← Back to list
Colossus (Invest Like the Best / Business Breakdowns)Podcast13 Aug 2019Source: traffic.libsyn.comHost: Patrick O'Shaughnessy

Zack Kanter – All Things Business - [Invest Like the Best, EP.142]

In plain words

This interview is about the deep logic behind business success. Guest Zack Kanter argues Amazon's greatness isn't any single move but its 'constitutional' system that makes it faster as it grows, defying the 'big companies slow down' myth. Key examples: Amazon (uses 'six-page memos' and 'two-pizza teams' to evolve itself), Walmart (impressive but limited by physical stores), and Stripe (a payment company you pay per use, like having a 'shadow army' working for you).

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

At a Glance

This episode features Zack Kanter, founder and CEO of Stedi. He and host Patrick O'Shaughnessy delve into the success models of Walmart and Amazon and their transferability, extending to underlying business logic such as the OODA loop, network health, team size, and leverage. The most weighty judgment in the entire episode is: Amazon's success does not stem from a single strategy, but from its construction of a self-reinforcing system akin to a "constitution," which allows it to accelerate as it scales, overturning the conventional wisdom that "large companies inevitably slow down."

Topic Sections

1. Walmart vs. Amazon: From "Optimizing Fixed Space" to "Boundless Network"

Zack Kanter argues that Walmart and Amazon are essentially the same type of "optimization problem," but with vastly different constraints.

  • Walmart's Optimization: Within a fixed physical space (approximately 100,000 square feet per store), optimize SKU selection, inventory levels, pricing, and product placement. This is a "defect discovery" process—managers constantly seek and correct defects like "overpriced" or "out of stock," because "the only good news is bad news." Walmart built a massive private satellite network and logistics fleet (7,000 tractors) for this, but everything serves a fixed physical model.
  • Amazon's Disruption: Amazon's assumption is "infinite shelf space." It removed the constraint of physical space, moving Walmart's model online. As a result, Walmart has about 100,000 SKUs, while Amazon has approximately 488 million SKUs (and growing). This creates entirely new problems, such as the "discovery" challenge (easier in physical stores, harder online) and managing a vast number of suppliers and inventory.
  • Key Difference: Walmart's achievement is the limit of "intentional coordination," whereas Amazon is more like an "organism." Zack points out that Walmart still has a price advantage today, but Amazon solved the dilemma of not being able to be "the world's best at everything" by building platforms (e.g., Marketplace, AWS).
2. Amazon's "Constitutional" Architecture: Merging "Intelligent Design" and "Evolution"

Zack Kanter proposes that Amazon's success stems from its unique system that merges "intelligent design" with "evolutionary" mechanisms, making it a self-reinforcing "autonomic nervous system" that does not rely on a single hero.

  • Mechanism Breakdown: Amazon achieves "mutation" through its "six-page memo" culture, "selection" through executive meetings focused on the fitness function of "customer value," and "amplification" through "two-pizza teams." Failed memos quietly disappear, while successful ones receive resources. This explains why new Amazon products (like Amazon Prime) often initially seem "half-baked," but the company can quickly allocate capital to projects that actually work.
  • Key Insight: Zack believes Bezos's critical insight was "not just selling services to others, but also selling services to yourself." With AWS, for example, Amazon is both its first and best customer. This forces Amazon to use the same tools and interfaces as its customers, thereby "being exposed to the customer's pain," creating the fastest feedback loop (OODA loop). This design means Amazon no longer relies on "serial heroes" (like Musk for Tesla), but externalizes its culture into a runnable "program." Consequently, few discuss "who will succeed Bezos," because Bezos himself is no longer critical to the company's survival.
3. The Core of the OODA Loop is "Orientation," Not "Observe-Decide-Act"

Zack Kanter reinterprets John Boyd's OODA loop, arguing its essence lies in "Orientation," while observing, deciding, and acting are implicit in high-performing teams.

  • Redefinition: In a fighter pilot scenario, once the pilot has the correct "orientation" (understanding relative position and energy state with the enemy) via "Energy-Maneuverability Theory," the decision (to fire) and action (pulling the trigger) are muscle memory, happening automatically. Similarly, in business, if you have good intuition, observing, deciding, and acting are continuous. The real "tempo" is "orient, orient, re-orient."
  • Software's Uniqueness: Zack believes that in the software industry, "orientation" is far more complex than in other industries. Software products are "multi-dimensional puzzles," with extremely intricate details (e.g., Slack's notification logic) that are "unobservable." A physical product (like a coffee mug) can still hold water even if slightly misaligned; but software, if off by one degree, can fail completely. Therefore, software teams must have extremely high "shared consciousness," ensuring that when team members work independently, their decision-making direction aligns with the team's overall intent.
4. Building "High-Leverage" Teams: From "Army" to "Navy SEALs"

Zack Kanter argues that modern companies should aspire to be "Navy SEALs" rather than an "Army," with the key being the use of a "shadow workforce"—third-party services that charge per transaction and are continuously optimized.

