Email and SMS for wellness studios: message the visit state, not the calendar

Most studio messaging is organised around the sender's calendar. A newsletter goes out monthly because the month ended, a promotion goes out because a holiday is coming, and a win-back campaign runs in January because January is when operators think about win-backs. Every one of those sends reaches the whole list, which means it reaches people who came in yesterday, people who have never come at all, and people who quietly stopped three months ago, with the same words.

A recovery studio has better information than that, and it is sitting in the booking system already. Every member has a visit history, and a visit history implies a state: how long since the last visit, whether a next one is booked, whether the interval between visits is widening, whether a membership payment failed. Those states change on their own schedule, and a message that fires when one of them changes arrives at a moment the member can recognise. That is the whole idea.

One caveat sits over the entire subject. Consent, identifying yourself clearly in the message, and honouring an opt-out are durable obligations in messaging, and the specifics differ by channel, by jurisdiction and over time. Nothing here states a legal requirement, and none of it is legal advice. Before you launch a program, confirm the current rules for the channels you are using with a lawyer who works on them.

Six states a member can be in, and what each one needs

Start by writing down the states rather than the campaigns. A state is a description of where somebody stands with your studio right now, derived from records you already hold. Campaigns come afterwards, one per state, and the discipline of that ordering prevents the most common failure, which is a message that exists because someone had an idea for a message.

The states below cover a recovery studio's floor without much left over. What matters is that they are mutually exclusive at any moment, because a member in two states gets two messages the same afternoon and reads both as spam.

• Never visited: signed up, bought a gift card, or gave you an address at an event, with no visit on record. Needs a first-visit path with a specific time to book, not a general welcome.
• First visit complete, no second booked: the most valuable state on the list, and the most time-sensitive.
• In rhythm: visiting at a stable interval with a next visit booked. Needs almost nothing, and the correct message volume is close to zero.
• Slipping: the interval since the last visit has stretched well past this member's own normal gap, and nothing is booked. Needs a low-pressure, specific reopening.
• Lapsed: past the point where the member has plausibly stopped. Needs a different message entirely, and needs to stop after a defined number of attempts.
• Payment at risk: a failed card or an expiring one on a membership. Needs an operational message, sent immediately, that is not dressed up as marketing.

Overdue is a per-member number, not a studio-wide one

The temptation is to pick one threshold for the whole list, and it produces bad messages in both directions. A member whose pattern is twice a week is meaningfully absent after a fortnight. A member whose pattern is a float every few weeks is not absent at all on the same day, and telling them you have missed them is an obvious sign that a machine is guessing. The threshold has to come from the member's own history.

The computation is simple enough to run in a spreadsheet if your platform will not do it. For each member with enough visits to have a pattern, take the gaps between consecutive visits and use the median gap as their normal rhythm. Then flag members whose current gap has passed some multiple of their own median, and use the same multiple for everyone. What you get is a list of people who are behaving unusually FOR THEM, which is the population where a message can still change something.

Members without enough history to have a pattern need a separate rule, and that rule is usually about the first visit rather than about intervals. Anyone with exactly one visit is not slipping in any statistical sense; they are undecided, which is a different problem addressed further down and with far more urgency.

• Compute each member's median gap between visits, and flag on a multiple of that member's own gap rather than a fixed number of days.
• Exclude members with too few visits to have a pattern and route them to first-visit logic instead.
• Recompute the gaps periodically, because a member who moves from weekly to fortnightly has changed rhythm rather than lapsed.
• Watch seasonality on the modality level: heat and cold demand does not behave identically month to month, and a threshold set in one season mislabels people in another.
• Suppress anyone with a future booking, no matter what their gap says.

The message that pays for the entire program

If you build only one automation, build the one that gets visit two on the calendar. A recovery studio's product is a felt effect that fades, so a single visit rarely produces a habit on its own, and the gap between the first and second visit is where a new member decides, mostly without noticing, whether this is a thing they do. Everything downstream is more expensive: a slipping member costs a message and a discount, a lapsed member costs a campaign, and a member who never came back a second time is usually gone for good.

The strongest version of this is not a message at all. It is a booking made at the desk while the member is still in the building, standing in front of the person who ran their session, with the message afterwards acting as a confirmation rather than a persuasion attempt. Automation is the backstop for the ones the desk missed, and it should fire within a day or two while the session is still a memory rather than a receipt.

Write it as the smallest possible next step. One suggested time, drawn from your actual availability, with one tap to accept and one to pick something else. Not a menu of services, not a promotion, not an explanation of the benefits of consistency. The member has already decided the session was fine; the only thing in the way is that nothing is on their calendar.

What email is for and what SMS is for

The two channels are not interchangeable and the split is about time sensitivity rather than length. SMS is for a message that needs an action inside a few hours and would be a failure if it were read tomorrow: a booking confirmation, a same-day reminder, a slot that opened, a card that just failed. Email is for anything that benefits from context and can wait: a new modality on the floor, a schedule change, a seasonal explanation of how members typically use contrast in winter.

The most common error is running SMS as a broadcast channel because the delivery rate looks better. It does look better. What misusing it costs you is not a lower open rate but an opt-out, which removes your only reliable channel for that person permanently. An email ignored costs almost nothing. A text ignored is a text somebody remembers receiving.

