Rehab Lab Case Study: Paid Media for a Cash-Based PT Clinic

Rehab Lab is a cash-based physical therapy practice run by Dr. Sam Wagner, with its original clinic in Green Bay/Appleton, Wisconsin and a second location in Milwaukee that opened in February 2025. It sells packages of $5,000 for 12 visits, $7,000 for 18 and $10,000 for 30, and most sales are the $5,000 package. Clinic MarketingOS has been its paid media partner since April 29, 2025. This is a paid media engagement: we run the ads, and the clinic's own team does the selling and the treating. The story below includes the month that went wrong, because that month taught more than the good ones.

What kind of engagement is the Rehab Lab case study?

The Rehab Lab case study is a paid media engagement, not a coaching install. Clinic MarketingOS has run Rehab Lab's ads since April 29, 2025, more than 15 months at the time of writing, while Dr. Sam Wagner's team sells and delivers the care. That distinction matters when you read the numbers. Revenue at a clinic moves for many reasons: staff, a new location, the owner's sales skill, the season. Ads are one input. Sam has been direct about that himself, crediting more staff for part of the clinic's year-over-year growth. What the ads can be judged on directly is cost per lead, cost per qualified lead, cost per new client and return on ad spend, and those are the numbers this page focuses on. The approach is the same one laid out in Facebook ads for physical therapy.

Where did Rehab Lab start in April 2025?

Rehab Lab started in April 2025 with about $85,000 invoiced and about $55,000 collected across both clinics in a partial month, and ads that had never really been dialed in. The practice was not small or struggling. It had two locations, a proven package model and a team. What it did not have was an ad account that reliably produced the right people, or anyone telling Sam what to make. On the first call he put it plainly: "I just need something to tell me what to do. Tell me what to create, what to put out. And then I can record the stuff." That request shaped the engagement. Sam would record, and the ad side would decide what he recorded, how it ran and what got cut. An owner on camera beats anything an agency can shoot.

What changed by October 2025?

By October 2025 Rehab Lab collected $72,000 and invoiced $110,000, Sam's third-best month. The gap between invoiced and collected is normal for a package model, where patients pay over time, and it is why we report both. Sam credits more staff for part of the year-over-year jump, which is the honest read: a clinic that adds clinicians can treat more people, and ads only help if there is capacity to see the leads they bring in. That month he also said: "I can't thank you enough for just all you've done already." We include it because it is real and on the record, not because a quote proves anything about results. The numbers do that, and November was about to make them louder.

How did Rehab Lab reach a $217,000 month in November 2025?

Rehab Lab reached a $217,000 revenue month in November 2025, beating Sam's own goal of $200,000, with a STEMwave shockwave campaign as the standout. That campaign returned $30,000 in revenue on about $2,400 of ad spend, from 96 leads and 17 plans of care. Our arithmetic on those figures: about $25 per lead, about $141 per plan of care, and roughly 12.5 times return on spend. The reason it worked is the reason wedge offers work generally. A specific device-based treatment for a specific kind of pain is an easier yes for a stranger than "physical therapy," and the clinic had a clear plan of care to put behind it. A plan of care is the course of treatment a clinician recommends after the evaluation, usually sold as a package of visits (more in our plan of care entry).

Why did the bank call November a fluke, and what was built for it?

In December 2025 Rehab Lab's bank called the November month a possible fluke and wanted proof it was repeatable, so Jake built a return on ad spend sheet for Sam to take into the bank meeting. One great month on a revenue chart looks like noise to a lender. A sheet that shows spend, cost per acquisition and return on ad spend month by month, split by location, looks like a system. The sheet reported return on ad spend two ways: on deal value (the package sold) and on cash collected (what actually hit the account), because a lender cares about the second. Return on ad spend is revenue attributed to ads divided by what the ads cost (see return on ad spend). It also tells you, not just the bank, which location and which campaign deserve more budget.

What went wrong in May 2026?

May 2026 went wrong because Facebook was optimizing for message volume instead of lead quality, and collections fell to about $40,000 two weeks into the month against about $70,000 the year before. Messages kept coming in, which is exactly what the campaign was optimizing for. But cost per qualified lead had climbed from about $11 to about $17, because Meta was finding people who like sending messages and fewer people who would buy a $5,000 package. The platform did exactly what it was told. For context, WordStream's 2025 Facebook benchmarks put the median cost per lead for Health and Fitness lead campaigns at $52.98 (WordStream), so even $17 per qualified lead was cheap by that measure, although a WordStream lead and a hand-checked qualified lead are not measured the same way. Cheap is not the point. The trend was. What caught it was not a dashboard. The clinic's admin tracked qualified leads by hand.

How was the May 2026 problem fixed?

The May 2026 problem was fixed in steps: first a qualifying question on the lead flow, then landing pages, then an AI setter, all built inside Rehab Lab's own GoHighLevel account. The qualifying question was the fast fix, because it stopped people who were never going to pay cash from counting as leads, which changed what Meta learned from. Landing pages added a deliberate step between the ad and the conversation, trading some volume for intent. An AI setter is software that replies to new leads and books them, so a lead gets an answer quickly at any hour without adding work for the front desk. Building everything in the clinic's own account was a choice, not a detail. If the engagement ever ends, the clinic keeps its pages, its automations and its data. Owners comparing tools can read about GoHighLevel for clinics.

What did cost per new client look like in June 2026?

In June 2026 Rehab Lab's Milwaukee clinic acquired new clients at $342.50 each and Appleton at about $500, while STEMwave clients cost $150 to $180 against a $2,850 plan of care. With a $5,000 package as the most common sale, $342.50 per new client is a strong number, and even $500 leaves plenty of room. The STEMwave figure was strong enough to act on: at $150 to $180 per client on a $2,850 plan, our arithmetic says each dollar spent was coming back roughly 16 to 19 times in plan value, so spend on that campaign was doubled. Different locations will always run at different costs, which is why the bank sheet splits by location. Compare these figures with what a new patient costs by channel.

What changed in July 2026?

In July 2026 two changes went in: an online program's daily budget was cut from $150 to $80, and new organic-style creative cut cost per click from $0.60 to $0.20. The online program had been spending about $4,500 a month and returning $5,000 to $6,000, a thin return of roughly 1.1 to 1.3 times spend by our arithmetic, so the budget moved toward the clinic campaigns that paid better. Cutting something that technically makes money is a hard call for most owners. It is the same call a coach makes when an athlete's accessory work is eating recovery the main lifts need. The creative change was the other half. Ads shot to look like normal posts, not ads, got clicks at a third of the old price. Cheaper clicks are not the goal by themselves, but they give the same budget more chances to find a buyer.

What can another cash PT owner take from the Rehab Lab case study?

Another cash PT owner can take four things from Rehab Lab: sell a specific offer, count qualified leads by hand, keep a return on ad spend sheet by location, and own the accounts. The STEMwave campaign beat the general campaigns because it was specific. The bad May was caught because a person counted qualified leads instead of trusting the cost per lead. The bank's "fluke" comment got an answer in the form of one sheet with every month, location and campaign in it. And the fix was built in the clinic's own GoHighLevel account, so it belongs to the clinic. None of this needed a huge budget or a clever trick. It needed weekly attention and a willingness to cut what was not working, including an online program with a thin return. For how the pieces fit together, start at the clinic ads hub.