One-time sends

A one-time send is a single message that you send to a group of people at once. One-time sends make it easy to broadcast a single message to a wide audience at a predictable time without setting up an automation workflow.

How it works

A One-time send is a single message that you want to send to a group of people. You can schedule a one-time send or send it immediately, providing a simple way to broadcast a message to your audience without setting up a complex automation.

You might send a broadcast to alert your audience to changes in your terms of service, or to a group of people when you onboard people belonging to an account.

Set up a one-time send

The steps to set up a one-time send are fairly simple, but each step includes settings that we’ve explained in the sections below.

1. Go to Broadcasts and click Create Broadcast.
2. Pick One-time send.
3. Set up your message Content.
4. Determine who gets the message in the Recipients step.
5. Schedule your message.
6. Set a Goal.
7. Review and activate your message.

Content: write your message content

You can send a broadcast for any channel— email, in-app, push, SMS, or webhooks. Select your channel of choice to begin.

While most channels are self explanatory, you can select Webhook to “broadcast” requests to your webhook server for each member of your audience. You might do this to trigger in-app messages or messages from a service outside Customer.io. When you elect to broadcast webhooks, you can personalize the URL or payload for each member of your audience using liquid .

How you create your message is up to you. Choose the editor that works for you, use customer data, and even create variations to test!

Set up an A/B test

You can add an A/B test to evaluate variations of your message and find the one that most resonates with your audience. You’ll set up a test and send it to a sample percentage of people. After you send your broadcast, you can track variation performance and send the most successful version to your remaining recipients.

Click Add test in Content step to set up an A/B test. You can send your tests to a sample percentage, and automate sending to the rest. Adding a variation will copy your first message, so tweaking one variable of the message should be easy!

1. Set your Sample Percentage: this is the percentage of recipients in the test. After the test, you can send the best-performing variant to the rest of your audience.

   Let’s say you want 20% of recipients to receive one of two emails with different subject lines. A randomly selected 10% of people will receive each email, and you can choose to automatically send the one which performs better to the remaining 80% once the test is over. Start by setting the Sample percentage to 20% using the slider.

   A/B test variants are split evenly within the sample

   You can set the Sample Percentage anywhere from 1–100%. Within that sample, recipients are split evenly across your test variants (2 variants = 50/50 each; 4 variants = 25% each, and so on)—there’s no option to weight individual variants unevenly.

   

2. Click Add variation to create the messages you want to test. While you can test anything, we recommend that you only change one variable per test to produce a reliable test. You can add up to 8 variations.

   

   Use a statistically significant test audience

   Results are most reliable when you have at least 500 recipients per variation. With fewer recipients, your test may not produce Conclusive Results.

3. (Optional) Enable Automatically send winner to remaining recipients to send the winning option to your audience as soon as we decide the winner. Select the metric you want to use to decide the winner and how long the test should run before sending it to the remaining recipients.

   

Find the results of your test

When you send your broadcast, you can start to monitor the results of your test on the Test tab of the broadcast’s Overview page.

From here, you’ll see the current metrics for each variation, with the metric chosen to track the winner in bold. The Details section of the page shows the timeline for the test:

• When you first sent the message
• To what percentage of your recipients
• How much longer the test will run
• When the winning message will be sent to the remaining recipients.

Modify your test

If you want to change how the winning variant is evaluated, go to the Test tab.

Under Details, click Edit test.

After you turn off “Automatically send winner”, you can’t turn it back on. You have the option to send to your remaining recipients manually.

Calculate statistical significance

There are a number of statistical significance calculators you can use with the information provided. The example below uses Neil Patel’s calculator ↗.

If you’re testing different subject lines on your emails, you likely defined success as the email with the highest open rate. To determine if the difference in open rates is significant, go to the Test tab for the one-time send you’re testing. Enter your test result numbers into your calculator following the guide below:

What does the result mean?

Each calculator will have a slightly different way of communicating results, but this calculator uses the phrasing, “I am 80% certain that the changes in Test “A” will improve your conversion rate.” This mean there is a 20% chance that the higher open rates in variation A were just random chance and not a pattern.

How to increase the likelihood of statistical significance

It always feels great to declare a winner with high confidence. So how can you increase the number of times your tests are statistically significant? The math that determines significance is driven by two primary factors: sample size and test duration.

• Sample size
  • The larger the test’s sample size, the quicker it will achieve the number of email actions required to meet your desired level of confidence. With marketing messages, the quicker you can achieve valid results, the better. If you don’t have a large sample size, the test will likely need more time to achieve statistical significance (and if the sample size is too small, it may never get there).
  • A rule of thumb is a sample size of at least 500 recipients per variant, but it’s worth taking the time to calculate sample sizes specific to your needs and confidence levels using a sample size calculator ↗.
• Test duration: Although it’s not used in statistical significance calculation, the duration of a test impacts confidence in the results because the more time people have to act on the message (open it, click it, etc), the more information you have for calculating the results. When planning how long the test should run, first consider the likelihood that people will complete action the action. For example, the amount of time it takes to gain confidence in the results of a test measuring open rate is lower than for one measuring conversions.

