This episode explains why Tracy Graham invests in small, traditional businesses in overlooked U.S. regions. His core idea: these companies are cheap (bought at ~8x profit) and have valuable customer data that almost comes for free. By analyzing that data, he can boost profits and growth. Three key holdings: an **oncology practice services firm** (owns 80% of local chemo data, margins improved); a **Midwest manufacturer** (potential target, $35M revenue, 25% margin); and a **managed service provider** (large stake, uses transaction data to spot profitable clients). The key: these firms are ignored by big-city investors, so competition is low and data is an untapped goldmine.
Tracy Graham (Founder of Graham Allen Partners) discussed in a podcast the strategy of investing in overlooked companies in second- and third-tier U.S. markets. The core argument is that technology-driven enterprises in these markets are undervalued due to a lack of attention, and transforming tradi
Tracy Graham is the founder and managing partner of Graham Allen Partners, focusing on acquiring and building technology-driven middle-market businesses. The core theme of this episode: traditional businesses in America’s second- and third-tier markets are undervalued due to their geographic location, and can generate outsized returns through data and technology transformation. The most impactful takeaway from the entire episode: Graham argues that middle-market businesses possess proprietary datasets that are not priced in—these data are valued at nearly zero at the time of acquisition, yet through data analytics and AI, they can drive significant margin expansion and growth—this is the "alpha generation engine."
Tracy Graham believes that mid- and low-end enterprises in second- and third-tier U.S. markets are systematically undervalued due to their geographic location, which represents the core arbitrage opportunity in his investment strategy.
Graham points out that these enterprises are typically located outside first-tier markets such as Silicon Valley and New York, lacking attention and competition. He describes the characteristics of typical target companies: annual revenue of $50 million to $100 million, EBITDA of $5 million to $15 million, 1-2 factories, family-run, and operating for over 15 years. These companies are "generally undervalued because of their geographic location" and "face little competition from other private equity firms."
Key data support:
Graham emphasizes that the largest unpriced asset in these companies is proprietary data. "I have never walked into a mid- or low-end market company and said, 'Can we talk to your data scientist?' only to have the company reply, 'He's in that office' — they usually don't have data scientists." This means data carries almost zero value in transactions, but Graham views it as a "free option."
Graham believes that acquiring and analyzing proprietary datasets can significantly improve target companies' margins, growth, and competitive positioning — this is the core differentiator of his investment strategy.
He illustrates the value of data using the oncology business as an example:
Three major directions for data usage:
1. Customer level: disaggregates aggregated data into transaction-level data to identify "profit drivers," "profit consumers," and "profit drags"
2. Marketing efficiency: improves target customer precision through clean data and customer understanding
3. Product improvement: collects product usage data at customer sites and feeds it back to engineering teams
Graham particularly emphasizes the value of transaction-level data: "If you only look at the product/service list, you'll find two or three customers paying the same price, but one calls customer service 50 times while another calls only twice — the support cost difference is huge, yet the pricing model completely fails to reflect this."
"Fool's gold" warning: clinical medical data appears attractive but is difficult to monetize — "On the clinical side, even if you achieve 90% accuracy, it's still worthless; on the administrative side, 90% accuracy already far surpasses 60% of competitors."
Graham views manufacturing companies as "data centers with machines," arguing that data can significantly improve product quality, customer experience, and supply chain efficiency.
He describes a typical Midwestern manufacturer: produces 1–2 core products, operates 1–2 factories, with annual revenue of approximately $35 million, an EBITDA margin of 25%, family-run, and now in its second generation.
Data application scenarios:
Graham notes that data transformation in manufacturing is "heavy lifting" — requiring sensor installation, integration of supplier data sources, and data cleaning and labeling. "Nobody has the answers. We're still developing it."
Graham argues that controlling software distribution channels (managed service providers) to tier-2 and tier-3 markets can create strong competitive moats.
He observes that in tier-2 and tier-3 markets, large software companies (such as Salesforce and Amazon) almost never sell directly — "In our region, we have never seen anyone from Amazon or Salesforce come to sell technology. There is always a middleman."
Key Insights:
Graham raises a disruptive question: "Do customers really want a tool or a service?" He leans toward the view that offering "full service" (software + implementation + support) rather than distributing through middlemen may be a superior model.
Graham directly applies the leadership principles he learned from Notre Dame football coach Lou Holtz to investment and business operations.
Holtz's three core principles:
1. Do right
2. Do the best you can
3. Treat others the way you want to be treated
Graham explains how these principles guide business decisions: "When things are going well, making decisions is easy. But the real test comes when decisions become difficult — will you still do what is right?"
Key leadership lessons:
Graham emphasizes that these principles directly apply to governance, conflicts of interest, and partnerships in investment decisions. "The people I started doing business with 20 years ago are mostly still doing business with me today."
| Position | Analyst Stance | Key Data |
|---|---|---|
| Oncology Post-Office Service Company (Acquired) | Bullish | Covers approximately 80% of chemotherapy data in the region; significant margin improvement; accelerating growth |
| Midwest Manufacturing Company (Unnamed, Potential Target) | Bullish | Annual revenue of approximately $35 million; EBITDA margin of 25%; single plant; family-run |
| Whirlpool | Neutral (Client Relationship) | Partnership of approximately 5–6 years |
| U-Haul | Neutral (Client Relationship) | Mentioned as a client of the manufacturing company |
| Managed Service Provider (Holds a Significant Stake) | Bullish | Analyzes client profitability through transaction-level data |
1. "Mid- and low-end market companies are undervalued due to their geographic location, which is the core arbitrage opportunity" (Tracy Graham) — Acquisition multiples are around 8x LTM EBITDA, far below comparable companies in first-tier markets; data assets are valued at nearly zero in transactions but can become an "alpha generation engine."
2. "Clinical medical data is fool's gold — 90% accuracy is still worthless; 90% accuracy on the administrative side already far surpasses 60% of competitors" (Tracy Graham) — Distinguish between commercializable data and data that appears attractive but is difficult to monetize.
3. "Manufacturing companies are essentially 'big data centers with machines'" (Tracy Graham) — Through sensors, connecting supplier data, and cleaning data, product quality, customer experience, and supply chain efficiency can be significantly improved; however, this is "heavy lifting," and no one has a complete answer yet.
4. "Transaction-level data reveals true customer profitability — two customers paying the same price, one calls customer service 50 times, the other calls 2 times" (Tracy Graham) — Breaking down from aggregated data to transaction-level data can identify "profit drivers," "profit consumers," and "profit drags," optimizing pricing and service strategies.
5. "Software is the most undervalued asset on the balance sheet — customers pay 100% of the cost but use only 10%-20% of the features" (Tracy Graham) — In second- and third-tier markets, customers lack data talent and need "full-service" solutions rather than "another tool."
6. "Lou Holtz's three principles directly apply to investing: do the right thing, do your best, and treat others the way you want to be treated" (Tracy Graham) — These principles guide governance, conflicts of interest, and partnerships; "when decisions get tough, you find out what people are really made of."
7. "By controlling the distribution channel for software into second- and third-tier markets (managed service providers), a strong barrier to entry can be built" (Tracy Graham) — Large software companies (e.g., Salesforce, Amazon) almost never sell directly into these markets; intermediaries control distribution.
8. "Giving a child 'exposure' can change the trajectory of generations" (Tracy Graham) — Citing the example of Chris Murphy flying him to Naples and spending 7 days together, emphasizing the power of "exposure" and "seeing a different world."