Class Demand Forecasting for Gyms: A Practical Guide

Class demand forecasting means predicting how many members will show up for each class, then building your schedule, room sizes and instructor staffing around that number. You do not need a data science team to do it. A spreadsheet, eight to twelve weeks of booking history and a weekly habit of comparing predicted attendance with actual attendance will get most studios most of the way there.

This guide covers what to forecast, which data to collect, three methods from simple to advanced, how to turn a forecast into schedule decisions, and how to tell whether your forecasts are any good.

Getfitnessflow

Make Class Scheduling More Informed

Fitness Flow brings class scheduling, bookings, check-ins, and analytics together to help gyms manage demand with less manual work.

Explore Fitness Flow

TL;DR

Forecast attendance for each class type and time slot, not for the gym as a whole. That is the level where scheduling and staffing decisions get made.
Start with your own booking, check-in, waitlist and cancellation history. Eight to twelve weeks is enough for a first baseline.
Sort classes into full, steady and struggling, then move, split, shrink or retire them. Add seasonality and no-show patterns once the basics work.
Measure accuracy every week by comparing predicted with actual attendance, by class. A forecast you never score is a guess.
Software that holds bookings, check-ins, waitlists and billing in one place makes all of this far less manual.

What class demand forecasting actually is
Which demand are you forecasting?
The data worth collecting
Three forecasting methods, from simple to advanced
Turning the forecast into schedule decisions
Handling no-shows, cancellations and booking lead time
How to measure whether your forecast works
Our perspective: forecast to protect the member experience
Where Fitness Flow fits
A four-week plan to get started
FAQ
Sources

What class demand forecasting actually is

The search results for this term are dominated by supply chain and retail guides about units, SKUs and inventory. The idea carries over to fitness, but the product is different. You are not forecasting how many items you will sell. You are forecasting how many people will occupy a fixed number of spots in a room at a particular time, with an instructor you have already paid for.

A practical definition: use attendance patterns and participation signals to predict demand at the level where decisions happen, meaning a specific class, time, instructor and location. “Mondays are busy” is an observation. “The 6 a.m. Monday spin class will draw 14 people in a 20-bike room” is a forecast you can act on.

This matters because most of a class’s cost does not move with attendance. Athletech News reported in August 2026 that, according to Playlist data, studios fill about 37% of class capacity on average, while instructor, facility and scheduling costs are largely fixed whether eight or eighteen people attend. If your average class is under half full, small improvements in matching supply to demand go straight to your margin.

It also matters for retention. Members who can reliably get the class type and instructor they want, without fighting through a waitlist, have less reason to leave. Matching the schedule closely to demand is as much a loyalty strategy as an efficiency one.

Which demand are you forecasting?

Before you open a spreadsheet, decide what the forecast number means. Four different numbers get called “demand”, and they are not interchangeable.

Registrations: how many people book a spot. This is the earliest signal, but it overstates attendance.
Check-ins: how many people actually walk in. This is what your staffing depends on.
Unmet demand: people on the waitlist, or who tried to book and found the class full. This is the demand your current schedule fails to serve.
Occupied capacity: attendance as a share of spots, which lets you compare a 12-person studio room with a 40-person group fitness floor.

Pick check-ins as your forecast unit and track the others as supporting signals. Staffing follows the people in the room, not the people who clicked “book”. Then forecast by class type and time slot, since that is how members choose. A 5:30 p.m. HIIT class and a 10 a.m. restorative class share a building and little else.

The answer to each class also tells you what to do with it. A class that consistently runs over capacity may need a larger room, a second session or a different instructor. One that runs under capacity may need a new time, a new format or retirement.

The data worth collecting

You likely already have most of what you need inside your booking and check-in system. The useful inputs fall into three groups.

Your own history

Bookings and check-ins by class type, day of week and time slot
Cancellations, and how close to start time they happen
No-shows and late cancels
Waitlist length and how often waitlisted members get in
Fill rate by instructor and by class

Member signals

Visit frequency by member, to separate regulars from occasional visitors
Feedback from surveys and front-desk conversations about which times people wish you offered
Membership growth or churn that changes your pool of potential attendees

Outside factors

Weather, holidays and school calendars
Local events that change commuting and routines
Seasonal swings across the year

Resist the urge to collect everything. Standardize what you have first: the same class names, the same time formats, and one consistent rule for what counts as an attended class. Messy data is the most common reason a forecast fails. If “Power Yoga” and “Yoga Power” are two rows in your report, your history is already split in half.

If attendance tracking is your weak spot, our guides on gym attendance tracking and class attendance tracking walk through setting it up properly.

Three forecasting methods, from simple to advanced

Trend projection, the method that shows up in the search results for this topic, is only one option. For a gym, the practical ladder looks like this.