Send to remaining recipients manually

If Automatically send winner to remaining recipients is off, you can manually determine the “winning” test and send it to your audience.

When you have conclusive test results, go to the Test tab and click Send to remaining. Then choose the variant you want to send and when.

Recipients

You can send a broadcast to everyone in your workspace or narrow your audience down:

• Specify conditions they need to match
• Add a list of recipients via CSV

Check the estimated count of recipients to make sure the volume matches your expectation. Click View list to see the list of expected recipients.

People matching conditions

On the Recipients step, you decide who will receive your message. If you select, “People matching conditions,” then you can define who will receive the message based on one or more segment, attribute, or message conditions.

Segment condition

People will receive your message if they belong or don’t belong to specific segments. Filter for people IN or NOT IN segments and join conditions using AND or OR logic.

Here’s a walkthrough of setting up a segment condition:

Attribute condition

People will receive your message if their profile attributes meet these conditions. You can filter for people whose attributes exist, don’t exist, are equal to, or are not equal to a value.

One-time send condition

People will receive your message if they previously received one or more messages.

Message condition

People will receive your message if they meet conditions related to any message in your workspace, like:

• people have never been sent any in-app message
• people have ever been sent the email “Thank You email” in “Anniversary Automation”

Import recipients

You might import recipient via CSV for the following reasons:

• You want to send the message as quickly as possible—assigning recipients by CSV upload means you bypass the time it takes to calculate audience conditions.
• You can’t narrow down your complete audience by shared conditions.

Instead of choosing to narrow your audience down by conditions, select Upload a CSV.

Your CSV must follow these rules:

• Size: no larger than 100 MB.
• Contains a single email or id column that identifies people who already exist in your workspace. You can’t add new people to your workspace through this CSV. Learn more about importing new people.

The import ignores any other columns; you can’t update your audience’s data through this upload. The import also skips rows that don’t match an existing person, instead of failing the import altogether.

Email: Skip duplicate addresses

If your workspace is ID only, multiple people in your workspace can have the same email attribute . The Skip duplicate email addresses setting prevents you from sending your email broadcast multiple times to the same address. It is on by default.

At your broadcast’s send time, we generate a delivery for each member of your audience. If this setting is on and a subsequent member of your audience has the same email attribute as someone who already generated a delivery, we will skip that person.

This setting only prevents duplicate deliveries; we do not affect your audience in any other way. For example, we don’t modify the Is a recipient segment criteria for your message—all of the people in the original audience, including people with duplicate email addresses who were skipped, will join the Is a recipient segment. This helps prevent you from accidentally sending repeat or follow-up emails to someone who already received your message.

While this setting prevents you from sending duplicate emails, you can deduplicate people in your workspace by enabling the email or ID setting, which makes email a unique attribute. This ensures that multiple people cannot have the same email address attribute.

SMS/WhatsApp: Skip phone number duplicates

The phone attribute isn’t required to be unique in Customer.io. If more than one person in Customer.io has the same phone number, you’ll end up sending duplicate outbound messages.

When you enable the Skip duplicate phone numbers setting, and we send a message to a phone number, we’ll ignore any subsequent messages to that phone number.

Set a subscription preference

If you’re using our Subscription Center feature, you’ll set your broadcast’s topic in the Subscription preference setting. If people are unsubscribed from the topic, they won’t get messages from the broadcast.

If you don’t want to set a topic, you can use the All subscribed and unsubscribed setting. You should use this setting sparingly—for things like transactional-style messages and important notices. Sending messages to unsubscribed people can violate their trust—or even violate local laws and regulations (GDPR, CAN-SPAM, etc)!

Send to unsubscribed users

You can opt to send your message to unsubscribed users—which you might do to notify users of important information like a change to your terms of service or scheduled maintenance. Go to the Recipients tab under Sending options:

Schedule your message

You can send your broadcast immediately or schedule it to send for a future date/time.

Click Schedule for later to set a future time. You should always schedule your message at least 5 minutes in advance, and note that the timezone defaults to your local time.

Send in recipient’s timezone

When you schedule a broadcast to send later, you can send it at a specified time in each recipient’s timezone. Learn more about sending in users’ timezones.

We calculate who receives your message at send time, so the audience count you see under Recipients when you schedule the message could differ from the final count we send to later.

You can’t schedule a timezone-matched one-time send while one of its segments is still building. Wait for the segment to finish building and then schedule your message.

