Latest / The Buyer & Seller Podcast with Fexingo: Business Brokers, Exits, and Private Sales Explained / How a Medical Billing Business Sold at 75x EBITDA with Software Data
Transcript
- Lucas: Seventy-five times EBITDA. That's the multiple a medical billing company out of Nashville closed at about six weeks ago. The business was doing roughly $2.3 million in profit, and the total consideration was north of $170 million. Luna: Seventy-five times is eye-popping even by the standards of this show. We've talked about 50, 60, even 70 times for waste haulers and packaging distributors with take or pay contracts. But a medical billing company? Lucas: Right, because on the surface, medical billing is a service business. You've got coders, billers, claim follow-up staff — it is people-heavy. And service businesses typically trade at 5 to 8 times EBITDA. The reason this one hit 75 is that it wasn't really a service business anymore. It had become a software business with a service wrapper. Luna: So the multiple was attached to the software, not the billing service itself. What was the software doing that was so valuable? Lucas: The company had built a proprietary revenue cycle management platform over about seven years. It automated claim scrubbing, denial prediction, payer-specific code optimization — and crucially, it aggregated data across every provider it served. By the time of sale, the platform had processed claims from over 4,500 providers and had a decade's worth of denial patterns, reimbursement rates, and code-level rejection data. Luna: That's the data moat we talk about. But in healthcare, that data is especially sensitive. Did the buyer care about HIPAA compliance as a risk or as a barrier to entry? Lucas: Both. The buyer — a mid-market private equity firm that focuses on healthcare services — saw the compliance infrastructure as a massive moat. Any competitor would have to build the same SOC 2 and HIPAA frameworks, recruit the same payer relationships, and accumulate years of clean data. That's not a twelve-month project. It's a five-to-seven year slog. And in the meantime, this company's dataset was getting more valuable every quarter. Luna: So the buyer was essentially paying for a ten-year head start. That starts to justify 75 times. But what about the recurring revenue piece? Was it month to month or multi-year contracts? Lucas: Mostly annual contracts with auto-renewal clauses, but the real stickiness was operational. Once a medical practice integrates its billing workflow into your platform — maps its fee schedules, trains its front desk, configures denial workflows — switching costs are brutal. The company had a 97 percent gross retention rate over five years. Net retention was around 112 percent because they upsold analytics modules and denial management add-ons. Luna: That net retention number is actually higher than most SaaS companies. And it's coming from a business that looks like a service business on paper. That's the magic combination for buyers right now. Lucas: Exactly. Private equity firms are desperate for what they call 'software-enabled services' — businesses where the labor component is real but the margin expansion comes from proprietary technology. The Nashville company had a gross margin of 62 percent, which is low for pure software but very high for a staffing-heavy service. And the buyer saw a path to 75 percent gross margin within 24 months by automating more of the denial appeal process. Luna: Let's talk about the denial appeal automation. That seems like a nuance, but it might be the core of the valuation. What was their approach? Lucas: They had built a natural language processing layer that read denial letters from insurers, extracted the reason code, matched it against historical appeal templates that had worked for similar denials, and generated a first-draft appeal letter. Human billers reviewed and submitted it. But the machine did 80 percent of the work. That cut the average appeal cycle from 14 days to 3 days, and improved the win rate by about 22 percent because they were using language that had worked before. Luna: So the data was compounding. Each denial and each successful appeal made the model smarter. That's a classic data network effect. And in healthcare, where denial rates are running at 10 to 15 percent for many practices, that's a huge value driver. Lucas: And the timing was perfect. Medicare Advantage denial rates have been climbing — some audits show initial denial rates above 20 percent for certain codes. Practices are desperate for any edge. This company's platform was effectively an insurance policy against revenue leakage. Buyers love that because it's counter-cyclical. When the economy slows and elective procedures dip, providers squeeze every dollar out of claims. The platform becomes more valuable, not less. Luna: That's a great point. I want to ask about the buyer's calculus on future interest rates. We're sitting here in July of 2026. The Fed has held rates steady for two meetings, but the market is pricing in a cut later this year. Does a 75x multiple still make sense if rates stay higher for longer? Lucas: The buyer's underwriting assumed a 10-year Treasury around 4.25 percent — roughly where it is today. They used a discount rate of about 11 percent. At 75 times, the implied earnings yield is about 1.3 percent, which looks terrible against Treasuries. But they weren't buying current yield. They were buying future cash flow growth. Their model assumed the company could double EBITDA in three years by