Leveraging interaction data
Connect certain Destinations as Sources to leverage additional signals about customer engagement.
Time to read:
Leveraging interaction data
Collecting data from your websites, apps, and servers is only part of the equation. Your users are giving you valuable signals when they open your emails, click on embedded links, start livechat conversations, answer a survey, or open a push notification. This lesson will show you how to collect and leverage interaction data to make your data-driven campaigns more intelligent.
Setting up Source integrations
There are a subset of Destinations on the Twilio Segment platform that also send data back into Twilio Segment. You can find these under Cloud Apps in our sources documentation. There are two types of cloud app sources, event sources, and object sources.
Event Sources
The setup for event sources usually involves creating a source in Twilio Segment, and then copying the write key into a partner's UI. The events that partners send back to Twilio Segment will depend on their product, but typically this will involve events from one of our email, livechat, or a/b testing specs. For example, Braze, a marketing automation platform that supports push notifications, in-app messages, and email supports sending the following events back into Twilio Segment:
Event Name |
Description |
---|---|
Application Uninstalled |
User uninstalled the App. |
Email Sent |
An email was successfully sent |
Email Delivered |
An email was successfully delivered to a User's mail server. |
Email Opened |
User opened an email. |
Email Link Clicked |
User clicked a link in an email. Email click tracking must be enabled. |
Email Bounced |
Braze attempted to send an email, but the User's receiving mail server did not accept it. |
Email Marked As Spam |
User marked an email as spam. |
Unsubscribed |
User clicked the unsubscribe link in an email. |
Push Notification Sent |
A push notification was successfully sent |
Push Notification Tapped |
User tapped on a push notification. |
Push Notification Bounced |
Braze was not able to send a push notification to this User. |
In-App Message Viewed |
User viewed an in-app message. |
In-App Message Clicked |
User tapped or clicked a button in an in-app message. |
News Feed Viewed |
User viewed the native Braze News Feed. |
News Feed Card Viewed |
User viewed a Card within the native Braze News Feed. |
News Feed Card Clicked |
User tapped or clicked on a Card within the native Braze News Feed. |
Each track event will also contain the following event properties:
Property Name |
Type |
Description |
---|---|---|
|
string |
The API Identifier of the App on which a user received a message or performed an action, if applicable. |
|
string |
The API Identifier of the Campaign associated with the event, if applicable. |
|
string |
The API Identifier of the Canvas associated with the event, if applicable. |
|
string |
The API Identifier of the Canvas Variation associated with the event, if applicable. |
|
string |
The API Identifier of the Canvas Step associated with the event, if applicable. |
|
string |
For Email events, the email address that the email was sent to. |
|
string |
For In-App Message Clicked events, the ID of the button the user clicked on. |
|
string |
For News Feed Card Viewed and News Feed Card Clicked events, the API Identifier of the News Feed Card. |
Once you've connected the source, send a test email and consult the Debugger in Twilio Segment to see the events coming through.
Object Sources
Object data sources will pull data from our cloud app partners and write them into a data warehouse, like Redshift or BigQuery. This could be non-user data like subscriptions in Stripe, accounts in Salesforce, tickets in Zendesk etc. The main use case for this kind of data is to do analytics and reporting based off of this data. Having this all in one place means you can make dashboards that combine these different cloud apps, along with the first party data you're collecting from your website, mobile apps. Our customers will often using a BI tool like Looker, Mode, or Periscope, that sits on top of this data warehouse.
There are some use cases where customers want to get the object data back into an event destination. For example, you might want to get a list of all users that have an open Zendesk ticket, and then push that list to your advertising and email tools to exclude them from marketing emails. Trying to upsell your customer when they're having a bad support experience is a great way to lose your customer's trust in your product or service.
To import this kind of data, we have a feature called SQL Traits part of our Personas product which will help you accomplish this. Please consult the lesson on leveraging SQL traits to learn more.


Using Source data
Once you have your cloud app events flowing back into Twilio Segment, it's time to leverage this data. The first use case we commonly see is for analytics. Marketing automation tools typically have some kind of reporting capability, but it's usually not their primary focus. To get more granular on understanding how your campaigns are performing, or how an email is doing, you want to be leveraging one of the analytics tools on our platform, or building your dashboard based on the events we're pushing into your data warehouse. Here's an example email funnel you can quickly build in Amplitude.


On the customer-facing side of the business, you might want to leverage this interaction data across your support, sales, and marketing teams. A good example is treating customers that are opening your emails differently from customers that haven't interacted at all. For example, you could create an audience of users that haven't purchased or visited your website in the last 6 months, yet have still opened an email from your team. If you are using Twilio Segment Personas, you can also then push this audience directly into an advertising platform like Facebook or Google Ads. You will get a better CPA on a re-engagment ad campaign if you include the email interaction data and filter out users that are completely inactive.