Send while a segment is still building

If your recipients include a segment that’s still building, you can still send or schedule your one-time send. Customer.io holds the message until its audience segments finish building, and sends when the segment is ready. While Customer.io holds your message, the one-time send shows as Queued.

If you try to send immediately and a segment is still building, the confirmation shows a provisional recipient count and a Send when ready button. The final count can differ, because Customer.io uses your segment to calculate the full audience for your message.

Customer.io checks your segments 2 minutes after the hold begins, doubling in time after each check up to 30 minutes between checks. For a scheduled send, the hold begins at your scheduled time, so your message can go out later than the time you set—if your segment is still building.

Customer.io stops waiting if your segments still haven’t finished building 6 hours after the hold began. It also stops waiting if a CSV import into one of the segments fails. If you delete that import or upload its CSV again, Customer.io also stops waiting. What happens next depends on how you sent the message:

• Send now: The status of the one-time send changes to Failed, and Customer.io doesn’t send any messages. To try again, duplicate the one-time send and send the copy after the segment finishes building.
• Scheduled: The one-time send goes back to Draft without showing a failure reason. Check your segments, then schedule your message again.

Customer.io doesn’t hold test sends, or one-time sends to imported recipients.

Limit send rate

Limiting the send rate is optional. If you don’t set a send rate, Customer.io sends your broadcast as fast as it can, unless new-account sending limits apply to your account.

Setting a maximum rate limit can help you maintain good deliverability. When you limit the send rate, we queue messages for delivery in batch sizes of the number you specify.

On the Schedule tab, turn on Limit send rate. You can choose between two rate limit types: a Fixed rate or a Daily ramp.

Fixed rate

Choose Fixed rate if you want to send a consistent maximum number of messages per minute, hour, or day.

1. Specify your Limit: the maximum number of messages and the time period.

2. Choose your Send mode: send evenly or send as fast as possible.

   Sending evenly across each period is the default to protect your downstream systems from getting overloaded. However, you might choose Send as fast as possible up to the limit in each period if your content is time-sensitive or you have automation conditions dependent on this message.

You can use a fixed rate and send your messages in the recipient’s timezone. Each timezone has a local send window—about an hour around the target local time.

If your broadcast sends in a recipient’s timezone AND has a daily fixed rate limit, any messages that fall outside the hour-long timezone window or surpass the rate limit are punted to the next day at the same local time in their timezone. This can repeat across days if the daily cap is too small to drain that timezone in one window. To reach people faster and prevent this spread, increase your daily limit.

When you send in a recipient’s timezone, a per-day fixed rate requires the Send evenly across each period mode. You can’t use a daily rate with the Send as fast as possible option in timezone mode. Per-minute and per-hour rates work with either send mode.

Daily ramp

Choose Daily ramp to exponentially increase your send volume over multiple days. This is ideal for domain warming; it automates the manual warm-up schedule so you don’t have to create multiple messages for each day.

When you select Daily ramp, configure the following settings:

• Starting Daily Volume: The number of messages to send on the first day. Start with a low number like 100. You can learn more about the recommended ramp-up rates in Domain warming.

• Target Daily Volume: Set a target that reflects the maximum number of messages you’ll expect to send on a daily basis. This is the volume the message will ramp up to sending by the end of the period. This is especially important if you’re using this message to warm up your domain. You don’t need to warm it up past your max daily send rate.

• Number of Days: How many days the ramp lasts (up to 60). The daily volume increases each day until the ramp period ends or all recipients receive the message. We’ll send across the day, too, not all at once.

  If recipients remain after the last day of your ramp, we continue sending at the final day’s rate until all recipients receive the message.

You can use a daily ramp and send your message in the recipient’s timezone. We only send to recipients during their local send window. If a day’s ramp volume runs out before a timezone’s window opens, those recipients roll over to the next day at their local time rather than sending outside their window.

Preview the chart to see the projected daily send volume and make sure your adhering to best practices for ramping. We’ll warn you if your daily ramp might not be optimized too.

After the ramp completes and all recipients have received the message, the broadcast’s status changes to Sent—just like any other message.

Rate limit constraints

• One-time sends must finish sending within 60 days. If your rate limit settings would take longer than 60 days to reach all recipients, you can’t activate the message send—adjust your settings accordingly.
• If you pause and resume a one-time send during a ramp, it picks up where it left off based on how many messages have already been sent—it doesn’t skip days.

Edit rate limit while sending

You may want to change the rate limit while your one-time send is sending to improve your domain warming outcomes. For instance, if your emails are getting deferred or engagement rates are lower than expected, you may want to slow your send rate. This allows mailbox providers more time to evaluate the traffic.

