This page is part of the Customer.io documentation. For the complete documentation index, see llms.txt.

30

Start sending automations and workflows

How it works

A message is what you want to send. But before you send a message, you have to determine who to send your message to and when to send it. You answer these questions when you set up an automation, broadcast, one-time send, or transactional message.

Each of these mechanisms has a different use case, but they all contain the same basic concepts/questions:

• How do I trigger messages?
• Who should receive messages?
• When should people get messages?
• What messages should people get?

When should I use automations, broadcasts, or one-time sends?

We’re glad you asked! There are a multitude of factors that can help you pick the right message mechanism for your needs. But, in general, you can use the following guidelines to help you pick the right one.

I want to control when my audience gets messages: You should send a one-time send or a broadcast. You’ll trigger one-time sends manually (or on a schedule), which makes it easy to determine who gets messages and when. Broadcasts are a bit trickier, but you base the send on an API call that you control—so you can send messages when something relevant to your business happens.

I want to send messages in response to my audience’s behaviors: You should send a automation or a transactional message. These things are very different, but they’re both predicated on your audience’s behaviors. You’ll send transactional messages when a profile performs an action that expects a message—like a receipt or a password reset. You’ll send automations when profiles meet certain criteria, like when they perform an event, gain certain attributesA key-value pair that you associate with a person or an object—like a person’s name, the date they were created in your workspace, or a company’s billing date etc. Use attributes to target people and personalize messages., and so on.

If you want to send multiple messages or perform multiple actions: you should send a automation or a broadcast. Automations contain a workflow that you can send profiles through to get multiple messages, or perform multiple actionsA block in an automation workflow—like a message, delay, or attribute change. over time. Broadcasts also have a workflow, but the whole workflow happens at once; this means broadcasts aren’t really ideal for sending multiple messages, but you can perform multiple actions—like changing your audience’s attributes or sending webhooks.

Try a one-time send if you’re just getting started

Many of the sending mechanisms in Customer.io rely on your data. If you’re just getting started, a One-time send, like the one in our quick start guide, is a great way to get started with Customer.io.

Where other kinds of messages rely on your data, API calls, etc, you can control all aspects of a One-time send from our UI. That can make it easy to test your implementation and get a feel for how our platform works.

Automations

Automations are the most powerful and common way to automate your messaging strategy with Customer.io. When profiles enter an automation, they start a journeyTypically, a person’s path through your automation. If the automation is triggered by a webhook, then a journey captures the webhook’s path, not a person’s. through your automation workflow. Your workflow can define different paths profiles can go down and different messages they can receive based on the actions they take and the attributes they have.

An automation consists of a trigger, workflow, and goal. When profiles meet your trigger criteria, they start a journeyTypically, a person’s path through your automation. If the automation is triggered by a webhook, then a journey captures the webhook’s path, not a person’s. through your workflow. They travel through the workflow until they:

1. Finish the workflow
2. Match your goal criteria (depending on your settings)
3. Stop matching the trigger criteria (depending on your settings)

Because profiles can match your trigger at different times and exit automations for different reasons, each profile’s journey is unique and personalized for them. This makes it easy for you to maintain an automated messaging strategy that resonates with your individual users.

Automation triggers

The trigger determines when profiles start an automation. You can set up automations to trigger when profiles perform an event, move into (or out of) a segmentA group of people who match a series of conditions. People enter and exit the segment automatically when they match or stop matching conditions., they reach a certain date, and so on.

In many cases, understanding the purpose of your automation and the kinds of messages you want to send can help you determine the trigger you want to use. For example, you might use date-triggered automations for things like birthday or anniversary messages. You might use event-triggered automations to send coupons encouraging your audience to complete their purchases when they abandon their shopping carts. Learn more about automation triggers.

But to set up automation triggers correctly, you’ll also need to know a bit about your data—what attributes and events are available to you. You can see the attributes and events in your system under Integrations > Data Index.

Workflows

The workflow is where you determine what messages profiles get and when. In a workflow, you can send profiles down different branches based on their attributes or behaviors. With a complex workflow, you can automatically send messages and respond to your audience in ways that feel personal and relevant to them.

For example, we have a workflow below based on a cart abandonment scenario with three paths. If a profile abandoned their cart for a certain amount of time, we send them a reminder message encouraging them to finish their purchase. But we don’t want to send them this email if they finish their purchase or they’re still shopping, so we bypass this message if the profile completes their transaction or they add new items to their cart!

Automation workflows also help you send messages at the right times with blocks like Wait until. These blocks let you delay a message for a set period of time, or until a customer does something relevant. For example, you might send a series of emails explaining how to use your platform when someone joins your service. You can set up Wait until blocks to send follow-up emails when a profile completes each step in the setup process.

Goals

Goals help you measure the success of your automations. Typically, we represent success when a profile performs an event or gains an attribute—like when someone completes a purchase or renews their subscription. If someone reaches your goal criteria in response to a part of your automation, we record a conversion, so you can see how effective your automation is.

Exit criteria

Exit criteria determine when someone ends their journey through the automation. For most automations, profiles will by default exit if they no longer match the trigger’s filter criteria; for segment-triggered automations, the default is when profiles no longer match the filter criteria OR the trigger conditions. Filters are additional criteria that profiles must meet to join your automation, giving you more granular control over who travels through your automation. We evaluate filters when someone enters an automation and before actions that affect them (like messages).

For example, if the purpose of your automation is for someone to buy something, you probably want to stop sending messages after someone actually completes the purchase. You could make the goal of the automation: “performs event purchase,” and your exit criteria They match the conversion criteria.