Method 1: Sort by fill rate

Pull eight to twelve weeks of bookings and attendance. For each class (day, time, type), calculate average attendance divided by capacity. Then sort:

CategoryTypical fill rateWhat to consider

Consistently full

80% and above

Add a second session, move to a larger room, or open more spots

Steady

40% to 80%

Leave alone, or test small changes

Struggling

Under 40%

Change the time, change the instructor, rework the format, or retire it

These cutoffs are a starting rule of thumb, not a law. A boutique reformer studio with eight machines will treat 75% very differently from a 60-person group fitness room. Adjust the lines to your own economics.

This method needs no software beyond a spreadsheet, and it often reveals the biggest wins immediately: a packed class that could carry a second session, and three half-empty classes that could be one.

Method 2: Trend and seasonality

Once you have the baseline, project forward. Take the average for each class and adjust for what you know is coming: a growing or shrinking trend over recent weeks, plus seasonal swings.

Seasonality is real and strong. A longitudinal study of a Leeds, England, physical activity program found that month was the most important determinant of daily attendance, with summer months producing the most visits. That is one English program, so do not assume your gym follows the same curve. Your own multi-year history will tell you when your January surge peaks, when summer travel hits evening classes, and when the post-holiday lull arrives.

Industry-wide data gives helpful context. The Health & Fitness Association’s FIT Tracker, built from an anonymized foot-traffic sample covering more than 10,000 locations, reported that July 2026 visits per location were 0.1% above July 2025. Flat national traffic can hide big local differences, which is why your own numbers beat any benchmark. If you want to see how seasonal and recurring rhythms differ between formats, this piece on boutique studio versus gym scheduling shows how members of each choose by coaching, schedule and cost.

A practical approach: compare each week to the same week last year, and note the ratio. If the third week of June ran 15% below your yearly average last year, expect something similar this year, then adjust for membership growth.

Method 3: Machine learning

At larger scale, or across many locations, machine learning models can combine more factors at once. These models are useful for separating bookings from actual attendance by weighing instructor popularity, weather, holidays and each member’s cancellation history. One 2026 industry piece describes forecasting that cuts the hidden drag caused by no-shows on instructor utilization and energy costs.

Be realistic about the payoff. A model is only as good as the data it learns from, and for a single studio with a few hundred members, a well-kept spreadsheet and Method 2 will often match it. Where machine learning earns its keep is multi-site operations with thousands of class sessions per month, where no human can track every pattern. Whatever you use, insist on forecasts you can explain. If you cannot say why the model expects 9 people on Thursday at noon, you will not trust it enough to change the schedule.

Turning the forecast into schedule decisions

A forecast has value only when it changes something. Here are the moves it should drive.

Move classes to where the demand is. If your 7 p.m. Tuesday class is packed and your 1 p.m. Tuesday class draws three people, the forecast tells you which slot deserves your best instructor. Our short guide on fixing class scheduling in a week covers the first three moves we suggest.

Set capacity deliberately. If a class is forecast to hit its limit most weeks, raise the cap where the room allows, or split it. If forecast demand is far below the limit, shrink the room or the cap so the class feels full, because a packed room of 12 feels better than a sparse room of 12 in a space for 30. Our walkthrough on class capacity planning shows how to set limits in a week.

Staff to the forecast. The schedule is your staffing plan. Peak slots get your strongest instructors and front-desk coverage, and low-demand windows get lean coverage or a different use of the room. Our piece on demand-driven staff scheduling explains how gym managers cut weekly scheduling time this way.

Manage the waitlist as a demand signal. A long waitlist is unmet demand you can measure. If the same class has a waitlist every week, that is the forecast telling you to add capacity. Meanwhile, a good waitlist strategy fills cancelled spots automatically so late cancels do not become empty spots.

Reposition before you retire. Before cutting a struggling class, try a different time and a different instructor, one change at a time, for four to six weeks. Otherwise you cannot tell which change worked.

A worked example makes this concrete. Say your Thursday 6 a.m. boot camp averaged 6 attendees over ten weeks in a 20-person room (30% fill). Your Thursday 5:30 p.m. boot camp averaged 19 and has a waitlist of five most weeks. Rather than cut the morning class, you move your best-rated instructor to the evening, add a second evening session at 6:30 p.m., and shift the morning class to a 45-minute express format with a smaller cap. After six weeks you compare actual to forecast and decide again. No prediction was perfect, but each decision was better informed.

Turning the forecast into schedule decisions — overview diagram

Handling no-shows, cancellations and booking lead time

Bookings are not attendance, and the gap between them is where forecasts go wrong.

Treat no-shows as a separate forecast. Track, by class type, what share of bookings turn into check-ins. A reformer class might lose 5% of bookings while a free community class loses 30%. Apply those rates to current bookings and you get a much better prediction of the room on the day.

Booking lead time matters too. Reservations made far in advance tend to be missed more often than last-minute ones. One 2026 gym industry blog reports that long lead time is the strongest predictor of no-shows. Treat that as a hypothesis to test in your own data: compare no-show rates for bookings made a week out versus the same day. If the pattern holds for your members, you can shorten the booking window, send a reminder 12 to 24 hours before class, or ask for confirmation on early bookings.

