Baillie Gifford is an Edinburgh investment partnership founded in 1908, famous for ultra-long-horizon, high-conviction growth investing — its early stakes in Amazon, Tesla and NIO are classics. Its "actual investors" philosophy holds world-changing companies on 5-10 year views; AUM is around $120bn. The Insights column carries its managers' investment views and thematic research.
This piece argues that active investing isn't about having more information, but about forming a genuinely different view of the future. Markets focus on surface events and short-term data, but the real edge lies in understanding the underlying system structure. The author uses BlackBerry as a cautionary tale: analysts thought the market was saturated because they defined it too narrowly (Wall Street workers needing mobile email), but the real denominator turned out to be everyone who could benefit from a connected device. He also admits a mistake on Peloton, where he mistook a temporary pandemic behavior shift for a permanent structural change. Key holdings mentioned: Shopify (the rails behind e-commerce transactions, becoming more valuable as shopping fragments across AI interfaces), Cloudflare (the gatekeeper of internet traffic, more scarce as AI-driven automated traffic grows), and Axon (the digital evidence platform, activated by policy changes and now a control point in law enforcement workflows).
One-sentence summary of the author’s current market view: The core value of active investors lies in forming independent judgments that differ significantly from market expectations, rather than in obtaining more information. [Neutral]
The author argues that the core value of an active investor lies not in acquiring more information or being smarter than others, but in forming an independent judgment about the future—one that must differ meaningfully from current market expectations and ultimately prove more useful. The author's original words: "If everybody has access to more or less the same information, what exactly is an active investor being paid to do?" The answer is not to collect more information, but to "develop an independent understanding of what the future might look like – one that differs meaningfully from the view embedded in today’s expectations." The author emphasizes that mere contrarian thinking is not an investment philosophy; only correct differentiated judgment holds value.
The author points out that markets excel at observing events and patterns but struggle to foresee the long-term changes brought about by systemic reorganization—and this is precisely the source of mispricing. The author writes: "Large changes rarely travel in straight lines. They create feedback loops. They remove constraints. They alter incentives." Using examples such as the internet transforming commerce, smartphones revolutionizing communication, cloud computing reshaping computing, and AI changing the way we work, the author notes that merely identifying these changes "isn’t much of an insight." The real challenge lies in asking "and then what?"—what new behaviors does the change unleash? Whose position is strengthened? What begins to self-reinforce? These are where investment opportunities reside.
Drawing on system thinker Donella Meadows' "iceberg model," the author divides thinking into four layers: events (visible), patterns (recurring behaviors), structure (how the system is organized), and worldview (the assumptions and values underpinning the structure). Markets are adept at analyzing the first two layers—measuring growth acceleration, comparing historical data, revising forecasts—but rarely delve into the structural level. Using the example of a company whose growth suddenly accelerates, the author notes: the event is "growth rising," the pattern is "sustained strength," but the real question is "what caused it?"—a temporary promotion, a cyclical recovery, a competitor's mistake, or a change in the underlying system (e.g., a bottleneck disappearing, a feedback loop strengthening, a new customer segment emerging)? The author concludes: "Events and patterns are outputs. They tell us what the machine has done. They do not tell us how the machine works." This is precisely why markets excel at assessing the present but struggle to imagine a structurally different future.
The author defines "structure" as "the non-fakeable physics of a business or system"—the underlying relationships and constraints that management cannot obscure with words. The author distinguishes between "story" and "machine": stories can be polished, but machines have mechanical principles. Management may claim a strong culture, customer love for products, or innovation as a core strength—but none of these necessarily reveal how the system will behave when conditions change. Structure asks different questions: Who has leverage over whom? Where does information flow? What do customers depend on? What is scarce? Where are the bottlenecks? What is self-reinforcing? How are incentives arranged? When difficult trade-offs arise, what does the organization repeatedly choose to optimize? These questions shift our focus from "what the company says" to "what the machine itself tends to do."
