A campaign can submit requests faster than its delivery path processes messages. Planning the send requires a capacity estimate for each stage.
How does API rate limiting differ from throughput?
An API rate limit controls requests. Delivery throughput controls how quickly a later stage can process messages or segments.
When Bird rejects a request with 429, follow the response headers and the rate-limit guide. Retrying immediately can repeat the same rejection.
Successful submission has a different meaning. A 202 confirms acceptance, while charging, network submission and delivery occur afterward. Track the message identifier through those outcomes.
Twilio queues messages when submissions exceed the sending rate and reports queue-overflow errors when its capacity is exceeded: “Customers sending large volumes of messages may encounter errors such as Queue Overflow”.
What determines the available rate?
The effective rate depends on the sender, destination, registration, provider configuration and applicable account limits.
A short code and a long code can have different capacity in the same destination. Registration can affect the traffic a sender may carry. Check the sender choices and destination requirements before planning volume.
A published maximum is a capacity limit, not a delivery promise. Confirm the unit, scope and conditions before applying it to a campaign.
How do I estimate the sending time?
Divide the amount of work by a rate measured in the same unit. A segment is the text carried in one network message. Longer texts use multiple segments.
For an illustrative route limited to 100 segments per second, 6,000 one-segment messages require at least 60 seconds of capacity. If each message needs two segments, the same work requires at least 120 seconds.
Those calculations assume continuous use of the stated capacity. Other traffic, pauses, retries and downstream conditions can extend the time. They do not predict when every recipient reads or receives the message.
Check personalized bodies with the segment calculator. A longer name or an emoji can change the amount of billable text.
Why are messages delayed when my requests succeed?
Queueing, downstream processing, network conditions and recipient availability can all separate acceptance from delivery.
Compare the accepted, submitted and reported delivery times. Inspect errors by sender and destination. A missing receipt does not by itself identify the cause.
An SMS message center can retain a message while the phone is unavailable. What is an SMSC explains that stage. SMS analytics explains how to investigate outcomes in Bird.
Do not resubmit accepted messages merely because a delivery receipt is late. That can create duplicate texts without resolving the delay.
How should I prepare a larger campaign?
Choose a sender and capacity plan that fit the traffic, then test a limited release before expanding it.
Discuss the countries, segment volume and required release window with the provider. Registration and sender preparation may be needed before a higher-volume rollout.
Spreading traffic across additional numbers to evade carrier limits is known as snowshoeing. The CTIA principles address that practice.
- Count rendered segments by destination.
- Confirm the rate, unit and scope for each sending path.
- Allow time for processing and downstream variation.
- Release a limited group, inspect outcomes and expand deliberately.
In short
API capacity and delivery capacity measure different stages.
A successful request does not establish that every message can immediately leave for the network.
Check the unit behind the limit.
Requests, messages and segments produce different capacity calculations.
Delayed delivery has several possible causes.
Inspect queueing, errors and delivery events before attributing a delay to throughput.
Plan capacity for the sender and destination.
Account limits, registration and carrier rules can also constrain the sending path.