SMS Marketing Australia: Worth It, or a Compliance Risk?

Almost every SMS pitch opens with a true sentence: nearly every Australian still reads text messages. That part is not in dispute. What the pitch leaves out is that this is also the most heavily policed marketing channel in the country, the most expensive attention to buy, and the one channel where a list you cannot evidence consent for turns a marketing send into a regulator's file.

I have spent thirteen years choosing channels for Australian businesses, and the pattern with SMS is almost always the same. It gets adopted in a panic during a quiet month, sent to a list nobody can evidence consent for, measured by nothing, then dropped after two sends. The channel gets blamed. The list is usually the problem.

So this runs the other way. What the channel is actually for. The two questions that decide whether you may send at all. The arithmetic of one send, run on your own numbers rather than an industry average. The four message types that earn their place, and the three that belong in email. What to measure, and what to stop paying for.

One boundary, stated once. This is marketing judgement, not legal advice. Nothing below decides whether a particular list, message or sender ID is compliant for your business, and nothing below is a review of SMS platforms. If your list is large or your sector is regulated, get the consent position checked by someone qualified to check it.

The Channel With the Highest Bar to Entry

What 96% reach actually means for a small business

The reach number should come from the regulator rather than from a vendor, and this one does. The ACMA's Communications and media in Australia report, in its executive summary and key findings published in February 2026, draws on the regulator's own consumer survey of Australians aged 18 and over and states: "In 2025, mobile phones remained the dominant communication device, with almost all Australians using them for calls (98%) or SMS (96%) in the previous 6 months".

Read that as an owner and the meaning is narrower than the number suggests. Reach is not attention. A message arriving on a lock screen sits beside a bank alert, a delivery update and a family group chat, and the person reading it owes you nothing. The figure tells you the delivery mechanism works on almost every adult in the country. It tells you nothing about whether your message deserved the interruption.

Why the same reach makes it the most competitive channel to buy

Competition shows up in what advertisers bid. In Semrush's Australia database for October 2026, "sms marketing" carries a cost per click of $30.72 and "text message marketing" sits at $24.48. Those are third-party estimates of what advertisers pay, not a claim about what any reader should spend, and a bid is not a value. What they do show is that a lot of businesses with real budgets are buying this audience. You will not win this channel on price.

The three businesses SMS genuinely suits, and the ones it does not

Three shapes of business get real value from SMS, and they share one trait: the message attaches to something the customer already did or already expects.

The first is a business whose revenue depends on attendance. Salons, driving schools, mechanics with a booked bay, allied health rooms, anyone running a schedule. The customer holds an appointment, so the reminder is a service rather than a promotion.

The second is a business where time changes the answer. A slot that opened at short notice, a delivery window that moved, a service interruption. The message is worthless tomorrow and useful in ten minutes.

The third is a business with an enquiry nobody answered. A missed call from someone ready to book, a form filled in on Friday and never followed up, a quote that went quiet three days ago. That is the highest-value send most small businesses never make.

The businesses it does not suit are just as specific. No appointment, no deadline and no open enquiry means there is nothing to nudge. A product that needs explaining should not be sold in 160 characters with no room for the explanation. And a business that cannot name who consented, when, and to what does not have a list. It has a liability with phone numbers attached.

What the Rules Require Before You Send Anything

The two questions that decide whether you may send at all

The rules turn on consent you can demonstrate, which reduces to two questions you answer from your own records. Did this person agree to receive marketing from us? Can we show what they agreed to, when, and how they stop it?

If you can produce that record for a number, you are in a position to message it. If you cannot, you are guessing, and guessing is the whole exposure. That is an operational test rather than a legal one, and it takes an afternoon: open the list, sort the numbers by whether a consent record exists, and count what is left.

Where the line sits for a specific list is not something a marketing article can settle. The ACMA publishes the rules; read them at source rather than taking my summary or anyone else's.

Why a purchased or scraped list is not a shortcut, it is the whole risk

A bought list fails the first question before anything is sent. Nobody on it agreed to hear from you, and no amount of copywriting fixes that. The seller will tell you the numbers are opt-in. Ask for the consent record per number and you get a spreadsheet with no provenance.

