Cadence: how often is too often, and who decides

Cadence questions are usually asked backwards. "How many emails a week should we send?" has no answer, because the right cadence for a weekly regular and the right cadence for someone who ordered once in March are different by a factor of ten. Cadence is a per-customer setting derived from their behaviour, capped globally so no one gets buried.

Anchor cadence to order rhythm

The workable rule: contact frequency should not exceed the customer’s own order frequency by more than about 2 to 1 for promotional messages. A weekly customer can comfortably take one to two promotional messages a week. A quarterly customer should hear from you roughly monthly at most, and even that is generous.

Below that ratio you are under-communicating with your best customers, which is the more common error and the more expensive one. Operators who fear over-sending typically under-send to exactly the people who would happily hear more.

The global cap matters more than the per-campaign rule

Fatigue is not caused by any single campaign. It is caused by six campaigns each individually reasonable arriving in the same week because nobody was counting across them. A birthday email, a win-back, a game-day blast, a loyalty milestone, an expiry reminder and a new-item announcement is a normal week in a busy programme and an unbearable week in a customer’s inbox.

Enforce a single cap across every channel and every campaign type, evaluated at send time, with an explicit priority order for what gets dropped when the cap is hit.

1. Transactional and operational messages: order confirmations, delivery updates. Never capped, never counted.
2. Expiry reminders for value the customer already holds. Highest promotional priority; they are the customer’s own money.
3. Triggered lifecycle: welcome, second-order, win-back, birthday. Time-sensitive, so they take precedence over calendar sends.
4. Segment campaigns: targeted offers to a defined group.
5. Broadcast: the whole-list send. Lowest priority, and the first thing that should be dropped.

Fatigue shows up before unsubscribes

By the time unsubscribe rate moves, you have been over-sending for a month. The leading indicators are quieter.

• Open rate decay within the same segment: compare a cohort against itself over time, not against your list average.
• Click-to-open collapse: people are still opening out of habit but no longer engaging. This is the clearest early fatigue signal there is.
• Rising time-to-open: messages that used to be opened within an hour now take a day. The customer has demoted you.
• Spam-folder placement drift: deliverability degrades ahead of complaints, particularly at Gmail.
• SMS opt-out rate above 1% per send: treat as an emergency, not a metric. SMS opt-outs are permanent and consent is expensive to reacquire.

Let the customer set it

A preference centre with real options (weekly, occasional, offers only, deals plus news) reduces unsubscribes more effectively than any cadence rule you can impose, because it converts a binary decision into a dial.

Honour it exactly. A customer who selects "occasional" and then receives a weekly send has learned that your preferences are decorative, and the next thing they click will be unsubscribe.

The unsubscribe is not the customer telling you they are done. It is the customer telling you they were done a while ago.

Questions