Reminder Cadence Across Email, SMS and WhatsApp at a Dental Group
A dental group with 41 clinics uses one Dotdigital account for appointment reminders and recall campaigns. Consent rates among active patients are email 78%, SMS 84% and WhatsApp 31%. Reminders go out 48 hours before the appointment, and every patient who has consented to all three channels receives all three, about 26% of the file. The no-show rate for that triple-contacted group is 8.1%; for SMS-only patients it is 9.0%. Cost per message is £0.032 for SMS, £0.018 for a WhatsApp template and £0.0004 for email. Each no-show costs roughly £85 in lost chair time. WhatsApp delivery failures run at 6%, with reasons mostly 'template paused' and 'user has not opted in to receive messages from your business'. The group now wants a two-way flow where patients reply 1 to confirm. Marketing has proposed removing email reminders entirely to simplify reporting and cut cost, and has asked whether the WhatsApp reminder should be built as a utility or a marketing template.
1. Design the reminder sequence across channels and justify the ordering using the numbers in the setup.
Read the worked answer
The numbers do not support triple-sending. Triple-contacted patients no-show at 8.1% and SMS-only patients at 9.0%. That 0.9pp gap is smaller than the selection effect, because patients who opt into three channels are more engaged to begin with. The true effect of the extra two messages is close to zero, while you pay £0.0504 per patient to deliver near-identical content three times.
Sequence by intent, not by channel count. The goal is a confirmed appointment, so make confirmation the exit condition.
- WhatsApp at 72 hours out, for the 31% who have opted in. It is the cheapest message (£0.018), carries the highest perceived urgency, and can host the one-tap confirm.
- SMS at 48 hours out, only to patients who have not confirmed. SMS has 84% consent and the best deliverability, so it is the reliable backbone rather than a duplicate.
- Email at 48 hours out, only for patients with no SMS or WhatsApp consent, roughly 16% of the file. It is effectively free and is their only reachable channel.
- No further reminder once a patient confirms. Confirmation is the product; a second reminder to a confirmed patient is pure cost.
Test it. Run the sequence against the current triple-send for four weeks on a random 20% split. Compare no-show rate and cost per attended appointment, not cost per message sent. If WhatsApp confirmation works as expected, the SMS fallback volume drops and the saving compounds.
2. Address the WhatsApp template category, the 6% failure rate, and the constraints on the confirm-by-reply flow.
Read the worked answer
Template category matters more than price. An appointment reminder is a utility template: it relates to an existing booking and can be sent on the strength of the appointment relationship plus the WhatsApp opt-in. A recall or 'time for your check-up' campaign is a marketing template and needs a separate marketing opt-in. Reusing the reminder opt-in for recall is the most common compliance failure on this channel, and it damages template quality as well.
Diagnose the 6% before scaling. 'Template paused' means the quality rating on that template fell, usually because recipients blocked or reported it. It is not a delivery glitch; it is feedback that content or frequency is wrong. Check the quality tier and block rate per template before sending more volume. 'User has not opted in' means the number exists in the WhatsApp Business Account but that patient never granted opt-in - a consent data problem in the contact records, not a channel problem. Reconcile the two lists before the next send.
Confirm-by-reply constraints. Building it is the right move, because it gives a clean confirmation signal to write back to the appointment record and reduces SMS fallback. Practical limits: replies only arrive inside the 24-hour customer service window, so the reminder must be sent early enough that a reply still lands inside it. You also need a route for replies that are not 1, 2 or 3, because unanswered or mishandled replies still count against quality.
Sequencing. Fix template quality and consent reconciliation first. Launching a two-way flow onto a paused template simply multiplies the failure rate.
3. Marketing proposes removing email reminders entirely. Evaluate that proposal and state what evidence would change your mind.
Read the worked answer
Removing email saves almost nothing and costs reach. At £0.0004 per message, email is under 1% of the per-patient messaging cost of the current triple-send. The saving is not the issue; the coverage is. Email reaches 78% of active patients, and roughly 16% have no SMS or WhatsApp consent at all, which makes email their only reachable channel. Removing it would drop a sixth of the file from all reminders and push no-show cost up - one prevented no-show at £85 more than covers the entire email bill for thousands of patients.
The real problem marketing is describing is reporting complexity, not cost. Three channels mean three engagement measures and no single view of whether a patient was reminded. Solve that with reporting, not by deleting a channel: build one reminder-event table keyed on appointment ID with channel, send time, delivery status and confirmation status. That gives a single answer to 'was this patient reminded and did they confirm' without touching the channel mix.
What would change my mind. If a four-week test showed that email reminders to patients who already received WhatsApp and SMS produced no measurable reduction in no-shows, and that patients with email-only consent were negligible in number or could be converted to SMS consent at low cost, then dropping email would be defensible. Neither condition holds in the current data.
Do not judge on opens. Email reminders are read on the day; open rate will look poor next to WhatsApp read rates and will not tell you whether the reminder did its job.