Gym ads on Meta: trial offers, lead forms and the cost per new member

Gym ads on Meta do not sell memberships. Nobody signs a year-long contract from a feed; they book a free trial, a first class or a consultation, and the membership is sold at the front desk or on the booking page after the visit. Gym Facebook ads and Instagram ads therefore have one job: to put a visit with a date in the calendar, close enough to the club and soon enough that the person still wants it when the day comes.

That changes what the ad says and what the campaign counts. The ad names the offer, the place and the next step: a free week, the first class, a consult with a coach, at this club, book a slot. The campaign counts booked trials, trials that turned up and memberships started, and reads each against spend. A low cost per lead on form fills that never visit is a cost with no member behind it.

Fitness ads from a national chain and personal trainer ads from one coach run on the same unit: the visit. The difference is scale. The chain repeats the structure per club, the coach runs one ad set in one radius, and both are judged on cost per new member, which the last sections of this guide build from the sourced figures.

A club's audience is the people who can get there. Meta's location targeting reaches people who live in, have recently spent time in or go often to countries, regions, cities and postal codes, and when you enter an address or use Drop pin, Meta chooses a radius for you by default that you then adjust; radius selection is not possible on ZIP or postal codes, states, regions or countries. The default option shows ads to people who spend time in your locations, such as people living there or recently there, which for a gym is the right one: a commuter who works near the club is as much a prospect as a neighbor.

Set the radius from where members actually come from, which the membership software knows by postcode, and test two radii as two ad sets rather than guessing. When you exclude a location, who sees the ad depends on whether their current location or their home location is excluded, so a club near a boundary should check both before cutting the neighboring town.

The calendar is the second lever. Clubs sell on the year's resolutions, the run-up to summer and the September return, and the January offer is the one campaign most clubs repeat. Meta's learning phase sets the cost of each restart: an ad set exits learning after about 50 results in the week after its last significant edit, and edits that meaningfully change how it might perform send it back in, so a January campaign rebuilt every week never leaves learning. Prepare it in December, launch once, and change the creative inside it rather than the ad set.

• Radius from the members' postcodes, not from the map
• Two radii as two ad sets when unsure
• One January campaign, launched once, creative rotated inside it

Meta's lead ads with instant form are designed to help you generate and qualify leads by asking people to fill out a form, with contact information such as name, email address and phone number prefilled when the person has already provided it to Meta, and at least one question required. Three form types exist: more volume, the default, designed to generate a larger number of leads; higher intent, which adds a review screen where people confirm their information and is recommended when follow-up or further qualification is required on each lead; and rich creative, which adds images and text about the business. A higher intent form is only delivered to Facebook Feed and Instagram feed on mobile devices.

For a gym the form wins on speed and volume and loses on the one thing a trial needs: a time. A booking page that shows the schedule and takes a slot produces fewer leads and more visits, and because the visit happens on your site, the lead, the click ids and the membership that follows can be tied together and sent back to Meta from your server. The form keeps the lead with Meta: downloadable for up to 90 days from submission, after which it is no longer available, and Meta recommends downloading leads often or using a CRM so none are lost.

The rule for gym lead generation: use the higher intent form when the front desk calls every lead within the hour and books the trial on the phone; use the booking page when the trial needs a slot, a waiver or a card, or when you want the membership credited to the ad that brought it. When unsure, run both as two ad sets for two weeks and compare cost per trial that turned up, not cost per lead.

A radius of a few miles is a small audience, and Meta will show the same people the same trainer again and again. The signs are the ones the creative fatigue guide on this site describes: frequency up, click-through rate down, cost per trial up on the same ad, read together rather than one at a time. A high frequency with a stable cost per trial on a retargeting audience of past visitors is expected; a falling click rate at low frequency is a weak message, not fatigue.

The remedy is a rotation of angles rather than a new logo treatment: the class at its busiest hour, one trainer to camera, a member's story told with their consent, and the price said plainly. The trail table sets those angles against the variable each one tests and how a verdict is read. Film at the club, with members who agreed in writing, and keep each ad to one angle so the test stays one variable.

Meta's health and wellness advertising policy sets the limits on the transformation angle: no statements of inferiority about physical appearance, no close-up on a specific body area with pinched fat, weight loss products and services for adults 18 and over only, with the ad allowed to show the impact of use or a clear indication of the time it takes to achieve noticeable results. A member's story framed on what they do now, in their words and with their permission, stays inside it.