There is a colder commercial argument too. A purchased list converts worse than your own, because the person receiving the message has no idea who you are. Every complaint and every unsubscribe is also a signal, and you paid for the list that generated it. Buying a list to avoid building one is paying to inherit somebody else's risk.

What a branded sender ID now requires, and what happens if you skip it

The operational change is dated and clear. The ACMA states: "From 1 July 2026, new rules mean texts sent with an organisation's name at the top of a message (known as a branded sender ID) will need to have the sender ID registered."

If your messages arrive under a business name rather than a mobile number, that registration is now part of the setup, and it belongs to the business rather than to whoever sends on its behalf. Check the ACMA's register page for your own situation. I would not rely on an article for a compliance decision, including this one.

What the recent penalties show about the regulator's threshold

Two enforcement actions make the scale concrete, and neither reads like a business that set out to break the law. The ACMA states: "Latitude Finance Australia has paid a $3.96 million penalty after the ACMA found the company breached Australia's spam laws more than 2.7 million times." The TAB matter is described by the regulator as follows: "it sent more than 217,000 marketing emails and SMS over a 16-day period to customers who had unsubscribed from specific marketing channels". That one was self-reported.

What I take from both is that the failure is usually a process failure rather than a decision to spam. An unsubscribe flag in one system, a send from another, a list restored from a backup last month. Nobody in that chain chose to break a rule, and the penalty does not care. If you message at any volume, unsubscribes have to be honoured in every system that can send, including the one everybody forgot about.

The Arithmetic of One Send

Start with your own list size, not an industry average

Every article on this subject hands you a response rate from somebody else's list. Ignore all of them, including any I could quote. A response rate is a property of a list, an offer and a moment. Yours is the only one that predicts your result.

Working out what one send is allowed to cost

Build the cost from the two things you actually pay. The first is time: the hours to write the message, check the list and handle the replies. The second is the platform or per-message fee, which moves with provider and volume. I am not putting a number on the second one, because I found no sourced Australian benchmark and an invented figure belongs in a sales deck rather than in an article.

Then set that cost against something specific on the other side of the ledger: what one recovered booking, one attended appointment or one reopened enquiry is worth to you. Not average customer value. The one event the message is designed to cause.

The two ways a send loses money: wrong list, wrong moment

Both failures cost the same and look different. Wrong list is sending to people who never agreed, or who agreed two years ago about something else entirely. Wrong moment is a useful message arriving when nothing in the reader's day connects to it: a reminder three days out, an offer on the day the invoice lands, a "we have missed you" to somebody who was in last week.

The second failure is the one blamed on the channel. A solid list and a badly timed message produce the same silence as a bought list, and the owner concludes that SMS does not work. Usually it means the message was not attached to anything the reader was already doing.

A worked example with the reader's own numbers in it

This is the arithmetic I would run with an owner in one sitting. I am leaving the values blank deliberately, because my figures describe somebody else's business and would tell you nothing about yours.

Five lines, written down before you spend anything:

• Your list size, counted by number rather than by record.
• The share you can evidence consent for, taken from your own records.
• Sends a month you could actually sustain, rather than the number the platform's plan allows.
• What one send costs you: the hours multiplied by what an hour is worth to your business, plus the fee.
• What one recovered booking or reopened enquiry is worth to you.

Then three sums. List size times the evidenced share is your real reach. Real reach times sends a month is what the channel delivers in a month. Cost per send times sends a month, divided by real reach, is what you pay to reach one consenting person.

The decision sits in that last number. If reaching one consenting person costs more than the bookings that reach produces, the channel is too expensive for your business at your list size, and no amount of rewriting changes it. If it costs a fraction of that, you have room to send more often, and the real constraint becomes how many useful messages you can honestly write in a month.

What the exercise usually exposes is not a message problem. It is a list problem. Most owners who run it find the evidenced share is far smaller than the list size, and the honest channel turns out to be a quarter of the size they assumed they had.

If you want the arithmetic modelled before you commit, the budget calculator takes your own list and spend rather than a generic benchmark.

If you would rather have the decision checked than made alone, my own figures are published rather than withheld: "SEO retainers in Australia typically start around $3,000 a month, diagnostic projects start around $5,000, and hourly consulting runs $100 to $300." That is what it costs to have someone check the numbers before you spend, and it is a different purchase from a per-message fee. For a shorter version of the same work, the three engagement models and what each one costs is the honest entry point.