expanding to larger hospital systems and layering in more automation. If that happens, the effective multiple on year-three EBITDA drops to the high 30s, which is more reasonable. Luna: So the bet is that the software takes over more of the work, margins expand, and the multiple compresses as earnings grow into it. That's a common PE thesis, but it only works if the growth actually materializes. Did they have a track record? Lucas: They did. Revenue had grown at a compound annual growth rate of 34 percent over the prior three years, and EBITDA had grown faster — 41 percent — because the software was taking share from labor costs. The company had also just signed its first two health system clients, each with over 500 providers. That was a step-change in scale. The buyer was essentially paying up for the right to ride that wave. Luna: It strikes me that this deal is a perfect example of something we've touched on before: the most valuable businesses are the ones that look like one thing but are actually another. A medical billing company that's really a data and AI company. Lucas: Exactly. And it's why we keep coming back to these case studies. The surface-level story — 'medical billing sold for 75 times' — sounds like a market anomaly. But when you peel it back, the logic is actually quite rational. The buyer identified a hidden asset class: proprietary healthcare claims data with a software layer that makes it increasingly indispensable. Luna: And a business that gets more valuable as the regulatory environment gets more complex. Medicare Advantage denials, prior authorization requirements, coding changes — every new complexity makes the platform more necessary. Lucas: Right. And speaking of things that get more valuable over time — this show. We've now done 93 episodes of these deep dives, and we deliberately don't run ads. No sponsors, no commercials. It's a choice we made early on because we wanted these conversations to be as clean and specific as possible. If those conversations have moved your work forward in some small way — given you an idea for how to position your own business, or helped you understand what a buyer might actually pay for — and you want to support that ad-free choice, the link is buy me a coffee dot com slash fexingo. Luna: Yeah, and it's not about rewards or perks. It's just a way to keep this exactly what it is. No pressure, ever. Lucas: Back to the deal. One more detail I think is instructive: the earnout structure. The seller — the founder who built the platform — took $50 million in cash upfront and the rest in a mix of seller notes and equity in the acquiring entity. The earnout was tied to retention of the existing client base, not to new sales. That tells you the buyer was most worried about client churn during the transition. Luna: Smart. Because the data moat only works if the clients stay. If even a few large practices leave, the denial prediction model starts to lose accuracy. The earnout aligned the founder on what mattered most. Lucas: Exactly. And the founder is still there as a chief technology officer. The buyer wanted him to keep building the software, not just coast through the earnout. So the earnout had a second component: a bonus for hitting product milestones on the hospital system integration module. Luna: That's a two-part earnout — one for retention, one for innovation. That's more sophisticated than what we usually see. Most earnouts are just a revenue target. Lucas: It reflects the buyer's thesis. They weren't buying a static revenue stream. They were buying a platform that could be the backbone of revenue cycle management for large health systems. The earnout was designed to keep the platform evolving while the founder was still fully engaged. Luna: So what's the one takeaway for a business owner listening who might have a similar hidden asset? Maybe a service business with proprietary data or software? Lucas: I think it's this: if your business generates data that is unique, hard to replicate, and gets more valuable the more clients you serve, you may be sitting on a much higher valuation than your EBITDA multiple suggests. But you have to package it. The Nashville company didn't sell itself as a billing service. It sold itself as a data platform with a billing service attached. That reframing was worth an extra 50 turns of EBITDA. Luna: And the data has to be clean, structured, and defensible. If it's scattered across spreadsheets and paper records, it's not an asset. Lucas: Right. The buyer's due diligence team spent three weeks just auditing the data architecture. They wanted to confirm that the denial history was properly tagged, that the code mappings were accurate, and that the data could be integrated into their existing portfolio companies. That audit passed, and that's when the multiple went from an initial indication of 45 times to the final 75. Luna: So the data diligence itself created value. That's a lesson for anyone preparing for a sale: invest in data hygiene before you go to market. Lucas: Absolutely. This company had a dedicated data engineer for two years before the sale, just cleaning and structuring historical claims data. That was a six-figure investment that returned tens of millions in valuation. Not a bad ROI. Luna: Not bad at all. And it's a reminder that in 2026, the businesses that command premium multiples are the ones that have already done the work to make their intangible assets tangible. Lucas: Exactly. And that's the kind of work we'll keep digging into on this show. Thanks for listening.