (Amazon S3), Amazon API Gateway, AWS Lambda and Amazon DynamoDB.

Solution: compose email from nested templates using AWS Lambda

The solution we built is fully serverless, which means we do not have to manage the underlying infrastructure. We use AWS Cloud Development Kit (AWS CDK) to deploy the architecture.

The figure below describes the architecture diagram for the proposed solution.

  1. The entry point to the application is an API Gateway that routes requests to a Lambda function. A request consists of an HTML file that represents a part of an email template and metadata that describes the structure of the template.
  2. The Lambda function is the key component of the application. It takes the HTML file and the metadata and stores them in a S3 Bucket and a DynamoDB table.
  3. Depending on the metadata, it takes an existing template from storage, inserts the HTML from the request into it and creates a SES email template.

Architecture diagram of the solution: new templates in Amazon SES are created by a Lambda function accessed through API Gateway. THe Lambda function reads and writes HTML from S3 and reads and writes metadata from DynamoDB.

The solution is simplified for this blog post and is used to show the possibilities of SES. We will not discuss the code of the Lambda function as there are several ways to implement it depending on your preferred programming language.

Prerequisites

Walkthrough

Step 1: Use the AWS CDK to deploy the application
To download and deploy the application run the following commands:

$ git clone https://github.com/quirionit/aws-ses-examples.git
$ cd aws-ses-examples/projects/go-src
$ go mod tidy
$ cd ../../projects/template-api
$ npm install
$ cdk deploy

Step 2: Create nested email templates

To create a nested email template, complete the following steps:

  1. On the AWS Console, choose the API Gateway.
  2. You should see an API with a name that includes SesTemplateApi.
    Console screenshot displaying the SesTemplateApi
  3. Click on the name and note the Invoke URL from the details page.

    AWS console showing the invoke URL of the API

  4. In your terminal, navigate to aws-ses-examples/projects/template-api/files and run the following command. Note that you must use your gateway’s Invoke URL. 
    curl -F html=@m-full.html -F "isWrapper=true" -F "templateName=m-full" -F "child=content" -F "variables=FIRSTNAME" -F "variables=LASTNAME" -F "plain=Hello {{.FIRSTNAME}} {{.LASTNAME}},{{template \"content\" .}}" YOUR INVOKE URL/emails

    The request triggers the Lambda function, which creates a template in DynamoDB and S3. In addition, the Lambda function uses the properties of the request to decide when and how to create a template in SES. With “isWrapper=true” the template is marked as a template that wraps another template and therefore no template is created in SES. “child=content” specifies the entry point for the child template that is used within m-full.html. It also uses FIRSTNAME and LASTNAME as replacement tags for personalization.

  5. In your terminal, run the following command to create a SES email template that uses the template created in step 4 as a wrapper.

Step 3: Analyze the result

  1. On the AWS Console, choose DynamoDB.
  2. From the sidebar, choose Tables.
  3. Select the table with the name that includes SesTemplateTable.
  4. Choose Explore table items. It should now return two new items.
    Screenshot of the DynamoDB console, displaying two items: m-full and order-confirmation.
    The table stores the metadata that describes how to create a SES email template. Creating an email template in SES is initiated when an element’s Child attribute is empty or null. This is the case for the item with the name order-confirmation. It uses the BucketKey attribute to identify the required HTML stored in S3 and the Parent attribute to determine the metadata from the parent template. The Variables attribute is used to describe the placeholders that are used in the template.
  5. On the AWS Console, choose S3.
  6. Select the bucket with the name that starts with ses-email-templates.
  7. Select the template/ folder. It should return two objects.
    Screenshot of the S3 console, displaying two items: m-full and order-confirmation.
    The m-full.html contains the structure and the design of an email template and is used with the order-confirmation.html which contains the content.
  8. On the AWS Console, choose the Amazon Simple Email Service.
  9. From the sidebar, choose Email templates. It should return the following template.
    Screenshot of the SES console, displaying the order confirmation template

Step 4: Send an email with the created template

  1. Open the send-order-confirmation.json file from aws-ses-examples/projects/template-api/files in a text editor.
  2. Set a verified email address as Source and ToAddresses and save the file.
  3. Navigate your terminal to aws-ses-examples/projects/template-api/files and run the following command:
    aws ses send-templated-email --cli-input-json file://send-order-confirmation.json
  4. As a result, you should get an email.

Step 5: Cleaning up

  1. Navigate your terminal to aws-ses-examples/projects/template-api.
  2. Delete all resources with cdk destroy.
  3. Delete the created SES email template with:
    aws ses delete-template --template-name order-confirmation

Next Steps

There are several ways to extend this solution’s functionality, including the ones below:

  • If you send an email that contains invalid personalization content, Amazon SES might accept the message, but won’t be able to deliver it. For this reason, if you plan to send personalized email, you should configure Amazon SES to send Rendering Failure event notifications.
  • The Amazon SES template feature does not support sending attachments, but you can add the functionality yourself. See part one of this blog series for instructions.
  • When you create a new Amazon SES account, by default your emails are sent from IP addresses that are shared with other SES users. You can also use dedicated IP addresses that are reserved for your exclusive use. This gives you complete control over your sender reputation and enables you to isolate your reputation for different segments within email programs.

Conclusion

In this blog post, we explored how to use Amazon SES with email templates to easily create complex transactional emails. The AWS CLI was used to trigger SES to send an email, but that could easily be replaced by other AWS services like Step Functions. This solution as a whole is a fully serverless architecture where we don’t have to manage the underlying infrastructure. We used the AWS CDK to deploy a predefined architecture and analyzed the deployed resources.

About the authors

Mark Kirchner is a backend engineer at quirion AG. He uses AWS CDK and several AWS services to provide a cloud backend for a web application used for financial services. He follows a full serverless approach and enjoys resolving problems with AWS.
Dominik Richter is a Solutions Architect at Amazon Web Services. He primarily works with financial services customers in Germany and particularly enjoys Serverless technology, which he also uses for his own mobile apps.

The content and opinions in this post are those of the third-party author and AWS is not responsible for the content or accuracy of this post.