The author proposes that once the underlying structure is understood, three questions should be asked, the first being: "What clock is the system running on?" The author writes: "Some structures strengthen slowly. Distribution becomes denser. Data accumulates. Customers become embedded. Trust builds. A cost advantage widens. What looks unremarkable over 12 months can become extremely important over 10 years." The author concludes: "Sometimes, time is part of the investment thesis." This means that active investors need to identify structures that self-reinforce over time, rather than chasing short-term volatility.
This article provides active investors with a systematic analytical framework: moving beyond visible events and patterns to deeply understand the structures, constraints, and feedback loops that generate them. The author suggests that current market pricing of transformative technologies such as AI and cloud computing may focus too much on short-term growth data, underestimating how these technologies reshape industry structures through feedback loops. Investors should seek companies undergoing changes in underlying structure (e.g., disappearing constraints, shifting incentives, forming self-reinforcing cycles), rather than merely chasing superficial growth. Institutional perspective bias: As a long-term growth investor, Baillie Gifford's framework naturally favors structural changes that take time to reveal value; readers should be aware that this may lead to overlooking short-term risks.
The core judgment of the article is: analyzing a company should not focus on "what it is" (hardware/software), but on "where it sits in the system." The author argues that the market's conventional classification frameworks (e.g., software vs. hardware, winners vs. losers) are "the wrong unit of analysis," as they only capture surface-level traits. True value lies in understanding the system structure a company occupies and which link becomes "scarcer and more indispensable" as technology changes.
The author states: "That question rarely respects category lines at all." This means: "That question (i.e., which part of the system becomes more valuable) almost never respects category boundaries." The author uses Shopify and Cloudflare as examples to illustrate two ends of the same logic:
| Company | Surface Classification | Structural Position | Value Change Under AI Impact |
|---|---|---|---|
| Shopify | E-commerce software | Transaction "rails": payments, fraud prevention, order fulfillment, returns processing | The more fragmented the shopping interface (chat, voice, agents), the more valuable the rails every transaction must pass through |
| Cloudflare | Cybersecurity | Traffic "gate": controlling access, filtering malicious activity, ensuring reliability | The more automated traffic from AI (agents, machine-to-machine transactions), the higher the scarcity of the gate |
The author introduces the concept of "activation," referring to the process of converting latent demand into actual adoption. Observed demand does not equal underlying demand—sometimes demand is real but suppressed by "friction": cost, inconvenience, regulation, habit, lack of trust, insufficient infrastructure, or the pain of adopting something new. This demand is "latent" rather than absent.
The author states: "If the relevant constraint disappears, latent demand can convert into adoption surprisingly quickly." This means: "If the relevant constraint disappears, latent demand can convert into adoption surprisingly quickly." The author believes that a seemingly stable market can suddenly expand, a seemingly niche area can attract new groups, and a seemingly linear adoption curve can steepen sharply. Therefore, the question he often asks is: "What is this system currently trying to do but cannot—and what would happen if the obstacle blocking it disappears?"
The author defines "leverage" not as financial leverage, but as "the ability to change outcomes under resistance." This ability is often hidden: a participant may today have relatively small visible economic power but occupy a system position that becomes extremely valuable when conditions change—such as holding a bottleneck, having key information flow through it, making it difficult for customers or suppliers to bypass, setting standards, or having a business model that competitors structurally cannot replicate in pricing, reinvestment, distribution, or expansion.
The author emphasizes: "A business model establishes the economic rules the company must live under." This means: "A business model establishes the economic rules the company must live under." These rules quietly constrain certain behaviors while enabling others. Competitors can copy features or even products, but "copying an entire machine" is far more difficult. Path and leverage connect here: identifying changes that create big opportunities is one thing, but it is better to think about "who is structurally positioned to capture the released value."
The author uses Axon as an example to illustrate how structural changes simultaneously activate demand and confer leverage. The adoption of body cameras accelerated not because the need for better evidence and accountability suddenly emerged, but because the "surrounding scaffolding of policy, regulation and institutional acceptance" changed. Friction decreased, and latent demand was activated. The same logic now applies to new tools like "Drone as First Responder"—the utility is easy to imagine, but adoption depends on legality, privacy rules, and integration with existing workflows.
