How often should you send marketing texts? Let the first 30 days tell you

A text lands harder than an email ever does

Day one always looks wonderful. The welcome text gets read within minutes, replies come in, somebody screenshots one for the client's Slack channel. Then a promo goes out at eight on a Sunday evening, queued in the same slot as the email, and by midweek the opt-outs have spiked.

That pattern repeats because a text isn't an email. An unwanted email gets ignored and quietly forgotten. An unwanted text interrupts a phone in a pocket at dinner. There's no ignore button on a text, only unsubscribe.

So the first month of a new SMS programme isn't a cautious ramp. It's a measurement window. You're finding the real ceiling for this list, on this brand, instead of borrowing a frequency off a benchmark, and the only way to find it is to watch every send on its own.

Log opt-outs against every single send, never the monthly average

A monthly opt-out average is close to useless for spotting the moment things went wrong. One bad send gets diluted by seven decent ones and vanishes into a number that still looks fine on the report. So you get to month two thinking the list is behaving. Then you repeat the exact send that did the damage.

So we log opt-outs against each individual send instead, timestamped, with that send's own context attached. Within a handful of sends a shape appears. Most lists hold flat through the early texts, then one send bends the curve, and the context tells you why.

The send that bent the curve is nearly always the same story. It landed outside a window the subscriber thought was reasonable: too soon after the last one, or at a bad hour. That spike marks your list's actual ceiling, and it's worth more than any industry chart.

The first pattern-breaking send is the test, not the welcome text

Everyone tests the welcome flow and almost nobody tests what happens after it. The welcome text is the safest message the programme will ever send. The subscriber has just opted in and expects to hear from you, so it tells you almost nothing about tolerance.

The real test is the first message that breaks the pattern the welcome flow taught. Somebody signed up expecting order updates and restock pings, and their fourth text is a generic discount blast. The deal they thought they'd made is broken. That mismatch triggers the opt-out far more often than frequency in the abstract.

So we look at that first category-changing send harder than send one, reading its opt-outs on their own against the baseline the early sends set. If it sits meaningfully higher, either the welcome flow has to set expectations honestly, or that segment should never get that message type at all.

Fix the timing before you touch the frequency

Before frequency becomes the conversation, timing usually needs fixing. A badly timed send at a sensible frequency does more damage than a well timed send at a slightly higher one. And the same timing mistakes turn up in almost every new programme we take on.

• Don't inherit the email schedule by default, because a slot that suits an inbox can feel like an interruption on a phone
• Keep promotional sends off weekend evenings, unless you've got real evidence that the segment welcomes them
• Separate transactional-feeling sends like restock alerts from promotional ones in the calendar, because subscribers tolerate very different frequencies for each
• Send against the subscriber's own time zone, not the account default, especially if a real share of your audience sits outside the UK

Fix the timing variable first, then read the curve. Once timing's controlled, the curve is measuring frequency tolerance and not a text that landed at a bad moment. Two different problems, identical on a report.

Build the ceiling from the list in front of you, not the playbook

The tempting shortcut is a frequency rule imported from another account, or off the platform's own benchmarking page. That number might suit a general apparel list and fail a supplements brand whose subscribers genuinely want frequent restock alerts. It fails the other way for considered homeware. A modest month of texts there feels like pestering.

So for the first month we deliberately send slightly under what we think the list can take, and watch where the curve stays flat. Frequency goes up one send at a time in week three or four. Never before, and only once there's a baseline to compare against. If the curve bends at send three, the list has answered early, at a far smaller cost in subscribers than running a borrowed cadence blind for a month.

The ceiling is a segmentation decision as much as a number, which is why the lifecycle side of the work matters more here than the SMS platform does. A subscriber who's bought several times, like the repeat buyers in the accounts we write up, takes more texts than somebody who came in off a pop-up and has never purchased. The first month finds both ceilings.

Sometimes the list is telling you it doesn't want texts

Now and then the first month tells you something less convenient than reduce the frequency or move the send time. The curve stays high at a genuinely conservative cadence, on well timed, well segmented sends. That's not a tuning problem. That's the list telling you SMS is the wrong channel for this audience.

Saying so out loud is part of running the programme honestly. It happens more with older audiences and considered purchases than with impulse DTC lists, and the right response is putting the budget back into channels the list does want. Quietly reducing volume forever and calling it optimisation is the dishonest version of the same call.