What mobile messaging decides
Messaging decides three things: who may be contacted, through which channel, and how often. The first is a question of permission. A push notification needs the operating system's opt-in; an SMS for marketing needs consent of the kind regulators and carrier codes describe; an in-app message needs nothing more than an open session, which is why it is the most underused of the three. The second is a question of fit: some messages need to arrive whether or not the person opens the app, others only make sense in the middle of a session. The third is the one teams get wrong most often, because every extra message borrows against the permission.
What messaging does not decide is the wider retention strategy or the tooling. The programme a message sits inside, from welcome to win-back, belongs to the lifecycle work that the mobile and app marketing guide frames as a loop. The journeys, templates and two-way flows that send the messages belong to the automation section, whose SMS and WhatsApp automation guide covers the workflow side. This section stays with the channel decisions: what each channel is for, how to earn and keep the permission, and how to read whether a message helped.
Push, in-app, SMS
The three channels differ in who they reach, what permission they need and what they are good for. Messaging apps such as WhatsApp sit close to SMS in this table: they reach a phone number, need explicit opt-in and are governed by the platform's own business policies.
| Channel | Who it reaches | Permission | Best at | Weak at |
|---|---|---|---|---|
| Push notification | People with the app installed and notifications allowed | Operating-system opt-in, revocable in settings | Timely, low-cost prompts back into the app: activity, reminders, things that changed | Reaching anyone without the app; long or complex content |
| In-app message | People currently using the app | None beyond the session; respect for the task in hand | Guidance, feature discovery, offers at the moment of use | Bringing anyone back; it waits for them to return |
| SMS | Anyone with a phone number who has consented | Marketing consent under direct-marketing rules and carrier codes | Urgent and transactional messages, reaching people who never installed or have lapsed | Frequent marketing; rich content; cost at volume |
| Messaging apps | Users of the app on that number who have opted in | Explicit opt-in and the platform's business policy | Conversational service, confirmations, two-way replies | Broadcast promotion, which the platforms restrict |
The See-Think-Do-Care framework is a useful way to place them. Messaging lives almost entirely in the Do and Care stages: it speaks to people who already have a relationship with the business. Using it as a See-stage reach channel, by buying lists or pushing promotions to people who never asked, is both ineffective and, for SMS, the kind of contact direct-marketing rules exist to stop.
The core guides
- Push notifications and in-app messaging covers the two channels an installed app unlocks: how to earn the notification permission rather than ask for it on first launch, the message types and the triggers that suit each, timing, frequency caps and quiet hours, deep links that land people on the right screen, and how to test and measure messages with holdouts.
- SMS vs push notifications sets the two interrupting channels against each other on reach, consent, cost, format, links and measurement, explains where each one wins, and shows how a single lifecycle uses both without sending the same message twice.
Read the push guide first if the app is live and the question is engagement. Read the comparison first if the question is whether to add SMS, or whether SMS is doing a job push could do more cheaply. Both core guides stop where the automation tooling starts, and both treat the permission as the asset being protected.
Further guides
Further guides take the first session, the lapsed user and the craft of each message in turn:
- App onboarding and permission requests — designing an app's first session: the path to first value, which permissions to ask for and when, soft prompts, onboarding messages, and reading changes by cohort.
- Re-engaging lapsed app users — defining a lapse for the app, who each channel can still reach, the reason to return, paid re-engagement, and holdouts.
- Designing in-app messages — banners, modals, tooltips, cards or an inbox, triggers from behaviour, accessible design, caps, and measuring with holdouts.
- Setting up notification preferences by topic — a tutorial for the settings people control: listing message types, grouping topics, defaults, the settings screen, honouring choices and opt-outs.
- Writing push notifications — a tutorial for a single push: naming the trigger, leading with what changed, making it specific, matching the deep link, previewing on devices, and testing.
- Building onboarding messages for new app users — a tutorial for the messages after the first session: the steps where new users stall, in-app first, push only for people who allowed it and left, one channel per nudge with the email flow, exits and a holdout.
