SFMC - Triggered - How it works

How Phrasee's triggered experiment technology is working in the background

Updated over a week ago

Phrasee's triggered experiments are an innovative way to optimize language on a recurring basis. With the Salesforce Marketing Cloud integration, it's easier than ever for Salesforce users to say it better with their triggered email experiments.

Simply put, the integration works thanks to a combination of custom data extensions, queries and API calls created and run in the background, which keeps the steps you do to a minimum. Awesome, right?

But for all our curious phriends, let's dive a little deeper!

After you've completed the getting started steps and worked with your Customer Success Manager to define your testing parameters, your triggered experiment goes to work in the background to create the following assets in Marketing Cloud:

  1. A code snippet

  2. Four data extensions

  3. A query

We'll go over each one below.

Code snippet

For every triggered experiment you create in the Phrasee platform, the integration will create a content block known as a code snippet. This code snippet will be named after the experiment you make in Phrasee.

This content block will be automatically created and placed in a folder named after your experiment in Content Builder. You can find it in your Content Builder folder structure under PhraseeData.

This content block contains AMPscript, Marketing Cloud's proprietary scripting language, that powers the experiment's ability to dynamically adjust the subject lines your subscribers receive. This AMPscript both uses and records information in the data extensions created when you activated your experiment.

It's very important you don't modify anything in this code snippet. Your triggered experiment will not work correctly if the code is modified.

Data extensions

Every triggered experiment will generate four data extensions. The experiment uses these data extensions to store information it needs to reference for subject line content, testing proportions, performance metrics and send records. These data extensions will all be named after your experiment ID and the type of information they store:

  • <your_experiment_ID>-AR
    Used to store aggregated send information from your triggered experiment

  • <your_experiment_ID>-batches
    Used to store the proportions of your audience each variant should receive by batch

  • <your_experiment_ID>-recipients
    Used to keep track of which recipient received which variant and when

  • <your_experiment_ID>-variants
    Used to store the actual copy variants

By using these data extensions, the experiment is able to dynamically populate, add and remove content on a recurring basis. Phrasee is able to push updated information to these data extensions automatically during your testing period via API calls.

Just like the code snippet, the data extensions are stored in a folder named after your experiment. You can find it in your existing data extension folder structure under PhraseeData.

Query

In addition to using some API calls to update information in the data extensions, the triggered experiment creates a SQL query activity in Automation Studio for a similar purpose.

It's taking information from the -recipients data extension and aggregating it so Phrasee can see how many people were sent each variant without sending any personally identifiable information about your subscribers.

The query will be named like your data extensions were: <your_experiment_ID>_query.

This query will be in a folder named after your experiment. You can find it under the PhraseeQueries folder in the SQL Query activity section of Automation Studio.

And there you have it!

Using those assets, some API calls and the power of the Phrasee Brain, triggered experiments are able to continually optimize your subject lines. All you have to do is drop the code snippet in your emails, sit back and watch the magic happen.

Did this answer your question?