Android SMS Gateway Scenario 56: SaaS onboarding two-way inbox in practice

Summary

SaaS onboarding on an Android two-way inbox is a support lane that happens to use SMS. Replies cost operator airtime. Unread HELP texts cost activation. Service pricing is based on device count and total SMS sent through the gateway. You need a working Android phone with a SIM and SMS credit from your mobile operator. Operator message costs are yours—we do not sell carrier SMS balance.

If nobody owns the inbox during the trial timezone, you do not have conversational onboarding. You have a outbound cron that ghosts customers.

Contrast rented aggregator numbers on Twilio vs Android SMS gateway. Here the FROM is a SIM you host, and inbound lands on that same radio.

Key takeaways

• SaaS onboarding SMS is a conversation: HELP and STOP need a human SLA, not a black hole.
• Login OTP must live on a different radio than the onboarding thread.
• Priced by devices and SMS send volume. You use your own phone and operator SMS credit. Free is 300 SMS lifetime, 1 device, 300 contacts.
• Auto-reply is a safety valve, not a support org.
• Unread inbound at 02:00 is a churn event at 09:00.
• Developer Center owns webhook field names; this page owns who answers.

Related reading

two-way SMS inbox, SMS auto-reply and STOP keywords, OTP and 2FA SMS, FCC unwanted-text guidance.

Context

Product teams text a day-0 checklist, then hope “reply HELP” is cheaper than in-app chat. The failure is not the send. It is a weekend inbound pile while signup OTP shares the same Xiaomi in a drawer.

Free (300 SMS lifetime, 300 contacts) proves the thread with staff numbers. A launch week is devices plus airtime plus a rota.

Core scenario guidance

Split three MSISDNs in your head even if you only buy two phones: OTP, onboarding conversation, and later marketing. STOP on the conversation number must not break login codes.

Auto-reply for STOP and a one-line HELP ack. Everything else waits for a human in the trial’s business hours. Do not let a bot invent billing promises.

Idempotency on signup_id for YES confirmations. Widget retries will double-start trials if you are naive.

Who owns which inbound

Cost and ownership

Priced by devices and SMS send volume. You use your own phone and operator SMS credit. Every CSM reply is a send. Budget airtime for conversations, not only the outbound drip.

Name who watches the inbox on launch Saturday. A “shared Slack channel” is not an owner.

Operations

Daily: unpaired minutes, inbox age of oldest HELP, OTP canary. After OEM updates, send and reply on a staff thread end to end.

Abort: pause drips, keep OTP, drain HELP. Marketing must not resume the drip without ops.

Security and compliance

Do not paste full session tokens or passwords into SMS. Rotate keys when a contractor growth hacker leaves. Verify inbound webhooks before they can flip account state.

Promotional day-7 upsells need consent and STOP. Day-0 setup texts are still identifiable traffic — say who you are.

Decision guide

Ship when a named rota covers the trial timezone and OTP is isolated. Delay if inbound is “we’ll check Monday.” If you refuse handset ops, put OTP on CPaaS and keep Android only if someone will actually sit the inbox.

Checklist

• OTP device pinned away from inbox SIM.
• STOP automated and logged.
• HELP ack plus human queue.
• YES confirmations idempotent.
• Rota matches trial locales.
• Airtime includes replies.
• Staff canary of a full thread.
• Webhook signatures on inbound.
• On-call for launch weekend.
• Developer Center checked for receive payloads.

Next steps

Use two-way SMS, device and SMS volume pricing, and SMS API documentation.

Deep dive: production hardening

OEM sleep kills inbound polling as surely as outbound schedules. Keep the inbox phone charged and exempted.

Dual-SIM “OTP plus inbox” units remap after reboot. Recheck slot maps or login codes land on the chat number.

Do not auto-reply a marketing paragraph into a thread that just sent a 2FA on another product.

Pace day-3 and day-7 drips. A launch list is not an excuse to ignore OEM ceilings.

Deep dive: scaling and failure modes

More seats means more replies, not a slogan about unlimited SMS. Devices still meter. Free and Developer still cap platform SMS.

Multi-region trials need inbox hours per locale. One UTC volunteer is how APAC HELP dies.

A support surge after an outage will contend with OTP if you merged radios. Fail closed: hold drips.

Deep dive: integration discipline

Map inbound to signup_id via the original outbound. Orphan replies need a search, not a second drip.

Feature-flag new drip copy. Canary staff and design partners before a 2,000-trial week.

Priced by devices and SMS send volume. You use your own phone and operator SMS credit. Live fields stay in Developer Center.

First-week inbox drill

Monday: staff thread — outbound, HELP, human reply, STOP. If any hop fails, do not announce SMS onboarding.

Tuesday: confirm OTP canary on the other phone while a drip is queued. OTP latency must not move.

Wednesday: oldest-HELP alert at 30 minutes in the trial locale. If nobody acks, page, do not send another drip.

Thursday: rotate a contractor key used by the growth worksheet. Inbox webhooks keep the scoped key.

Friday: hold weekend drips unless the rota is staffed. Ghosting Saturday signups is worse than silence.

Abort sentence: pause drips, keep OTP, empty HELP. If marketing can override that, you do not have a two-way programme.