Multi-location swim school

Meta Ads

Budget scaling test

The ad platform said doubling spend barely worked. The bookings said otherwise.

A multi-location kids’ swim school wanted to know if spending more in a market actually brings more families through the door. I doubled Meta spend in one four-school market, then measured total trial bookings against markets that held their budget flat.

+9%

Trial bookings vs last year in the test market, July to September

−19%

Same period, markets that held their budget flat

85–105

Added trial bookings in 90 days that wouldn’t have happened otherwise

2.1x

Meta spend in the test market vs the spring baseline

The question

The owner asked a fair question: is spending more a viable path for our other locations? Not a dashboard screenshot. They wanted to know if the extra money created bookings that wouldn’t have happened anyway.

Why the ad platform got it wrong

Meta’s reporting made the test look like a miss: spend up 2.1x, platform-reported bookings up 6%. But that summer every market slowed down. A new membership-hold option was routing returning families around the trial step, school calendars shifted, and consumer sentiment dipped. Platform numbers can’t tell “the ads stopped working” apart from “everything got harder.”

How I measured it

Count every booking, not just the ones Meta claims

Total trial bookings by state from Google Analytics, across every traffic source. If the spend worked, it shows up here.

Compare against markets that didn’t change

Markets with flat budgets show what the slowdown alone did. The gap between them and the test market is the lift.

Price the added bookings, not the average

Divide the added spend by the added bookings. That’s the real cost of growth, and the number that decides the next market.

Data pulled through the Windsor.ai MCP and worked through with Claude. Every number checked back against the source.

The test market kept growing while the rest of the system slid

Total trial bookings vs the same month last year

Test market (4 schools)

Markets with flat budgets

Spend increase in effect

+16%

+5%

Apr

+17%

−4%

May

+6%

−13%

Jun

+10%

−29%

Jul

−3%

−20%

Aug

+24%

+1%

Sep

+8%

−33%

Oct*

*October is month-to-date, projected to a full month. Source: Google Analytics, trial bookings by state, all traffic sources.

Before the increase, the test market ran about 18 points ahead of the flat-budget markets. After, the gap widened to about 28 points.
That works out to 85 to 105 added trial bookings in 90 days, roughly a 10 to 13% lift over where the market would have landed.
Each added booking cost about the same as an average booking in the brand’s largest market. Growth at the system’s going rate, not a premium.

More spend kept paying, until it didn’t

Meta cost per trial booking at each weekly spend level, indexed to the starting cost

Starting budget

1.0x weekly spend

1.0x

Step 1

1.8x weekly spend

1.5x

Step 1, weeks 6–12

1.9x weekly spend

2.1x

Step 2

2.7x weekly spend

3.4x

Source: Meta Ads, test market prospecting campaign.

Up to about 1.9x the starting weekly budget, the extra spend kept adding bookings.
Past that point, cost per booking jumped to 3.4x the starting level. That’s where this market topped out.
The ads didn’t break. Click-through rate rose 15% and cost per click rose only 6%. What ran out was nearby parents ready to book: the booking rate after the click fell by about half. Four schools only have so many families in range.

A second market, same result

A month later I raised Meta spend 2.3x in a second market. Its trial bookings swung from down 20% vs last year to up 18%, then up 30%. Two markets, same pattern: more spend, more bookings.

Second market: trial bookings vs last year

Before

After the increase

−24%

Jun

−20%

Jul

+18%

Aug

+30%

Sep

The playbook I handed the client

Step up 30 to 50% at a time

Hold each step 3 to 4 weeks. You see where a market starts to flatten before you blow past it.

Judge the added bookings, not the average

The first dollars in a market are always the cheapest. A blended average hides what the next dollar buys.

Read the market, not the platform

Total bookings against a comparison market. The same method behind this case study.

Fix the booking flow first

Scale markets that convert after the click. More traffic into a leaky funnel just costs more.

Stop at the value line

When an added booking costs more than an enrollment is worth, the market is full. Move the budget.

“If you only read the ad platform, you’d have cut a budget that was working.”

Bradley Zeller, Zeller Media

Client details are anonymized. Figures come from Google Analytics and Meta Ads, April to October 2026. Added bookings are estimated against markets that held their budget flat.