Telehealth No-Shows: What Your Ad Data Explains

Telehealth no-shows are partly a marketing signal, because channel, creative promise, and booking lead time all show up in whether a patient attends, and Curve is the HIPAA-compliant attribution layer that lets you break no-show rate down by campaign without sending appointment data to an ad platform. The reason this matters is structural: a booked appointment is a promise and an attended one is revenue, so a channel that books cheaply and attends poorly is more expensive than it looks. Curve does the measurement server-side with a signed BAA on every plan.

Why no-shows are a marketing metric

Most telehealth companies treat no-show rate as an operations number owned by the care team, reported as a single practice-wide figure, and addressed with reminders.

That framing hides the useful information. No-show rate is not uniform across sources. It varies with what the patient was told before they booked, how far ahead they booked, what they expected to pay, and how much friction stood between the click and the calendar. Every one of those is a marketing decision.

The financial consequence is direct. If channel A books at $40 and attends 80 percent of the time, and channel B books at $30 and attends half as often, channel B is the more expensive channel despite looking cheaper on every dashboard the ad platform shows you. Teams reallocate toward B routinely, because cost per booking is the number sitting in front of them.

And it compounds. Feeding booking events to a bidding algorithm teaches the platform to find people who book. Some of those people are enthusiastic bookers and indifferent attenders. Optimize on the wrong event for long enough and the algorithm gets very good at finding them.

What actually drives no-shows on the marketing side

Booking lead time

The strongest and most consistent driver. An appointment booked for tomorrow attends at a very different rate than one booked three weeks out, and the difference is large enough to dominate most other factors.

This is a marketing-adjacent problem because ads generate demand at a rate your schedule may not absorb. When spend rises and available slots do not, average lead time stretches and attendance degrades quietly. The metric to watch is not the no-show rate alone but the no-show rate against lead time.

Expectation mismatch in creative

What the ad promised and what the intake required have to line up. An ad implying a quick assessment followed by an intake asking for twenty minutes of history produces people who book to see and then reconsider.

Price is the sharpest version. Creative that leads with a low entry price attracts bookings from people who discover the real cost during intake, and a meaningful share of those simply do not show rather than cancelling.

Friction and its inverse

Extremely low friction booking increases volume and decreases commitment. A one-tap booking flow with no payment step and no meaningful intake will book more people and attend fewer of them than a flow requiring a card on file.

Neither design is universally right. What is wrong is optimizing the flow for booking rate without watching the attendance rate move in the other direction.

Channel intent

Search traffic for a specific service generally attends better than broad social traffic, because the person arrived with a formed intention. That is not a reason to abandon social; it is a reason to expect different attendance rates and price the channels accordingly rather than comparing them on cost per booking.

Measuring it without PHI

Here is the tension. To break no-show rate down by campaign you have to connect an appointment outcome to an ad click, and appointment outcomes are health information.

The resolution is that the analysis happens inside your systems and only a neutral signal travels outward.

Inside your infrastructure, where PHI is permitted and a BAA is in place, you can join attribution to appointment records freely. Campaign, creative, lead time, attended or not. That is where your no-show reporting lives, and it can be as detailed as you like.

Outward to ad platforms, only the attended appointment goes, as a neutral conversion with a hashed identifier and the original click ID. The platform learns which clicks produced attendance. It never receives an appointment time, an appointment type, a no-show flag, or anything naming the service.

One thing to be careful about: do not send no-shows as a negative or distinct conversion type. A conversion event that exists specifically to mark a missed appointment tells the platform that a particular person had a healthcare appointment and failed to attend it. Send only the positive outcome and let the absence of an event carry the rest.

Keep the reporting dimensions coarse

For channel analysis you need lead time in buckets (same day, this week, next week, later), attended or not, and the acquisition source. You do not need the appointment time, the clinician, or the service line in the same view.

Coarse buckets are also what keeps the analysis honest. Fine-grained slicing of a modest appointment volume produces confident conclusions from very few observations.

How Curve measures attended versus booked

Curve is HIPAA-compliant ad tracking, marketing attribution, and analytics for healthcare, and telemedicine is a core vertical. The tracking script installs in place of the Meta Pixel and Google tag, so events land on Curve's US-hosted infrastructure rather than going directly to ad platforms, which is what makes the split above enforceable.