The Four Messages That Earn Their Place

The appointment or booking reminder

The simplest message and, for most appointment-based businesses, the highest return. Send it close enough that the person can act on it and far enough out that they can move it. It should contain the time, the place, the name and one way to change the booking. It should not contain an offer, a newsletter or a "while you are here".

One test: would the message still be useful with your logo removed? A reminder passes. A promotion dressed as a reminder does not, and readers work that out on the second send.

The time-critical notice that changes what someone does today

Something has moved, broken, opened or closed, and the person needs to know before they leave the house. Weather-affected work, a cancelled class, a delivery window that slipped. This is where SMS beats every other channel, because email arrives when it arrives.

The discipline is that only real events qualify. Manufacture urgency to get a read and you are training people to ignore the one message that mattered.

The recovery message after a missed call or abandoned enquiry

The send most small businesses never make, and the first one I would fix. Somebody called and nobody picked up. Somebody filled in the form and the follow-up never happened. A short message that names the miss and offers a time is the cheapest new business available to you.

It also has the shortest shelf life. A recovery message an hour late is service. A recovery message a week late is an apology.

If missed enquiries are the pattern rather than the exception, the problem sits upstream of the message, and the enquiry path itself is worth rebuilding before you spend anything on reminders.

The short sequence to people who asked to hear from you

Two or three messages at most, to people who opted in for exactly that. A service in their area, a season they buy in, a date they need to know. Keep it short enough that you could justify each message on its own, because that is the test a customer or a regulator will apply.

The Three Messages That Should Be an Email

Anything with a link, a paragraph and a decision in it

If the reader has to read, click and consider, that is email. A quote, a proposal, an explanation of what changed and why, a comparison of two options. SMS can carry a link, but the format punishes anything that needs more than one thought.

Anything that needs an image to make sense

A layout, a map, a menu, a before and after, a floor plan. If the message does not work without the picture, it is not an SMS.

Anything you would be embarrassed to send twice

The best filter I know. If you would not be comfortable sending the same message to the same person next month, it is probably not a useful message, and it is definitely not an SMS.

Where Email Wins, Honestly

What the owned-list advantage actually is

Email is the one channel where you hold the list, the history and the right to reach the same person again without paying per message. SMS buys attention. Email keeps a record. If a platform changes its terms next quarter, the email list still works, and that is why I treat email as the channel a business owns rather than as one option among several.

Why email carries the explanation and SMS carries the nudge

Different jobs. Email holds the paragraph, the proof, the image and the decision. SMS holds one instruction and about 160 characters. A business that turns email into a nudge gets ignored, and a business that turns SMS into an explanation gets one-word replies and unsubscribes.

How the two channels share one list without doubling the unsubscribe

One permission, two channels, in the words people actually use. When somebody opts in they should be able to see which channels they are choosing and change it in one step. In practice that means the preference is stored once and honoured everywhere, including the system that sends the SMS. Most unsubscribe complaints I see come from a preference stored in three places, two of which nobody maintains.

What to Measure, and What to Ignore

The one number that matters per send

Decide it before you send. A reminder is measured by attendance. A notice is measured by how many people acted on it. A recovery message is measured by enquiries back from that group. Record the list size and the number in the same place, every time.

Why an open rate is not available to you and a click is

SMS gives you no open. There is no pixel, and a delivery report tells you the message arrived rather than that it was read. What you have is what happens next: a click, a reply, a booking, a call. Measure the response and treat the delivery report as a technical check rather than a performance number.

How to tell a channel problem from a message problem

Three sends to the same list with one variable changed will tell you. If every send produces nothing, look at the list first and the moment second. If one message type works and the rest do not, the channel is fine and the message is the problem. That is why I would run one message type for a month rather than four in a fortnight: four half-read results decide nothing.

And if the honest answer is that you cannot say what the last send returned, that measurement problem comes before the next send.

What I Would Do in the First Ninety Days

Month one: evidence consent for the list you already have

Sort the list into two piles: numbers with a consent record you can produce, and numbers without. Message the first pile only. For the second, send a re-permission email instead, which is cheap, lawful and tells you who still wants to hear from you. Expect the list to halve. The half that remains is worth more than the half that went.

