← Back to list
Colossus (Invest Like the Best / Business Breakdowns)Podcast4 Jul 2023Source: joincolossus.comHost: Patrick O'Shaughnessy

Dave Yuan - A Primer on Vertical Market Software - [Invest Like the Best, EP.335]

In plain words

This piece is about investing in vertical market software (VMS)—software built for one specific industry, like restaurants or auto repair. Guest Dave Yuan says the moat comes from being the "last system a merchant would turn off"—the one holding the most important data or daily workflow. The real home run is expanding from one part of the value chain to another, e.g., from serving insurers to serving repair shops. Key holdings: Toast (restaurant POS, helped eateries survive COVID via online ordering), CCC (from auto insurance to repair shops, building an online parts marketplace), and FareHarbor (activity booking platform, charged consumers 6-7% fee, sales reps got 50% of first-year revenue as commission, later sold to Booking).

AI SummaryAI-generated · may contain errors · verify against the original

Dave Yuan (Founder of Tidemark Capital) systematically elaborated on the investment framework for vertical market software (VMS) in a podcast. Core thesis: VMS serves small and medium-sized businesses (SMBs) through industry-specific functionalities, with its moat derived from control points (e.g.,

~12 min full read · 11 sections
Deep Analysis

Dave Yuan – Vertical Market Software Investment Framework

At a Glance

Dave Yuan (Founder of Tidemark Capital, former TCV Partner, board member of Toast and Karbon) systematically articulates an investment framework for Vertical Market Software (VMS). Core thesis: The moat of VMS stems from the "Control Point"—i.e., the system that stores the most critical data, hosts the core workflow, and is the "last system a merchant would shut down"; true excess returns come from the ability to extend from a single control point upstream and downstream along the value chain.


I. Definition and Core of VMS: The Control Point

Dave Yuan argues that the essence of VMS is "software built for a specific end industry," with its core concept being the "control point."

The control point consists of three types of "gravity":

  • Data Gravity: Storing the most critical data for that industry
  • Workflow Gravity: Systems that merchants must use daily
  • Account Gravity: The last system to be shut down if all systems must be turned off

> "If you have to shut down all systems to close for business, which system is the last one you turn off? That is your account gravity." — Dave Yuan

Key Distinction: VMS is not the same as SMB. While most VMS solutions serve local merchants (restaurants, hotels, retail), there are also cases serving large enterprises—such as CCC (where Dave Yuan previously served on the board), which sells to the largest auto insurance companies in the United States.


2. Three Elements of Market Assessment: TAM, Competitive Intensity, and Industry Growth Dynamics

Dave Yuan points out that evaluating VMS opportunities requires prioritizing three variables, which determine the company's structure and investment strategy.

Assessment Dimension Core Question Impact on Strategy
TAM (Total Addressable Market) 800,000 to 1 million restaurant locations in the U.S. vs. laundromats? Large TAM resembles a horizontal market, requiring aggressive expansion; small TAM demands meticulous cultivation
Competitive Intensity Number of suppliers and market structure Large TAM typically faces fiercer competition, positively correlated with TAM
Industry Growth Dynamics Speed of industry change, frequency of merchant software replacement In stable industries, only about 5% of merchants reassess software annually, making customer acquisition opportunities scarce

Key Insight: A VMS public company CEO once described his market—"Only about 5% of merchants reassess their software each year. My job is to see every sales cycle and win 75% of them." This means VMS customer acquisition strategies are "measured," unlike the S-curve expansion of horizontal SaaS.


III. Trust: From Single-Layer to Cross-Value-Chain "Moonshot" Variable

Dave Yuan identifies "trust" as the most nuanced yet highest-return variable in VMS investing—it determines whether a company can expand from serving one link in the value chain to serving the entire chain.

Mechanism Breakdown:

  • Phase 1: Win the control point, sell adjacent products
  • Phase 2 (the true home run): Extend from one layer of the value chain to another
  • Case in point: CCC expanded from serving auto insurers to serving auto repair shops—the latter derives 70-90% of its revenue from insurance claim customers

Why Trust Matters:

  • Cross-layer selling means "selling software to both sides of a transaction"
  • Requires long-established industry trust, endogenous participation rules, and industry standards
  • The outcome gap between winners and losers can be "orders of magnitude"

> "If you go from auto insurance to auto repair shops, you are now selling software to both sides of the transaction. This industry has existed for a long time, with various interaction stages between the two sides, requiring you as a service provider to build trust over the long term." — Dave Yuan

Reader Note: Dave acknowledges this is the "most frontier" variable, difficult to underwrite, but with enormous returns once successful.