Both channels need a real reply path. Members answer a reminder by asking to move the appointment, and a number that cannot receive that reply, or an inbox nobody watches, converts a service message into a small insult. Decide who reads replies and inside what window before you turn anything on.

Metered cost is a budget, and the constraint is doing you a favour

Messaging is priced per send, per segment or per contact, and a studio list large enough to matter is large enough for that meter to be visible on the monthly statement. The instinct is to see this as a limit to be worked around with a cheaper provider. The more useful reading is that the meter forces the question you should be answering anyway: given a fixed budget, which members are worth reaching this month?

Do the comparison explicitly rather than in your head. A send costs a known amount. A retained member is worth their monthly rate multiplied by however long they stay, and the entry-level monthly membership across listings that publish one is counted below from our own directory rather than asserted here. Set the two side by side and the arithmetic is not close: a message that recovers one slipping member pays for a great many sends. What it does not justify is spending the same budget on people who were coming in anyway, which is exactly what a whole-list monthly newsletter does.

Segmented, state-triggered messaging usually reduces total volume while increasing total effect. Fewer people are messaged in any given week, each message is about something that actually happened, and the members in a stable rhythm mostly hear nothing, which is the correct treatment of somebody who is already doing the thing you want. The budget line goes down and the booking line does not.

Consent, identification and opt-out as standing obligations

Three things are durable across messaging channels regardless of how the specifics are written where you operate. You need the recipient's consent to message them for the purpose you are messaging them for. You need to identify clearly who is sending, so a member knows the text is from your studio and not from an unknown number. And you need a working opt-out that is easy to use and honoured promptly across every list you hold. Requirements around all three differ by channel and by jurisdiction and they change, so confirm the current position with counsel rather than inheriting another studio's setup.

The practical failure in studios is usually record-keeping rather than intent. A phone number collected so the desk could confirm a booking is not obviously the same thing as agreement to receive promotions, and once both live in one contact record nobody can tell them apart. Keep the provenance: when the contact was captured, on what form, with what wording shown, and which channels and purposes it covers. That record is cheap while you are collecting it and impossible to reconstruct afterwards.

Treat an opt-out as global by default. A member who stops your marketing texts should not keep receiving them from a second tool that holds its own list, and the fastest way to get that wrong is running messaging out of three systems that do not talk. Operational messages about a booking someone actually made are a different category from marketing, and keeping the two clearly separated in your own systems is what allows you to honour an opt-out from one without breaking the other.

Writing the messages, and keeping the volume defensible

Each message should carry one action. If a member has to decide between booking a session, reading an article and following you somewhere, the most likely outcome is that they do nothing, and the second most likely is that they unsubscribe. Name the state in the first line so the message explains itself: it has been a while, your card did not go through, your session is tomorrow.

Avoid manufactured urgency, and avoid outcome language entirely. A message that promises a physiological result is advertising with a substantiation problem attached, and the same rules that govern your website copy govern an automated text. Describe the session, the time, and what happens when they arrive. Personalisation should use records you actually hold, because a message that references the wrong service is worse than one that references nothing.

Keep the whole program small enough that one person can hold it in their head. The share of listed locations that are independent rather than part of a chain is counted below from our own directory, and at an independent location the person writing these messages is also running the floor, covering shifts and ordering towels. Six automations that get re-read every quarter survive that. A twenty-sequence lifecycle map built over one enthusiastic weekend does not, and its failure mode is not that it stops working: it keeps sending, long after anybody remembers what it says.

Praxium is not a messaging platform, and none of the sending described above happens there. It is a directory a studio can be listed in for free, plus an optional protocol layer that records the plan a member agreed to, which is one of the states these messages describe. The sending stays wherever you already do it.

• One action per message, named in the first line, with the link or the button doing the work.
• No health outcome claims in any automated message; the compliance standard for a text is the same as for your homepage.
• Cap the number of attempts per state, especially for lapsed members, and stop cleanly rather than tapering forever.
• Suppress anyone with a future booking, anyone who visited in the last few days, and anyone already in another sequence.
• Re-read every automation once a quarter; sequences written two years ago are still sending, and the studio has changed.

Measuring it without fooling yourself

Open rates have become close to useless as a measure of interest, and click rates only tell you that a link was interesting. The only outcome worth counting for a studio is a booked and attended visit inside a defined window after the send, attributed to the state that triggered it. That number is computable from the same booking records you used to define the state in the first place, which is the quiet advantage of building the program this way.

The strongest available check is a holdout. Withhold the message from a randomly chosen slice of each triggered segment and compare their return rate to the messaged group over the same window. Without it you will credit the message for members who were coming back regardless, which is most of them in a healthy segment, and that misattribution is how studios end up convinced a campaign works while the booking line stays flat.

Watch the cost side with equal attention. Opt-out rate per send, complaint rate, and the number of contacts you can no longer reach are the running charges of the program, and they do not recover. A sequence with a good booking lift and a rising opt-out rate is borrowing from a channel you will want later, and the right response is to send it to fewer people rather than to rewrite the subject line.