.svg)
A multichannel programme is a set of delivery controls as much as a set of messages. Frequency caps decide which sends survive when several qualify at once, and a capped send in a Canvas does not stop the journey. Design the caps alongside the journey or your sequence has holes nobody can see.
Key Takeaways
- Global frequency capping covers push, email, SMS/MMS/RCS, webhook, WhatsApp and LINE, and does not apply to triggered in-app messages.
- Caps count per dispatch, so a campaign or Canvas step send counts once however many variants or platforms sit inside it.
- A user who is capped at a Canvas Message step skips that message and still advances to the next step.
- You find the sends that never happened with message aborted events or a Segment Extension, so decide which before launch.
- When sends compete in the same window, the ones processed first count toward the cap, which isn't the same as the most important ones.
- A send throttle can be changed on an active Canvas, and the new limit applies to new messages.
Which channels does a global cap cover?
Global frequency capping applies to six channels:
- push
- SMS/MMS/RCS
- webhook
- LINE
It does not apply to triggered in-app messages.
That split is the first thing to settle when you build across channels. In-app messages and the capped channels follow different rules, so a user who has reached their cap can still receive an in-app message.
Used deliberately, that's a design tool. When the capped channels are exhausted, the app is still available for the same message, and it reaches someone who is already in the app.
What counts as one send?
Frequency capping applies per dispatch. Each campaign or Canvas step send counts toward your caps, not each variant or platform inside that send.
So a multivariate campaign is one dispatch, not one per variant.
The counting window matters too. Global frequency capping uses the user's time zone and counts by calendar day, not by rolling 24-hour periods. As an example, a cap of one a day would permit a message late at night and another shortly after midnight.
We set caps against what a person will feel as well as against dispatch counts, because the two can differ. That's our approach, and Braze doesn't frame it that way.
Why did a Canvas step send nothing?
This is the behaviour to build around. A user who reaches a Canvas Message step while over the global cap doesn't receive that message, and still moves on to the next step.
Braze documents it directly: Message steps advance users when a message isn't sent because of global frequency capping.
In our view that's a sensible default, because holding people in a step until a cap resets would stall journeys. It does mean a sequence can lose its middle and still reach its end, and the missing send is easy to overlook unless you go looking for it.
So we plan for it on any Canvas that carries a capped channel. Decide before launch how you'll detect the gaps, and what the journey should do when a message is skipped. That's our practice rather than a Braze recommendation.
How do you find the sends that never happened?
Braze points to two routes.
- Track message aborted events where
abort_typeisfrequency_capped. - Create a Segment Extension to analyse users who entered the Canvas but didn't receive the expected message.
Pick one before launch and wire it up, so the check is already running when the first capped send happens rather than being built afterwards.
Webhooks have their own rule. A webhook counts toward the cap when Braze records a successful delivery, and unsuccessful responses, such as 4xx or 5xx status codes, don't count. A failing endpoint therefore uses none of the user's allowance while delivering nothing.
Which message wins when several qualify?
When multiple sends compete in the same window, the messages processed first are the ones that count toward the cap.
Processing order isn't editorial priority. If a promotional send and a service reminder both qualify on the same morning, the cap has no way of knowing which you would have chosen.
That's why we treat a cap as a blunt control on its own and put the sequencing work upstream. Decide what may compete before you decide how many may get through.
| Control | Scope | What it decides |
|---|---|---|
| Global frequency cap | The six capped channels | How many dispatches a user can receive in the capped window |
| Canvas step advancement | Any Canvas Message step | Whether a capped user continues the journey (they do) |
| Send throttle | A single Canvas | The rate at which messages leave, adjustable on an active Canvas |
When you increase or decrease a Canvas rate limit, the updated limit takes effect for new messages, and there may be a brief delay before the change shows across the Canvas. Throttling is therefore the control to reach for when a live Canvas needs slowing down.
How does this change campaign design?
Caps are an editorial decision expressed as a number, so the number should follow the plan.
Onboarding is where the tension shows first. An onboarding journey in Canvas sends more in week one than most programmes, and the activation campaigns running beside it compete for the same allowance. Both are sends you least want dropped.
Retention work sits at the other end, where a win-back message losing to a promotional send is a poor trade. And because caps limit how many messages arrive rather than how good they are, personalisation and dynamic content are what make a smaller number of sends worth receiving.
Frequently Asked Questions
1. Do in-app messages count toward a global frequency cap?
No. Global frequency capping covers push, email, SMS/MMS/RCS, webhook, WhatsApp and LINE, and leaves triggered in-app messages outside it.
2. Does a frequency-capped user drop out of a Canvas?
No. They don't receive that step's message, but Message steps advance users when a send is withheld by global frequency capping, so they continue through the journey.
3. Does each variant in an A/B test count separately?
No. Frequency capping applies per dispatch, so a campaign or Canvas step send counts once, however many variants or platforms sit inside it.
4. How do I see which users were capped?
Track message aborted events where abort_type is frequency_capped, or create a Segment Extension to analyse users who entered the Canvas but didn't receive the expected message.
5. Can I change a send throttle while a Canvas is running?
Yes. The updated limit takes effect for new messages, and there may be a brief delay before the change shows across the Canvas.
Sources
- Braze, Frequency capping FAQ. Retrieved 01 Oct 2026.
- This post draws on a single documentation page. Treat it as a reading of that page plus CustomerIK judgement rather than a survey.
- Written by CustomerIK, a Braze implementation partner.
CustomerIK is a Braze implementation partner working across onboarding, technical integration, marketing operations and customer data management.
If your journeys are losing steps and you can't see where, let's talk.
.svg)