4. Customer Acquisition & Retention: VMS’s Unique Approach

4.1 Marketing: Local Network Effects and Account-Level Marketing

Dave Yuan argues that VMS marketing strategies differ fundamentally from horizontal SaaS—the core lies in leveraging local network effects.

  • Leads cannot be over-consumed: If there are only 100,000 positions nationwide, with 10,000 in Boston, a few SDRs could exhaust all leads within a year.
  • Stronger product-market fit: Building products for specific customers yields a clearer value proposition.
  • Local network effects: Like an online marketplace, “conquering city by city”—winning a flagship client creates a halo effect.
  • Opposite to horizontal SaaS: Horizontal SaaS gets harder the deeper it goes (must move beyond ICP), while VMS gets easier the deeper it goes.

4.2 Retention: Control Point-Driven, Individual Retention is Key

Dave Yuan distinguishes between gross retention and net retention, noting that VMS retention logic spans both enterprise software and consumer businesses.

Retention Type Definition Typical VMS Range
Gross Retention Customer count retention rate SMBs typically below 90%
Net Retention Gross retention + expansion from existing customers Must exceed 100% to change the economic model

Three drivers of high retention:

1. Control point gravity: The stronger the data/workflow/account gravity, the less likely customers are to churn.

2. Individual retention: If Joe fails at his lemonade stand and switches to a taco stand, retaining Joe himself counts as retention—this is a “consumer mindset.”

3. Broad and cheap customer acquisition: If gross retention is 60-65% but acquisition costs are extremely low (e.g., linear TV ads), it is still acceptable—consumer businesses often have only 20% retention.

> “If you have 60% retention, that’s catastrophic in the enterprise market—a leaky bucket. But if you have broad and cheap acquisition channels, it’s more like a consumer business. Any consumer business would trade a 20% retention rate for 60%.” — Dave Yuan


5. Multi-Product Strategy: From "One System" to "Operating System"

Dave Yuan emphasizes that VMS founders must adopt a multi-product mindset early on — this contradicts the traditional advice of "focusing on one thing at a time."

Meaning of Multi-Product:

  • Not multiple standalone software products, but "Better Together"
  • Case in point: Toast's POS system integrated with payments — no need for reconciliation, 1+1=5
  • Natural expansion direction: Front-of-house (payments, billing) → Back-of-house (employee scheduling, payroll) → Suppliers (procurement, payments) → Consumers (CRM, loyalty, demand generation)

Pricing Strategy:

  • Large TAM: Tend toward lower pricing to create a compounding flywheel
  • Small TAM: Can charge higher prices (customer receives $5 in value, pays $4)
  • The key lies in "building" ARPU — actual penetration rates for each add-on product vary significantly (e.g., payment attachment rate may be only 20%, not 100%)

6. Case Studies: FareHarbor and a Commercial Casting Tool

Dave Yuan uses two cases to illustrate the "art" rather than the "science" of VMS — the creative combination of pricing, product, and GTM.

FareHarbor (Tour Activity Booking Platform)

  • Pricing innovation: Charges consumers a 6-7% buyer fee (tourism is a low-frequency transaction, making consumers less price-sensitive)
  • Product innovation: Builds websites directly for clients during the sales cycle (traditional competitors sell software for $30/month, while FareHarbor earns tens to hundreds of thousands of dollars through buyer fees)
  • GTM innovation: Sales representatives receive half of the first year's booking revenue as commission, with almost no base salary
  • Outcome: Some sales representatives slept in vans while traveling around Hawaii until they signed up all tour operators in the region — ultimately sold to Booking Holdings for hundreds of millions of dollars

Commercial Casting Tool (Unnamed LA Company)

  • Strategy: Provides free tools to casting directors (scarce supply side), while charging actors to upload materials
  • Analogy: Similar to Ariba's model — serving the "thick end" (large pool of actors) while offering tools for free to the "thin end" (casting directors)
  • Insight: Always look for scarce resources in the ecosystem and serve the scarce side first

7. Valuation and Investment Maturity

Dave Yuan argues that the core of VMS valuation lies not in exit multiples, but in the "Location × ARPU" construction of the P&L.

