Deliverability is the share of the messages you send that actually reach a phone, as opposed to being filtered, blocked or silently dropped on the way. It is decided less by the network than by the paperwork: whether the sender is registered, how the brand was scored, which carriers the provider connects to directly, and what happens when a phone cannot take the message format you sent.
Deliverability is not the delivery rate
The two are often used as one word, and they are not. The delivery rate is a measurement: the percentage of a given batch that returned a delivered receipt. Deliverability is the underlying condition that produces that number, the registration, reputation and routing that decide whether the next batch will get through. A batch can show a healthy delivery rate while deliverability is already deteriorating, because carrier filtering tends to arrive as a step change rather than a slope: traffic passes until a trust score is re-evaluated or a filter is triggered, and then a large share stops at once.
What moves it
Four things account for most deliverability problems in US business messaging. Unregistered or under-registered A2P traffic on a 10-digit number, which carriers filter as suspected spam. A low trust score that throttles throughput below the volume you are sending. Content that trips carrier filters, such as public URL shorteners or prohibited categories. And the absence of a fallback path, so that a rich message to a phone that cannot render it is lost rather than downgraded.
Reading the signals
| What you see | What it usually means | Where to look first |
|---|---|---|
| Delivered rate drops sharply on one carrier | A filtering or registration event on that carrier, not a network fault | 10DLC campaign status and the carrier's vetting for the sending number |
| Messages accepted by the API but never delivered | Silent carrier filtering, the most common failure on unregistered local numbers | 10DLC registration and the brand's trust score |
| Throughput stalls at a fixed messages-per-minute | A throttle tied to the trust score or number type | The campaign's assigned tier; a higher-vetted brand gets a higher ceiling |
| Only links fail | A shortener or domain on a carrier block list | Replace public shorteners with a branded domain |
| RCS messages "disappear" on some phones | No fallback configured, so phones without RCS received nothing | Automatic fallback to SMS |
| Every message bounces for one number | The handset has opted out, is unreachable or the number was reassigned | The opt-out list and a number lookup before the next send |
The first row is the one to act on fastest. A sudden drop on a single carrier almost always means a registration or filtering event, and RCS messaging blocked walks through the usual causes in order.
Deliverability on RCS
RCS adds two mechanisms SMS lacks. A capability check before sending tells the platform whether the handset can receive RCS, and automatic fallback to SMS carries the message anyway when it cannot. Delivery and read receipts then report what happened per message, so a deliverability problem is diagnosable rather than inferred from silence. The sender itself is also verified: a verified sender carries a brand name, logo and checkmark that the carrier and Google have reviewed, which is the opposite of the anonymous 10-digit number that carrier filters are built to distrust.
Before a first send
- Register the 10DLC brand and campaign before any application-to-person traffic goes out on a local number; what is 10DLC explains what the registry wants to see.
- Keep the consent record for every number, in the form the carriers and the TCPA expect, so that a complaint can be answered. How does RCS consent work covers the RCS-specific part.
- Send from a branded domain, not a public shortener, and keep the content inside the use case the campaign was registered for.
- Leave fallback on. A rich message that cannot render should arrive as SMS or MMS, not vanish.
- Start at a volume the campaign's tier allows and raise it as the trust score settles, rather than sending the whole list on day one.
Measuring it
Track delivered rate by carrier and by number type, and watch for a sudden drop, which almost always means a registration or filtering event rather than a network fault. On SimplyRCS the sender verification and the 10DLC filing are done as part of onboarding, fallback to SMS is automatic, and the per-message delivery and read events appear in the analytics, so the measurement above is available without building it.