Nine locations, one lead engine, and a show-up rate that held

By Caleb Nguyen, co-founder & account operations lead · Published June 30, 2026 · Updated August 10, 2026

A fitness franchise with nine locations ran Meta lead campaigns the way most franchises do: one campaign per gym, each managed by whoever had time. Cost per lead averaged $43 against a 2025 cross-industry Facebook average of $27.66, and every location swore its market was just expensive. Five weeks later the blended CPL was $26, and the leads still showed up.

The numbers, before and after

The row that mattered most to the franchise owner is the one that did not move. A falling CPL with a collapsing show-up rate is a spreadsheet win and a front-desk loss; we tracked both from day one.

Why nine campaigns underperformed one

Nine campaigns meant nine budgets too small to learn. Meta's delivery system calibrates per ad set, and most locations generated too few weekly leads for the system to exit its learning phase. It also meant nine copies of the same creative fatiguing in parallel: Meta's research puts the conversion drop at roughly 45% once a person has seen the same ad four times, and in a nine-gym metro footprint the audiences overlap heavily.

We rebuilt to two campaigns: one Advantage+ leads campaign, the format Meta launched in February 2025, covering all nine trade areas, and one small campaign for the two locations with genuinely different offers. Location assignment happens by the lead's stated preference on the form, not by campaign walls.

How lead quality was defended

Instant forms convert cheaply and attract tire-kickers just as cheaply. Three changes kept quality flat while volume grew: a conditional question ("Which location would you actually visit?") that forces one deliberate choice, a phone-number field with verification on, and a CRM webhook that feeds show-up outcomes back into our two-ledgers report weekly. Optimization stayed on qualified bookings rather than raw form fills.

Caleb Nguyen has run franchise lead programs since 2019, and his rule from that work applies here: report cost per shown-up visit to the owner, and cost per lead only to the ad platform. The two numbers discipline each other.

How the counting was set up

Before launch, the owner and both founders signed the same one-page counting rules every engagement gets: a lead counts when the CRM records a booked visit with a verified phone number, the show-up rate is measured against booked visits, and the attribution window stays at 7-day click plus 1-day view for the duration. The CRM webhook made those outcomes visible weekly, so "cheap leads" could be tested against "people who walked in" from the first report onward. Without that plumbing, the price-led static below would have looked like the winner.

What did not work

Two of the twelve creatives never earned their slot. A price-led static ("memberships from $X") pulled cheap leads whose show-up rate sagged to 44% in its first 200 leads, so it died in week 3 despite the best CPL on the board; the counting rules exist precisely to kill that ad. And an AI-generated avatar walkthrough, tested out of curiosity about Meta's 2026 creative tools, hooked well at 28% and converted poorly. Members apparently want to see the actual gym. Both kills are in the client's log with dates, because a track record that only lists wins is a brochure.

Creative volume did the rest

The account went from 4 recycled creatives to 12 live ones, refreshed on the weekly calendar: member-story video, trainer-led walkthroughs, and offer-led statics, tested on hook rate and cost per qualified booking. The best performer, a 20-second member story, held a 32% hook rate against the 25% average from published creative panels.

The structure follows the same consolidation logic as the furniture restructure, applied to lead generation. The full method lives on the management service page, and the lead-cost context is in the August 2026 benchmarks.