Key valuation variables:

  • Location growth: The ICP (Ideal Customer Profile) is extremely important—competitive dynamics and value propositions vary significantly across sub-sectors
  • ARPU construction: The difference between a 20% payment attach rate and a 100% attach rate can turn a good deal into a mediocre one, and vice versa
  • Time horizon: The VMS market turns over slowly; the best investors need an extremely long time horizon—some of the best VMS companies have been operating for 30 years

Market maturity assessment:

  • On the entrepreneurial side: There remains ample "greenfield"—industries not yet served by software or only served by 1990s desktop software
  • On the investment side: Some growth-stage investors are beginning to treat all VMS uniformly, which Dave believes will lead to poor returns
  • A "natural tide" currently exists—when returns fall short of expectations, capital will exit, leaving behind investors who truly understand the differentiation

Mentioned Positions

Position Guest Stance Key Data
Toast Bullish (Board Member) Helped restaurants survive during the pandemic via online delivery; POS + payment integration, 1+1=5
CCC Bullish (Former Board Member) Expanded from auto insurance companies to repair shops, building an online parts marketplace—the "strongest expression" of value chain extension
SiteMinder Positive Case Integrates with 97% of global hotel PMS, disrupting traditional control points through an "integration and encirclement" strategy
FareHarbor Positive Case Charges consumers a 6-7% buyer fee; sales reps receive 50% commission on first-year booking revenue; eventually sold to Booking
AppFolio Positive Case Extends from property management to equity holders and creditors, building a "single source of truth" across the value chain
MindBody/ClassPass Research Case Integration logic of operating system + marketplace, but complex monetization issues in actual execution
ZipRecruiter Research Case From job distribution tool → building own candidate pool → heavy investment in consumer brand, proving "no shortcuts"
Square Mentioned Instant deposit feature (allowing merchants to receive funds immediately instead of waiting 3 days); very low risk but highly profitable
Intuit/Xero Mentioned Serving "one-person businesses" (FDE of 1); naturally lower gross retention, but compensated through individual retention
OneStream Mentioned Horizontal business, but the "single source of truth" concept also applies to VMS

Judgments Worth Remembering

1. Control Point Definition (Dave Yuan): The core of VMS is not "industry software" but "the last system a merchant would shut down"—composed of data gravity, workflow gravity, and account gravity.

2. 5% Rule (Dave Yuan quoting a CEO): "Only about 5% of merchants reassess their software each year"—VMS customer acquisition strategies must be "measured," unlike the S-curve expansion of horizontal SaaS.

3. Trust as a Moonshot Variable (Dave Yuan): Extending from one layer of the service value chain to another requires long-established industry trust—the gap between success and failure can be "orders of magnitude."

4. Individual Retention (Dave Yuan): If Joe fails at a lemonade stand and switches to a taco stand, retaining Joe counts as retention—VMS requires a consumer mindset, not a purely enterprise mindset.

5. FareHarbor's Triple Innovation (Dave Yuan): Pricing (charging buyer-side fees to consumers) + Product (building websites for clients during the sales cycle) + GTM (sales reps earning 50% commission on first-year booking revenue)—the combination created an advantage competitors could not replicate.

6. Multi-Product Thinking Must Be Early (Dave Yuan): VMS entrepreneurs need to "think several steps ahead"—this contradicts the traditional advice of "doing one thing well at a time," but it is an inherent requirement of VMS.

7. ARPU Construction Determines Deal Quality (Dave Yuan): The difference between a 20% vs. 100% payment attachment rate can turn a good deal into a mediocre one—investors must deeply understand the actual penetration rate of each add-on product.

8. Time Horizon Determines Success or Failure (Dave Yuan): The VMS market turns over slowly, and the best investors need an extremely long time horizon—some of the best VMS companies have been operating for 30 years; otherwise, you will be driven by multiples rather than P&L.