Cutting no-shows by rewriting one reminder

A metric everyone tracked and no one owned

No-show rate showed up in every leadership deck, but it belonged to no single team — scheduling blamed reminders, reminders blamed patients, and the product itself hadn't changed the appointment reminder flow in over two years. I was asked to take a fresh look at a feature that seemed too simple to be the problem.

Confirmation wasn't the same as commitment

The existing reminder asked patients to confirm their appointment, and most did — confirmation rates were above 90%. But confirming a text message and actually restructuring your afternoon around a doctor's visit turned out to be two very different acts of commitment, and the product treated them as the same signal.

• The reminder arrived too early to matter. A single message 48 hours out gave patients no reason to act on it in the moment.
• Confirming felt costless. Tapping "yes" took less effort than actually canceling, so patients defaulted to confirming even when unsure.
• Rescheduling was buried. The path to changing an appointment was harder to find than the path to silently not showing up.

Talking to the patients who confirmed and still didn't come

• Pulled six months of no-show data and cross-referenced it against confirmation timestamps to find the pattern: most no-shows had confirmed, most had confirmed early.
• Recruited and interviewed 22 patients who'd no-showed in the past quarter, focused on what changed between confirming and the appointment day.
• Ran a lightweight diary study with 10 patients across a full reminder-to-appointment cycle to capture real-time friction, not recalled friction.
• Mapped the emotional register of each touchpoint — where the flow felt transactional versus where it felt like it was actually about their health.

Make the reminder do less confirming, more deciding

• Moved from one reminder to a short sequence — an early heads-up, a day-of nudge, and a two-hour window check-in, each asking a different question instead of repeating the same one.
• Replaced the binary confirm with a real choice: keep it, move it, or cancel it — surfaced with equal visual weight so rescheduling wasn't the hidden option.
• Tested copy tone directly against no-show rate, not just click-through, since the two didn't move together in early experiments.

What shipped

A three-message reminder sequence with a one-tap reschedule path built into every message, rolled out gradually across clinics with an A/B test held against the original single-reminder flow to isolate the actual effect of the redesign from seasonal variation.