When you update a daily ramp, the ramp starts over. The day you save your changes is the first day of the new ramp, and Number of Days counts from that day, not from when the message was originally scheduled to send.

1. From your one-time send, click the Recipients tab.
2. Click Edit rate limit. To edit the rate limit, your message must follow these conditions:
   • The message state must be Sending.
   • The message must have a configured rate limit. You can’t add a rate limit if the message is sending as fast as possible with no limit.
3. Modify the settings as you see fit. You can change between Fixed rate, Daily ramp, or choose Unlimited to remove the limit.
4. Click Save changes.

The daily schedule will show the daily rate for the messages that have already been sent followed by the estimated schedule for the new rate limit.

Your one-time send keeps sending on the day you save your changes. Customer.io doesn’t count the messages you sent earlier that day against the new ramp, so the rest of the day sends at the new ramp’s first-day rate.

Check your send rate in Reports

The Send volume panel on the Reports tab of your one-time send charts how many messages the send sent over time. When a limit applies to your send, the chart draws it as a dashed line:

• Rate limit is the most messages your fixed rate or daily ramp allows in each bar of the chart.
• Deliverability limit can appear while an email send is in progress. It shows your account’s sending limit for new accounts when that limit is lower than your own rate limit, or when your send has no rate limit of its own.

When your sends reach a dashed line and stay flat, that limit is holding your send back. If it’s the Rate limit line, you can edit the rate limit to send faster. If it’s the Deliverability limit line, your account’s sending limit for new accounts is the cap, and editing your rate limit won’t speed the send up. The chart draws these lines only for a time range longer than 2 hours, so you might need to pick a longer range to see them.

If your one-time send went out before September 25, 2026, the panel shows an hourly chart of your sends instead.

Goal: determine what makes your message successful

You track the success of your one-time send based on whether a person performs an event, enters a segment, or exits a segment after they receive your message.

If a person performs your goal criteria within the time frame you configure, the message is marked converted. You can set goals to determine the success of your messages, and follow conversion rates over time to improve your message strategies. See Goals for more information about conversions; one-time sends have the same goal options as automations.

We track conversions for the following message/delivery types:

Slack and Create or update person actions are often internal or used for analytics purposes; they don’t always send messages to end-users. For that reason, we don’t attach conversions to them. You can enable webhook conversions on individual webhook actions.

In the Goal section of the workflow, you can decide if you want to track opens and clicks (on by default).

Review your one-time send

After you set up your one-time send, go to the Review tab to activate your message.

Check the high-level view of your settings, content, and schedule. When you’re ready to send, click Send or Schedule at the top.

Pause and resume a one-time send

You may want to pause a one-time send when you notice a typo after sending or need to make some other change to your message. You can pause if the message hasn’t finished sending to all of your recipients. Any messages already sent cannot be recalled, though.

1. In Broadcasts, click a one-time send with a status of Sending. (If the status is “Sent,” then all recipients have received the message and it cannot be paused.)

   

2. In the top-right, click Actions then select Pause sending from the dropdown. This action immediately pauses any further messages from sending.

   

   Back on the Broadcasts landing page, the broadcast’s status will change to Paused.

   

3. Go to the Content tab, click Edit content and make your changes.
   • If editing a message with A/B variations, choose the variation then proceed to edit.
   • You can modify any message content (email, in-app, etc) or field (subject, to, from, etc). You can’t modify your recipient conditions, tracking parameters, goals, or send options (rate limiting).
4. After you make your changes, click Resume sending in the top-right of the one-time send then confirm your action. The message will continue sending to people who have not received the message yet; people who already received the message will NOT receive another version.

   

   The broadcast’s status will update to “Sending” then “Sent” when all recipients have received it.

Send to an individual person

After you schedule and send your message, you can send the content to an individual person who wasn’t in your original audience.

You can only send your message to an individual this way. If you want to re-send it to a large number of new recipients, you should copy the one-time send and re-send it to a new segment of people.

1. Go to Broadcasts and select the broadcast you want to re-send. The broadcast must have a sent status.
2. Go to the Content tab.
3. Select the Recipient you want to send your message to. The recipient must be a person already in your workspace; you cannot send a broadcast to a person you haven’t identified.
4. Click Send newsletter to the recipient.

   

Copy one-time sends to another workspace

If you use similar messages in different workspaces, you can copy a broadcast or a one-time send to another workspace. Go to Broadcasts, find the one-time send or broadcast that you want to copy, and click Copy to. Then select the workspace you want to copy your items to, and click Continue.

Note that when you copy one-time sends and broadcasts, we’ll reset settings that don’t exist in the destination workspace—like layout, audience, etc. You should also check your liquid syntax in your destination workspace. You may not have the same attributes or other variables in your destination, so you may need to update liquid statements to make sure that your messages render and send properly.