The options for exit criteria include:

• They match the conversion criteria
• They stop matching the filters (or trigger if it’s a segment-triggered automation)
• They match the conversion criteria or they stop matching the filters (or trigger if it’s a segment-triggered automation)

Broadcasts and one-time sends

Broadcasts and one-time sends are grouped together in Customer.io under Broadcasts. You determine whether you’re going to send an API-triggered broadcast or a One-time send when you set up your message.

Unlike automations, you trigger broadcasts and one-time sends—they aren’t triggered by your audience. And, where automations can run for a long period of time and profiles can start or end automation journeys independently of each other, everybody in your broadcast or one-time send audience gets your messages at roughly the same time.

API-triggered broadcasts

API-triggered broadcasts let you trigger the broadcast with an API call—which helps you send messages to your audience when something relevant happens. For example, you might trigger a broadcast to your audience when a new product launches, a new coupon becomes available, or when tickets become available for a concert.

Unlike one-time sends, you can set up a workflow for your broadcast and perform multiple actions. The workflow in a broadcast is designed to help you send a message and perform other non-message actions, like updating a profile’s attributesA key-value pair that you associate with a person or an object—like a person’s name, the date they were created in your workspace, or a company’s billing date etc. Use attributes to target people and personalize messages., sending webhooks, or adding profiles to segmentsA segment is a group of people in your workspace. Use segments to trigger automations, track membership over time, or fine-tune your audience. There are two types of segments: data-driven and manual. Data-driven segments automatically update when people start or stop matching criteria. Manual segments are static..

In general, you shouldn’t send multiple messages in a broadcast because when you trigger a broadcast, the whole workflow happens at once. If you send multiple messages as a part of your broadcast workflow, a profile will get them at roughly the same time.

Like other messages, you can personalize broadcasts using your audience’s attributesA key-value pair that you associate with a person or an object—like a person’s name, the date they were created in your workspace, or a company’s billing date etc. Use attributes to target people and personalize messages.. But, you can also pass data in your API call that you can use to personalize your messages. For example, you might send a broadcast to let profiles know that a new product is available, and you can use the product’s name in your message. Below is an example API trigger.

curl --request POST \
  --url https://api.customer.io/v1/campaigns/{broadcast_id}/triggers \
  --header 'Authorization: Bearer REPLACE_BEARER_TOKEN' \
  --header 'content-type: application/json' \
  --data-raw '
    {
        "data": {
            "headline": "Roadrunner spotted in Albuquerque!",
            "date": 1511315635,
            "text": "We received reports of a roadrunner in your immediate area! Head to your dashboard to view more information!"
        },
        "email_add_duplicates": false,
        "email_ignore_missing": false,
        "id_ignore_missing": false,
    }'

One-time sends

One-time sends are the easiest way to send a single message to a large group of profiles. You can target a message to segmentsA segment is a group of people in your workspace. Use segments to trigger automations, track membership over time, or fine-tune your audience. There are two types of segments: data-driven and manual. Data-driven segments automatically update when people start or stop matching criteria. Manual segments are static. or a list of profiles. Unlike automations, which profiles enter when they meet certain criteria, or broadcasts, which you trigger with an API call, you trigger one-time sends manually—which makes them predictable and ideal for one-off communications to large groups.

You can also A/B test one-time sends to improve your communications over time. While you can also A/B test messages in automations, one-time sends have a much easier time reaching statistical significance: it’s much easier to tell which of your A/B test messages worked better when you send them all at once.

Transactional messages

Transactional messages are messages that you send to a specific profile when they perform an action that expects (or demands) a message—like a receipt when a person makes a purchase or a link to reset a password when someone has trouble logging into your service.

You’ll send a transactional message via API call, similar to the way you trigger a broadcast, but each trigger is unique to a specific profile. It’s a one-to-one message.

Like other messages, you can create your transactional message in our UI and personalize it with your audience’s attributesA key-value pair that you associate with a person or an object—like a person’s name, the date they were created in your workspace, or a company’s billing date etc. Use attributes to target people and personalize messages. and other data from the API call that triggers the message. But, if you want to maintain complete control of your message, you can also send the entire raw message in your API call if you want. This gives you complete, programmatic control over the message you send—the content, the audience, and the personalize data you send.

Here’s an example of a transactional message payload:

{
  "transactional_message_id": 44,
  "to": "cool.person@example.com",
  "subject": "Reset your password",
  "identifiers": {
    "email": "cool.person@example.com"
  },
  "message_data": {
    "password_reset_token": "abcde-12345-fghij-d888",
    "account_id": "123dj"
  },
  "cc": "cc@example.com",
  "bcc": "bcc@example.com",
  "disable_message_retention": true,
  "send_to_unsubscribed": true,
  "queue_draft": false
}

Personalizing messages with liquid

As you write your messages, you can use your audience’s attributesA key-value pair that you associate with a person or an object—like a person’s name, the date they were created in your workspace, or a company’s billing date etc. Use attributes to target people and personalize messages., event properties, and other data to personalize your messages. This helps you generate highly relevant, personalized communications with your audience at scale.

The message personalization syntax we use is called liquidA syntax that supports variables, letting you personalize messages for your audience. For example, if you want to reference a person’s first name, you might use the variable {{customer.first_name}}..

For example, if you wanted to reference a profile’s first name in a message, you might use something like this:

{{customer.first_name | capitalize }}

You can use liquid almost anywhere in your messages. For example, when you write an email, you can use liquid to personalize your subject line, the recipient’s email address, and the content of your email.