  • Historical Context: The driver of human progress is "using today's best tools to invent tomorrow's tools." Giants like Google and AWS built infrastructure; Stripe built payment systems on top; and Stedi stands on the shoulders of all these giants.
  • Mechanism Breakdown: Zack uses Stedi's use of AWS Cognito (user authentication service) as an example. Stedi doesn't build its own login system; it pays "fractions of a cent" per user login. This means AWS has "hundreds or thousands of people" working on this part of Stedi's code 24/7, without Stedi incurring any debt, security responsibility, or upgrade maintenance. This is akin to having a "shadow workforce paid per transaction."
  • "Build vs. Buy" Decision Framework: Zack proposes that companies should view all operations as a "surface area," where every part is either getting better or worse. The decision criterion is: Is there an external company that can "compound" this component faster than you can? If yes (e.g., Stripe for payments, AWS for infrastructure), buy it. If no, build it yourself, or don't do it at all. This "don't do it" option is key to preventing system bloat.
5. Network Health: A Leading Indicator Beyond GMV

Zack Kanter believes the best question a VC can ask is "How do you measure network health?" because it reveals the true state beneath surface growth data.

  • Origin of the Question: Sequoia Capital's Alfred Lin (former Zappos COO) asked Zack in a meeting: "Stedi is a network. How do you measure the health of the network?" Zack found he couldn't answer with GMV or "take rate."
  • Case Analysis:
  • Airbnb: Might focus on the "percentage of listings that have never been booked." These are "unhealthy nodes" that waste user browsing time.
  • Uber: Focuses on "the number of zeros"—the proportion of users who open the app but find no nearby cars, and driver idle time.
  • Core Value: Zack argues that GMV and take rate are "outcome metrics," while network health is a "leading indicator" and a "canary in the coal mine." By defining and measuring network health (e.g., node quality, link frequency/strength), companies can proactively manage and improve the network, rather than passively waiting for financial data to deteriorate. However, beware that wrong metrics (like Facebook's "engagement") can lead to short-term behaviors like "desperate emails."

Position Moves

Position Guest Attitude Key Data
Walmart Highly appreciates its achievements, but believes its model has a physical ceiling Logistics fleet of 7,000 tractors; approximately 100,000 SKUs
Amazon Views it as a model of "constitutional" organization, bullish on its self-reinforcing ability 488 million SKUs (at time of interview); 17 leadership principles; AWS is the "first and best customer"
Stripe Views it as a model of "high leverage" tools, a case for "buy" in the build vs. buy decision 1,800 engineers; focused on the single, deep business of payments
NetSuite Views it as an example of "we can no longer do hard things," believes it is the best in its industry but has fallen behind Founded in 1998; acquired by Oracle for $11.5 billion; outdated API, running on its own data centers
Tesla Views it as an example of the "more for less" value proposition of "more value, lower price" Outperforms peers, lower price
Slack Views it as an example of success through "depth" rather than "breadth" Focused on the single function of messaging
Flexport Mentioned as an early investor, but no in-depth business analysis provided Received $1 billion in funding from SoftBank

Judgments Worth Remembering

1. "Amazon is the only company that gets faster as it gets bigger." — Zack Kanter. Its "constitutional" architecture (six-page memos + two-pizza teams + customer value fitness function) combines "evolution" with "intelligent design," achieving self-acceleration.

2. "The core of the OODA loop is 'Orientation,' not 'Observe-Decide-Act.'" — Zack Kanter. In high-performing teams, observing, deciding, and acting are implicit; the real tempo is continuous, rapid "orientation"—constantly updating the cognitive model of a complex environment.

3. "Trust is not a cause, but a result. It comes from everyone in the team having 'Fingerspitzengefühl' (fingertip feel) for their own work." — Zack Kanter. Citing German Blitzkrieg theory, he points out that trust is an emergent property of the fact that "everyone is good at their job," not something built through team-building activities.

4. "The decision criterion for build vs. buy is: Is there an external company that can 'compound' this component faster than you can? If not, build it yourself, or don't do it at all." — Zack Kanter. This "don't do it" option is key to preventing system bloat and maintaining focus.

5. "The best question a VC can ask is: 'How do you measure network health?'" — Zack Kanter. GMV is an outcome; network health (e.g., Uber's "number of zeros," Airbnb's "percentage of never-booked listings") is a leading indicator that reveals real problems beneath surface growth.

6. "Most common business advice is wrong because it's backward-looking." — Zack Kanter. Using "Blitzscaling" as an example, it was the right answer for the past decade, but once it becomes common knowledge, it's no longer an advantage. Companies need to "weave the right strategic puzzle," not copy popular theories.

7. "Software is a 'multi-dimensional puzzle'; being off by one degree leads to complete failure. A physical product (like a coffee mug) can still hold water even if slightly misaligned." — Zack Kanter. This explains why "orientation" and "shared consciousness" are more critical in the software industry than any other.

8. "The best question is: 'What words are not being spoken?'" — Zack Kanter. As a founder, his job is to "harvest the collective unconscious of the team," bringing unspoken concerns, doubts, or conflicts to the surface, thereby dismantling "systemic paranoia."

~11 min full read
Deep Analysis