Month two: one send, one message type, one number

Pick the message type that matches the business. Appointment reminder if you hold appointments. Recovery message if you have missed calls. Send it once to the evidenced half, and write the number down before you look at it.

Month three: decide to continue, change or stop

With one real result recorded, the decision becomes cheap. Continue if the cost per consenting person reached is a fraction of the value of one recovered booking. Change if the list is fine and the response is flat, because that is usually timing or wording. Stop if the evidenced list is too small for the arithmetic to work. Stopping is a legitimate answer, and it is cheaper than a subscription you keep for two more sends out of habit.

What Not to Spend Money On

Do not buy a list, ever

No exception, no test campaign, no "just to see". The first send defines you to people who never asked, and the complaints arrive before the revenue does.

Do not buy a platform for a list of two hundred people

At that size the subscription is doing more work on your profit than the channel is. Write the message, send it from something you already pay for, and put the money into building the list instead.

Do not run SMS and email to the same people in the same week

Two channels, two interruptions, one inbox, and the unsubscribe usually lands on the wrong one. The exception is a reminder for an event the email announced, which is the sequence people expect.

Where the channel decision sits inside a plan for the year rather than beside it, that is how I choose which channels a business should actually run.

Frequently asked questions

Is SMS marketing legal in Australia?

Yes, and it is tightly regulated. The starting point is consent: you need to be able to show that each person agreed to receive marketing from you, and that they can identify you as the sender and unsubscribe. Buying a list, or scraping numbers, cannot satisfy that. On top of consent, a branded sender ID now has to be registered, and the requirement took effect on 1 July 2026. I am not giving you legal advice here and I would not rely on a blog post for a compliance decision: read the ACMA's own pages on the Spam Act and the sender ID register, and if your list is large or your sector is regulated, get it checked properly.

How much does SMS marketing cost?

Two costs, and most businesses only count the first. There is the per-message or platform cost, which varies by provider and by volume, and there is the cost of the list: the time spent evidencing consent, keeping the unsubscribes honoured and cleaning out numbers that no longer exist. The second one is the cost that catches small businesses, because a list that has not been maintained will produce bounces, complaints and, in the worst case, a regulator enquiry. I have not quoted a per-message price here because no Australian benchmark was sourced for it, and a made-up figure would be worse than none.

How many people should I expect to respond to a send?

I do not publish response rates, because a rate from someone else's list tells you nothing about yours. What I would do instead is run one send to your real list, measure one thing (a click, a booking, a reply), and record the number against the list size. That number is your baseline, and it is the only one worth comparing future sends to. Two sends to the same list, with one variable changed, will tell you more than any benchmark article.

Should we drop email and move to SMS?

No. They do different jobs and they should share one list rather than compete for it. Email carries the explanation, the proof, the paragraph and the decision. SMS carries the nudge: the reminder, the change, the short notice that alters what someone does today. A business that replaces email with SMS loses the only channel where it owns the audience and the history. A business that uses SMS as a second notice on top of email usually finds it is the notice that actually gets read.

We have a list of a few thousand customers from the last five years. Can we just message them?

Only for the people who consented, and the fact that someone bought from you is not by itself the evidence you need. Start by separating your list into people who opted in with a record you can produce, and people you cannot evidence. Message only the first group. For the second, do not guess and do not "test" it: run a re-permission email instead, which is cheap, lawful and tells you who still wants to hear from you. That exercise alone often halves a list, and the half that remains is worth more than the half that went.

How do we know whether it worked?

One number per send, decided before you press go. If the send was a reminder, the number is attendance. If it was a notice, the number is how many people acted on it. If it was a recovery message, the number is enquiries back from that group. Record the list size and the number in the same place, every time, and after three sends you can make a decision on evidence rather than on how the send felt.

The Honest Summary

SMS is a good channel with a high entry cost and no room for a bought list. It suits businesses that hold an appointment, face a deadline or have left an enquiry unanswered, and it punishes everyone else. It is not a substitute for email and it should not be run as one.

The order of work is not glamorous. Evidence the list, run one message type, record one number, then decide. If the evidenced list turns out to be too small to justify the send, that is a finding rather than a failure, and it is much cheaper to learn in month one than after a platform contract and a penalty notice.

If you want a second read on the arithmetic and the message types before anything is sent, tell me what you sent last quarter and to whom, and I will tell you what I would stop.