But activation is only half the story. Evidence.com is increasingly becoming a control point in the digital evidence workflow: data originates from Axon devices and flows into a unified system for storage, processing, and related tasks. More processes run through this hub, making switching harder, and new products attach to the installed base. Thus, two layers interact—a changing worldview activates demand, and Axon's structural position gives it leverage, determining where demand flows. The author concludes: "It isn’t enough to identify the pool of opportunity. You want to understand the plumbing: where the resulting value flows." This means: "It isn’t enough to identify the pool of opportunity. You want to understand the plumbing: where the resulting value flows."
The operational implication of the article is: investors should abandon the habit of screening companies by industry/category, instead mapping the system each company occupies, identifying where bottlenecks are shifting, and betting on those that become new control points amid change. The author explicitly criticizes market consensus (e.g., "software exposed, physical security safe") as causing "real damage." Readers should note: this is a position-holder perspective—Baillie Gifford holds Shopify, Cloudflare, and Axon, and the article itself contains narrative elements justifying these holdings.
The author uses the BlackBerry case to illustrate: when the underlying structure shifts, addressable market (TAM) estimates based on old definitions become completely invalid. The original text notes that BlackBerry's penetration rate appeared high at the time because the market was defined as "Wall Street employees who need mobile email." However, the author argues: "The problem was the denominator." — the true denominator later became "people for whom carrying a networked computing device is valuable," a scale difference of orders of magnitude. The author emphasizes: "Markets are systems. They expand when infrastructure changes, prices fall, friction disappears, business models change or a product becomes legitimate to an entirely new cohort." The core judgment: A shift in the denominator can render seemingly mathematically irrefutable conclusions almost irrelevant.
The author admits to making a mistake with Peloton — mistaking a temporary behavioral shift for a structural system change. During the pandemic, the author believed home fitness habits would become entrenched, benefiting Peloton. But the author reflects: "What I underappreciated was the deeper structure of the fitness system." — Gyms are not just equipment; they provide social interaction, identity, coaching, community, and other functions, and a key constraint (gym closures) was only temporarily removed. When gyms reopened, the old ecosystem immediately returned. The author sums up the lesson: "A shift in behaviour may not represent a change in system structure." Depth does not equal certainty; frameworks should sharpen questions, not amplify convictions.
The author argues that active management is not about predicting the next quarter or knowing more facts, but about understanding the causal structure behind visible events. The original text states: "The opportunity, as I see it, is to understand the causal structure underlying what everybody can already see." This includes: distinguishing events from the machine that generates them, identifying latent demand before constraints are removed, understanding where leverage lies (i.e., where value will flow when the system changes), and tracking these consequences far enough until the outcome differs significantly from current expectations. The author stresses this is not about contrarian positioning — consensus is often correct; the key is whether the "why" can take you somewhere not yet obvious.
The article does not point to specific companies or themes but proposes an analytical framework: look for things that "want to happen but currently cannot," identify release points when constraints disappear, and then ask "what next?" Institutional perspective bias: The author uses the Peloton failure as a self-reflection, showing that this framework is not foolproof — deep analysis can still misjudge system structure. Readers should note that as a long-term growth equity investor, Baillie Gifford's framework naturally favors finding nonlinear opportunities from structural changes, potentially underestimating short-term mean reversion risks.
| Ticker | Direction | Author's One-Sentence View | Key Data |
|---|---|---|---|
| Shopify | Hold & Observe | As a transaction "track," the more fragmented the shopping interface, the more valuable the track every transaction must pass through | No specific data |
| Cloudflare | Hold & Observe | As a traffic "gate," the more automated traffic AI generates, the higher the scarcity of the gate | No specific data |
| Axon | Hold & Observe | Worldview shifts activate demand; Evidence.com becomes a control point in the digital evidence workflow | No specific data |
| Peloton | Not specified (reflection case) | The author admits a mistake, misjudging a temporary behavioral change as a systemic structural shift | No specific data |
| BlackBerry | Not specified (negative case) | When the denominator shifts from "Wall Street employees needing mobile email" to "people for whom carrying a connected computing device is valuable," the TAM estimate completely breaks down | No specific data |