Consent and frequency as the constraint
Every mobile channel depends on a permission that is cheap to withdraw and expensive to win back. A person who turns off notifications rarely turns them on again; a person who replies STOP to an SMS has ended that channel for good. That asymmetry is why this section treats FrequencyAverage times a person saw an ad. as the constraint rather than the lever. Sending more messages produces more taps in the short run and fewer reachable people in the long run, and the loss does not appear in the report for the message that caused it.
Consent itself is a legal question as well as a practical one. The UK Information Commissioner's Office publishes direct-marketing guidance that sets out when electronic marketing needs prior consent and what a valid opt-out looks like, and the CTIA's messaging principles describe what carriers expect of business messaging in the United States. Both are general direction, not legal advice, and both point the same way: ask clearly, record what was agreed, honour opt-outs at once, and keep marketing separate from service messages. Treat the rules for your markets as a floor and the audience's tolerance as the real ceiling.
The practical controls follow from that. Cap messages per person per period across all channels, not per campaign. Respect quiet hours in the recipient's time zone. Prefer triggered messages, which respond to something the person did, over broadcasts, which respond to the calendar. Watch opt-out and uninstall rates alongside click-through (CTRShare of impressions that result in a click.) and conversion (CVRShare of visits or clicks that convert.), because a message that wins the click and loses the permission has cost more than it earned.
Lanternhatch Street Pharmacy moves reminders off SMS
Lanternhatch Street Pharmacy, an invented chain with an ordering app, sends every prescription-ready alert by SMS to 42,000 customers at £0.04 a message, about 90,000 messages a month. Its app has 18,000 users with push enabled. It switches those 18,000 to push for the ready alert and keeps SMS for customers without the app and for collection reminders after three days. Monthly SMS volume falls to about 51,000 and collection within two days holds steady at 74%. It also stops a weekly SMS promotion that drew 1.1% clicks and 0.6% opt-outs per send, judging that the opt-outs were costing the channel it needs for urgent alerts.
Common mobile messaging problems
Most messaging problems show up as one of a few symptoms. The table maps each to the guide that diagnoses it.
| Symptom | Usual cause | Guide |
|---|---|---|
| Few users allow notifications | Permission asked on first launch with no reason given | Push notifications and in-app messaging, App onboarding and permission requests; diagnosis: push notification opt-in rate low |
| Opt-outs and uninstalls rising after campaigns | No cross-channel frequency cap; broadcasts sent to everyone | Push notifications and in-app messaging |
| Taps that land on the home screen and stop | Messages without deep links, or links that break for logged-out users | Push notifications and in-app messaging, Writing push notifications |
| SMS costs growing with nothing to show | SMS used for engagement that push could do for app users | SMS vs push notifications |
| Urgent messages missed | Critical alerts sent only by push to people who disabled it | SMS vs push notifications |
| The same message arriving twice on two channels | Channels run by separate teams with no shared rules | SMS vs push notifications |
| Nobody knows whether messages add anything | Opens and taps reported without a holdout group | Push notifications and in-app messaging |
| Journeys hard to build or change | Workflow tooling, not channel strategy | SMS and WhatsApp automation |
| New users leave before the app has shown its value | An account and a run of permission prompts before anything useful; tours instead of tips | App onboarding and permission requests |
| People turn off every notification to stop the promotions | One switch for everything; promotional and service messages not kept apart | Notification preferences by topic |
| Lapsed users do not come back | Push sent to people who uninstalled or turned it off; a reminder instead of a reason to return | Re-engaging lapsed app users, Setting an app's lapse threshold |
| In-app messages dismissed unread | A modal on every open; triggers on app open rather than from behaviour; no exit condition | Designing in-app messages |
| New users who stall get the same reminder on every channel, or reminders for steps they already finished | Onboarding sent on fixed days by each team, with no step map, exits or shared plan | Building onboarding messages for new app users |
For a symptom-first diagnosis, the app marketing troubleshooting section works through each cause with a check, a fix, a verification step and a prevention habit.
