Class Schedule A/B Testing: Running a Schedule Experiment Without Upsetting Regulars

Studios that change class times without a test phase lose regulars. A parallel run — new slot added alongside the existing one for four to six weeks — generates reliable data before any commitment is made. The experiment protects your regulars and answers the question with evidence, not a guess.

Why Is Schedule Testing Necessary Before Making Changes?

A schedule change based on instinct is a commitment made without evidence.

Owners change class times for reasonable-sounding reasons: "our morning class should start at 6am instead of 6:30 because clients keep requesting it." But client requests are a biased signal — the clients who ask for 6am are vocal; the silent majority who chose the studio specifically because of the 6:30 start time don't say anything until they leave.

The parallel run solves this. Instead of moving the existing 6:30 to 6am, add a 6am class temporarily. Track attendance at both for a month. The data tells you what the requests could not.

The 34% vs 8% difference is the cost of skipping the test. A studio with a 60-person attendance at a class that drops 34% loses 20 attendees per week. That is real revenue, and it often takes months to rebuild.

What Can You Test in a Class Schedule Experiment?

Three variables are worth testing; testing more than one at a time is usually a mistake.

Time slot. Moving a class to a different time of day or day of the week. Most common test. Requires a parallel run where both times operate simultaneously for the test period.

Instructor. Adding a new instructor to an existing slot or testing whether an instructor moves demand with them when assigned to a new time. Tests instructor-specific vs time-specific demand.

Class format or level. Adding a new format (e.g., adding a "High Intensity" version of an existing class type) to an existing time slot. This tests format demand rather than time demand.

How Do You Protect Regulars During a Schedule Experiment?

The test period should be invisible to your regulars.

Regulars who attend a 9am Tuesday class should experience no change during the test of a new 8am Tuesday slot. They still see their 9am class in the schedule. The 8am appears alongside it as an addition, not as a replacement.

The operational requirements:

• The test slot should be bookable for all clients, not just new ones.
• The test slot should be staffed with the same quality instructor as the existing slot.
• The test slot should be tracked in the same way as any other class — attendance, no-show rate, waitlist activity.

What Metrics Determine Whether a Test Slot Succeeds?

Four metrics, tracked weekly.

Average attendance. The primary metric. A new slot should reach at least 60–70% of the existing slot's average attendance by week three to signal genuine demand. A slot at 30% of the existing slot's attendance after four weeks is not finding its market.

Booking-to-attendance conversion. What percentage of clients who book actually attend? If a new slot has high bookings but low attendance, clients are curious but not committed. This is a signal about the fit between the slot and the actual schedule of the clients booking it.

Waitlist activity. Are clients hitting the waitlist? A test slot that fills and starts generating a waitlist within 3–4 weeks is a clear success signal. A test slot with zero waitlist pressure after four weeks is struggling.

Repeat attendance. Are the same clients coming back to the test slot week over week? Or is it rotating through different clients with no repeat visits? Repeat attendance is the leading indicator of whether the new slot will sustain itself.

How Do You Communicate the Decision After the Test?

The post-test communication is as important as the test itself.

If the new slot wins and you're making the change:

"After a successful four-week trial, our Thursday 7pm class is joining our permanent schedule starting February 1st. The existing Thursday 8:30pm class will move to a new schedule — here's how to find your new time."

Clear, factual, gives clients time to adjust.

If the new slot underperformed and you're not changing:

"We tried a Thursday 7pm class for the past four weeks and the demand wasn't there to sustain it. We're keeping our Thursday 8:30pm and will explore other time options in Q3. Thanks to everyone who tried the early slot."

Transparency about failed tests is actually a trust signal. It shows you test before committing.

If the test was inconclusive:

Extend by two weeks and communicate that the decision is pending. Better than a premature commitment in either direction.

How Do You Avoid Cannibalizing Existing Classes With Test Slots?

Cannibalization — where the test slot draws attendance from existing classes rather than attracting new or different clients — is the main risk in parallel testing.

Signs of cannibalization:

• Existing slot attendance drops by more than 15% during the test period.
• The same client names appear in both the test slot and the existing slot in alternating weeks.
• Total weekly attendance across both slots is equal to pre-test single slot attendance, not higher.

If cannibalization is occurring, it means the test slot is not expanding your reach — it is splitting the same demand across two times. This could indicate you're testing the wrong variable (time change, not format change) or that the two times are in direct competition for the same clients.

The cannibalizing scenario is not always bad news. If a 9am and a new 8am together produce 80% of the 9am's original attendance each, and the 9am's attendance dropped 40%, you have evidence that some clients prefer 8am. The question is whether the 8am demand is large enough to replace the 9am entirely — or whether running both permanently is viable.

For the full scheduling framework, see the scheduling software playbook. The peak hour class optimization guide covers demand analysis by time slot, and the class capacity optimization guide addresses the fill rate mechanics that inform testing decisions.