Meta restricts what websites in some categories may share through the pixel and the Conversions API. Its help page on data sharing restrictions based on source categories describes three levels: core setup, which restricts the sharing of custom parameters and parts of URLs following the domain; a restriction on certain standard events, which restricts the sharing of specific mid and lower funnel events; and full restrictions, which fully restrict the sharing of all events in specific regions or all regions. Meta assigns categories itself and may assign a data source to a category that comes with restrictions.

The page does not name fitness, gyms or studios, so whether a club's site is treated as health and wellness is Meta's classification, not a rule you can read off a list. Check it in Events Manager: open the dataset, the Settings tab, then Data restrictions, where core setup shows as on or off; if Meta turned it on you cannot turn it off, and if you think the category is wrong you can ask Meta to review it. Meta's page on prohibited information names diseases, medical conditions and injuries, medical procedures and treatments and mental health among the data it does not want sent, and says event names and custom conversion criteria must not reflect or imply them.

For a gym this means three things. Send standard events, Lead, StartTrial, Subscribe and Purchase, with value and currency, which core setup leaves alone as predefined parameters. Put nothing about a condition, an injury or a program for it in an event name, a custom parameter or a URL path: a class built around an injury is a class on your site and a plain Lead in the dataset. And check the dataset's category before a launch rather than after a week of Lead events that never arrived.

LocaliQ's 2026 Facebook advertising benchmarks give the price of a lead. On the leads objective, the sports and recreation row converts 6.75% of clicks at $1.08 each for a cost per lead of $15.49 with a 3.16% click-through rate; the health and fitness row converts 7.98% of clicks at $1.84 for $27.11 at a 3.09% click-through rate; the average across industries is $27.39 at an 8.54% conversion rate. LocaliQ's 2026 search benchmarks put sports and recreation on Google Search at $2.77 per click, a 7.69% conversion rate and $44.26 per lead, and health and fitness at $6.17, 6.94% and $67.36, so a trial request from Search costs close to three times what it costs on Facebook in the sports row, for a person who typed the search.

A lead is a trial request. The formulas above take it to cost per visit by the share of trials that turn up, and to cost per new member by the share of visits that join, both from your own front desk. Then the ceiling: the monthly fee, times the months a member stays, times the share of the fee left after running costs, is the most a new member can cost before the ads lose money, and the target sits below it. Nothing in a benchmark tells you the two shares, which is why a club that tracks only form fills cannot know whether $15.49 is cheap.

Budget follows from the learning phase. Meta's help page says an ad set exits learning after about 50 results in the week after its last significant edit, and recommends a budget large enough to get enough total results while avoiding frequent budget changes; at $15.49 per lead that is about $770 a week for one ad set, at $27.11 about $1,360. A club that cannot fund that on one ad set should run one ad set, not three, since Meta also notes that with many ads and ad sets the delivery system learns less about each one.

• Weekly budget for one ad set = cost per trial × 50
• Cost per new member = cost per trial ÷ show-up share ÷ join share
• Ceiling = monthly fee × months retained × share left after running costs

Meta learns from the event you send. If the dataset receives a Lead for every form fill, the delivery system buys form fills; if it receives the trial that turned up and the membership that started, it buys those. The Conversions API exists for events the pixel cannot see, a visit at the front desk or a direct debit set up a week later, and a matching event_id lets Meta count a pixel event and a server event as one.

Adrails does the server side. The script on the club's site credits the visit to the campaign and ad and keeps the click ids for 90 days; your server, or the handler that receives the membership software's webhook, sends a trial_started event when the trial is booked and a subscription event with the monthly fee as its value when the membership starts, each signed with a server key that never goes in a page. Adrails sends them to the dataset you chose as StartTrial and Subscribe with the event_id, the value and currency, fbc and fbp and the hashed email and phone, holds back anything over a value limit or a daily limit per platform, and, when memberships are billed through Stripe, reads what each member pays next to the spend that brought them. The gym use case page walks through the per-club workflow.

Read five numbers once a week, in order: spend, trials booked, trials that turned up, memberships started, and cost per new member, for the last 7 days against the previous 7. Cost per trial up with frequency up is fatigue: rotate the angle. Trials flat and visits down is the follow-up or the booking flow: call faster, shorten the form. Visits flat and memberships down is the front desk or the offer, and no ad change fixes it.

Compare with the published benchmark once a quarter, not every week, and compare with your own January last year every January. Keep one ad set per offer per club so the 50 results are not split, and change creative inside it rather than rebuilding it. A club that holds this reading for a quarter knows its two shares, and from then on a cost per lead benchmark is a reference rather than a target.