Set push frequency around user expectations, message value and observed negative signals. There is no universal number of notifications that fits every app, audience and lifecycle stage.
Separate message types
An expected transaction update is different from a promotional reminder. Classify messages by purpose and urgency before setting frequency rules. A single global cap may either suppress important information or allow too much marketing.
Document which events justify a message and when the message becomes stale. A reminder sent after the user has already completed the action is a failure of coordination, even if the copy is excellent.
Give the system suppression rules
Suppress messages when the target action is complete, the user has opted out or another journey already covers the same need. Consider local time and the user’s natural usage pattern. Make controls understandable rather than hiding them in a complex settings screen.
Firebase provides Cloud Messaging documentation for its delivery infrastructure. Delivery capability is only part of the problem; relevance and consent still need product decisions.
Test incremental value
Where practical, compare a message strategy with a group that does not receive the promotional intervention. Examine meaningful actions and retention as well as opens. Users who would have returned anyway can make a campaign look more effective than it is.
Monitor opt-outs, complaints and uninstalls where observable. A short-term engagement lift can conceal longer-term damage to trust.
Review the whole message portfolio
Individual campaigns can look reasonable while the combined experience becomes exhausting. Maintain a shared calendar or orchestration view across product and marketing.
Frequency is an operating decision that should evolve with evidence. Start conservatively, define what success and harm look like, and increase only when the added message earns its place in the user’s day.