Not every member behaves alike either. Research on gym participation has identified two broad patterns: regular attenders and intermittent exercisers, with life stress and barriers strongly predicting attendance. In practice, that means a class with many occasional visitors will be more volatile than one built around a core of regulars, and should be forecast with a wider margin.

Handling no-shows, cancellations and booking lead time — overview diagram

How to measure whether your forecast works

A forecast you never score is a guess. Build a weekly routine:

Record the forecast before the week starts. Write down predicted check-ins for every class.
Compare to actual check-ins at week’s end. Calculate the difference for each class, not just the total, since errors can cancel out across classes and hide problems.
Track error by class type. If you consistently over-predict evening yoga and under-predict morning cycling, you have found something to fix.
Set a baseline to beat. The simplest baseline is “same as last week” or “average of the last four weeks”. Any fancier method must beat it, or it is not worth the effort.
Watch the supporting numbers. Utilization, waitlists, unmet demand and no-show rates tell you whether the forecast is leading to better outcomes, not just closer numbers.

A useful mindset: treat attendance tracking like website analytics. Pinpoint where people drop off between booking and walking in, and you can fix scheduling problems with confidence instead of hunches.

Space utilization has long been recognized as important to fitness studio revenue growth, as a benchmarking report for US fitness studios noted. That report covers 2014 data, so use it for the principle, not for current numbers.

Our perspective: forecast to protect the member experience

We talk to gym owners every week, and the same pattern comes up. They know their busiest classes by feel, but the decisions that cost them money are in the middle: the class at 45% that nobody dares to touch, or the instructor assigned by habit rather than by demand. A forecast does not replace your instincts about your community. It tells you where to point them.

Our view is that the best forecast is the one you actually run every week. That is why we built Fitness Flow around keeping bookings, check-ins, waitlists, billing and the member app in one system. When your data lives in five tools, forecasting becomes a monthly export project that nobody does. When it lives in one, the weekly comparison takes minutes, and the conclusions flow straight into scheduling and retention workflows.

We also do not think forecasting is only about efficiency. A class that is consistently full with a waitlist frustrates members, and a class that is consistently empty drains energy from instructors and the people who show up. Getting the schedule right protects the feel of your community, and that is what keeps members coming back. Start simple, score your forecasts, and let the method grow only when the data justifies it.

— Louis

Where Fitness Flow fits

If you are running forecasting out of exported spreadsheets, it is worth seeing what changes when scheduling, bookings, attendance, billing and member engagement sit in one platform. Fitness Flow offers a Solo plan at $149 per month and a Studio plan at $349 per month, with Multi-site available for multi-location operators. You can see the plans on our pricing page, or look at how the platform works for gyms before you decide.

A four-week plan to get started

Week 1: Clean the data. Standardize class names and time slots. Export eight to twelve weeks of bookings, check-ins, cancellations and waitlists.

Week 2: Build the baseline. Calculate average fill rate for each class and sort into full, steady and struggling. List the three biggest mismatches between demand and capacity.

Week 3: Make the first changes. Pick no more than three. For example: split one overfull class, change the time of one struggling class, and reassign one instructor to a peak slot.

Week 4: Score and repeat. Compare predicted with actual attendance, record the error, and adjust. Add seasonality and no-show rates once the weekly habit is in place.

If you also sell on-demand or streamed classes, our guide on selling on-demand classes shows how to give members a way to train when the live schedule cannot fit them.

FAQ

What is class demand forecasting in a gym?

It is the practice of predicting how many members will attend each class so you can set the schedule, room capacity and instructor staffing accordingly. The forecast works best at the level of a specific class type, time slot and location. It turns your class schedule into a staffing plan.

How much data do I need to start?

Eight to twelve weeks of bookings and check-ins is enough for a first baseline, as long as the data is consistent. To capture seasonality you need at least a full year, and ideally two. If you have less history, forecast conservatively and review weekly.

Do I need AI to forecast class demand?

No. A fill-rate sort plus a seasonal adjustment in a spreadsheet beats guessing, and it often performs close to more complex tools at single-studio scale. AI-assisted models become useful with many locations and thousands of sessions, where they can weigh weather, holidays and cancellation behavior at once. Whatever method you use, it should be explainable and scored against actual attendance.

How do I forecast for a brand new class with no history?

Use a comparable class as a proxy: similar format, time of day and audience. Start with a conservative capacity, track the first four to six weeks closely, and treat early results as a rough guide. Replace the proxy with real data as soon as you have it.

How often should I update my forecast?

Compare predicted and actual attendance weekly, and revisit your schedule decisions monthly or at each season change. Reviewing weekly keeps errors small and helps you spot a trend before it becomes a problem.

Class Demand Forecasting for Gyms: A Practical Guide