Save the Customer’s Attention: Frequency Caps, Quiet Hours, and Channel Budgets

A customer signs up on Monday. By Thursday they have received the welcome email, a browse-abandonment email, a push notification about the same product, the weekly newsletter, and an SMS about a sale. Five messages from five scenarios, each of which was individually reasonable, and an unsubscribe on Friday that the reports attribute to the newsletter.

Nobody sent five messages. Five automations did, and no rule sat above them. That rule is a contact policy. It sets how many messages one person can get in a period, across all channels and campaigns. It sets when nothing is sent at all. And it sets how much of the customer’s attention each channel may use. It’s one page long and lives in the platform, not in a calendar. It’s the difference between a program customers put up with and one they leave.

Why it is a budget

Every message spends a little of the customer’s attention, and attention doesn’t refill on your schedule. The sign that you’ve sent too much arrives late and looks like something else. Open rates fall and subject lines get the blame. Unsubscribes spike and the last send gets the blame. Spam complaints hurt sender reputation for every campaign after.

Validity’s 2025 benchmark puts global inbox placement around 84%, so roughly one in six emails never reaches the inbox. Over-sending shows up first in sender reputation, and sender reputation is one of the inputs mailbox providers use to place a message. The contact policy is a deliverability instrument before it is a courtesy.

Fig 1 · A budget for attention · Caps, quiet hours, and priority on one page

Part one: the cap, across everything

A frequency cap per campaign isn’t a cap. The customer doesn’t experience campaigns; they experience a day. The cap that works is per customer, per period, across all scenarios and all channels, enforced by the platform at send time.

Three numbers to set:

• Per day, all channels. Usually one or two. A day with a transactional message counts differently; see the exception below.

• Per week, all channels. Usually three to five for a retail brand, fewer for anything considered.

• Per channel per week. SMS and push get their own, lower cap, because they interrupt. One SMS a week is a lot.

Set them from the data, not from a template: pull the unsubscribe and complaint rates by the number of messages received in the previous seven days. There is a step where the rate jumps. The cap sits just below the step.

Part two: quiet hours

Quiet hours: a window, per customer’s local time, in which nothing goes out on any channel except transactional messages. Typically 21:00 to 08:00, wider for SMS. Several countries regulate marketing SMS hours; the policy should match the strictest market you send to instead of keeping a rule per country.

Quiet hours belong to the platform, not to each scenario. A wait node that fires at 23:40 holds the message until 08:00. The scenario designer shouldn’t have to remember.

Part three: channel budgets and priority

When two messages qualify on the same day and the cap allows one, which one goes? The policy answers with a priority order by message type. A message triggered by something the customer just did beats a scheduled one. A lifecycle step beats a newsletter. A reactivation beats a cross-sell. The platform picks the highest-priority message and defers or drops the rest.

Channel budgets apply the same idea to each channel. SMS costs a lot, in attention and in money, so save it for messages that earn it: a delivery update, a back-in-stock alert the customer asked for, or an abandoned cart above a set value. Push is cheaper but easier to switch off permanently. Email carries the rest.

The exception that proves the policy

Transactional messages, the order confirmation, the shipping notice, the password reset, sit outside the cap and the quiet hours. They are expected and they are the ones the customer would complain about not receiving. The policy has to define which messages are transactional, because that’s where marketing content sneaks in. A shipping notice with three recommended products is a marketing message in disguise, and regulators treat it that way.

Suppression is part of the policy

The cheapest message is the one you don’t send to someone who already said yes. Recent buyers out of acquisition offers; converters out of the cart flow; customers with an open complaint out of everything promotional. The paid-media version is the suppression audience; the owned-channel version is the same rule enforced inside the platform.

Writing it down

The policy fits on a page: the three cap numbers, the quiet-hour window, the priority order, the channel budgets, the transactional list, the suppression rules, and the date it was last reviewed. It is owned by the platform owner and reviewed quarterly against the unsubscribe and complaint curves.

Then it is configured in the platform: the cap on the project, the quiet hours on the project, the priorities on the message types. A policy that lives in a document and is applied by the campaign calendar is a suggestion. One that the platform enforces is a policy.

What to measure

The number isn’t open rate. It is unsubscribes and complaints per thousand delivered, by week, next to the messages-per-customer distribution. When the cap is right, the curve is flat. When it is wrong, the curve tells you which week and which channel.

The second number is revenue per message sent, which usually rises when the cap is introduced, because the removed messages earned the least. That is the number to bring to the person who thinks fewer messages means less revenue.

The takeaway

Caps across all channels, quiet hours the platform enforces, a priority order for the days when two messages qualify, budgets per channel, a transactional list nobody abuses, and suppression of the people who already said yes. One page, configured once, reviewed quarterly. It is the program’s retention policy in disguise, and it costs nothing but the messages that were hurting you.

Read next

• Open rate is a diagnostic, not a success metric. Why the falling open rate is the policy’s early warning.

• Journey orchestration across channels. The journeys the policy sits above.