• Click ID capture and server-side persistence. The gclid, fbclid, and msclkid are captured at landing and stored, so an outcome that happens two weeks later can still be tied to the click that caused it.
• Bridge tokens. When booking happens on a separate tool such as IntakeQ, Calendly, or Jane App, a token carries attribution across the handoff. Without this, most telehealth funnels lose the chain at exactly the point where booking occurs.
• Incoming webhooks. Your scheduling system or EHR posts attended and cancelled outcomes back, matched by email, click ID, or bridge token. Incoming data cannot override protected core attribution and contact fields, so a scheduling integration cannot corrupt your attribution.
• Offline conversion uploads. Attended appointments upload in bulk with automatic click ID matching, which suits an outcome that arrives late and in batches.
• Neutral event aliases. The ad platform sees a neutral conversion name rather than the visit type or service line.
• Per-destination field mapping. Only explicitly mapped fields forward to a given destination and the default is that nothing goes, so appointment metadata arriving on a webhook stays behind.
• SHA-256 identifier hashing. Contact identifiers hashed per each platform's conversion API requirements before forwarding.
• PHI-pattern detection. Payloads containing PHI-shaped values are flagged as a monitoring signal, useful when a scheduling webhook starts carrying more than it used to.

A signed BAA is included on every plan. For related mechanics, see HIPAA-compliant lead routing from ad click to CRM and server-side Enhanced Conversions without PHI leakage.

Acting on what the breakdown shows

Once attendance is visible by source, three moves follow, in order of impact.

Change the optimization event. Switch campaigns from booking to attended appointment where volume allows the algorithm to learn. This single change realigns the platform's objective with your revenue and usually matters more than any bid adjustment. Where attended volume is too low to train on, keep bookings as the optimization event and use attendance as the value weight.

Reprice channels on cost per attended appointment. Rebuild your channel comparison on the attended number and expect the ranking to change. Channels that looked mediocre on cost per booking often move up.

Fix creative that overpromises. If one ad set attends far below its peers with the same lead time, look at what it promised about price, speed, or effort. Attendance is where an overstated promise is repaid.

There is also an operational move that belongs to the care team but is discovered through marketing data: if attendance degrades as spend rises, the constraint is capacity, and buying more demand into a stretched schedule converts marketing spend into no-shows. That finding is only visible when lead time and attendance are read together.

What not to conclude

Two cautions, because this analysis invites overreach.

No-show rate is not purely a marketing metric. Reminder cadence, time zone handling, technical failures joining the call, and the patient's own circumstances all contribute, and some of those dominate in a given practice. Marketing explains part of the variance, not all of it.

And attendance differences between channels are not automatically a quality judgment on the patients those channels bring. A channel serving a population with less schedule flexibility will attend differently. The right response is usually to adjust lead time and reminder design for that segment, not to abandon the channel.

Frequently asked questions

Can we send no-show events to Google or Meta?

No. An event marking a missed healthcare appointment for an identifiable person is a health disclosure, and neither platform signs a BAA for its advertising products. Send only the attended conversion.

Should we optimize campaigns toward attended appointments?

Yes, when attended volume is high enough for the bidding algorithm to learn from. Below that threshold, keep booking as the optimization event and use attendance as the value signal that guides budget.

How do we connect an appointment weeks later back to the click?

Capture the click ID at landing, persist it server-side, carry it across any booking handoff with a token, and match the outcome back through a webhook or an offline upload with click ID matching.

Is appointment time itself sensitive?

Attached to an identifiable person, yes. Appointment scheduling detail is health information. Keep it in your own reporting and never map it to an outbound destination.

Does a deposit or card on file reduce no-shows?

It generally increases commitment while reducing booking volume. Whether that trade is worth it depends on your economics, and the only way to know is to measure cost per attended appointment on both sides of the change rather than cost per booking.

What lead time buckets should we use?

Same day, within this week, next week, and beyond. Four buckets are enough to see the pattern and coarse enough that no bucket describes a small group of identifiable people.

Can we run remarketing to patients who did not show?

Not through an ad platform audience, because building that audience discloses that those people had healthcare appointments. Follow up through your own BAA-covered channels instead.

Where to start

Get attendance flowing back into your attribution first. Without it, everything above is theory. That means confirming click ID capture at landing, closing the booking handoff so attribution survives the click-out, and wiring the scheduling system to post outcomes back.

Then rebuild your channel comparison on cost per attended appointment and read no-show rate against booking lead time. Those two views usually change at least one budget decision in the first month.

Curve supplies the pipeline: server-side collection to US-hosted infrastructure, click ID capture and persistence, bridge tokens across booking handoffs, incoming webhooks and offline uploads for attended outcomes, neutral event aliases, per-destination field mapping, hashed identifiers, PHI-pattern monitoring, and a signed BAA on every plan.

Run the free compliance scanner against your booking funnel, read the HIPAA-compliant conversion tracking setup, or visit curvecompliance.com to connect attendance to spend.

Reviewed August 2026. Ad platform conversion APIs and healthcare advertising policies change frequently. Verify current requirements before implementation.