城堡还是窝棚？Hamilton Helmer 讲哪些竞争优势真能守住

Helmer says his own view on timing has changed over the past five years. He used to think strategy should wait until a startup had found product-market fit, because there was little to do with it earlier. Talking to early-stage founders convinced him otherwise: the strategic conversations had a richness and relevance he did not expect, and founders, being practical people with little time, clearly found them meaningful.

His answer to when to think about power is simply "always". That does not mean a fully articulated plan with named competitors and pricing. At the earliest stage there are far more degrees of freedom, so the useful question is which characteristics of a business proposition might tilt it toward power being available. He stresses that you are tilting probabilities, not producing a determinate answer.

Once there is product-market fit, he says, you must understand your source of power to establish your competitive position, and in a stable phase you need to know it to defend it. It is also the foundation for what comes next, because iconic businesses often have a second or third act, such as AWS, Intel moving into CPUs or Apple moving into the iPhone. Each new act starts the process over.

Helmer calls himself a concept person and says concepts must be constrained by usefulness. Because "strategy" means almost anything in business, he narrows it to the fundamental determinants of business value, which comes down to the net present value of expected cash flows. That makes strategy a long-time-constant concept. He uses Pearl Harbor as an example of a big tactical success that was a strategic disaster for Japan, because it solved Roosevelt's problem of rallying the American public.

Power, in this frame, is an economic structure that gives durable returns, a refuge from competitors who want to eat your lunch. It has two parts: a benefit, such as lower cost or higher price, and a barrier that makes that advantage durable. He calls this the "to be or not to be" test, and says that if you pass it, it translates into good margins into the foreseeable future.

His example is Netflix and scale economies. Content is a very large fixed cost, about 50% of the cost structure each year, and it can be spread over more subscribers than a smaller rival has, which lowers cost per subscriber. At the same subscription prices, Netflix is therefore more profitable. He calls this a common type of power, but notes that these advantages are hard to achieve.

Helmer says the path to power is where the rubber meets the road, and it is the subject of his forthcoming second book. His idea of power progression holds that certain powers are available only at certain times in a business's life. Branding and process power only really become available in a late stability phase, so a startup should take them off the table.

He says brand recognition can be bought, for instance with a Super Bowl ad, and that is not power. Process power is "operational excellence on steroids" and is usually imitable. Resource power, owning something valuable by law or secrecy such as a pharmaceutical patent, applies to a different class of businesses. That leaves counter-positioning, scale economies, switching costs and network economies.

Those four are sequenced. Nearly every startup begins with counter-positioning, because product-market fit is mostly substitution, meeting an existing need in a novel way that creates more value, as Amazon did against bricks-and-mortar stores or Google against Yahoo. Without counter-positioning, an incumbent with the right capabilities can simply extend its product. Scale economies, switching costs and network economies then depend on your scale relative to competitors, and Helmer urges founders to think hard about whether they have counter-positioning to start.

Helmer agrees that the moat slide in a typical startup deck is often wrong, but he says to be kind about it. Founders need optimism to get through what they are doing. Understanding whether power exists is also genuinely hard: he says it can take him and his colleagues weeks to answer for a single well-known company, because it depends on the underlying industry economics.

Among the common mistakes, he lists claiming brand as power and claiming scale economies through data. Data scale economies are possible but rare, he says, because existing competitors often operate at scales large enough that their cost per unit is close to each other's. The curve flattens, since the typical scale economy is a fixed cost spread over more volume. He adds that his team laughs whenever someone says they have a flywheel, because the key question is materiality, not whether the effect exists.

For Helmer, network economies are network effects that are large and material, enough to create a price difference that changes future margins. Many platforms have modest network effects without network economies. He guesses Uber and Lyft fall in that group, and that Uber won through a well-played war of attrition with modest, geographically specific scale economies after an early misstep of treating the business as international. He offers this as an unformed guess, not a studied conclusion.

Helmer says a moat is roughly synonymous with the barrier part of power but not with power itself, since you can dig a moat around undesirable property. He admires Warren Buffett and Charlie Munger and thinks their way of thinking is good, but says 7 Powers is more systematic and comprehensive. He recalls that Buffett once told Bill Gates in correspondence that he could not invest in Microsoft because he did not understand the moat.

Buffett's line about economic castles protected by unbreachable moats handles the benefit by saying "castle", Helmer notes, but you still need to understand why something is a castle and not a shack. Netflix's heavy investments in UI, A/B testing and recommendations were important, but they could largely be mimicked. When Blockbuster finally moved toward a mail-DVD business, its site looked just like Netflix's because it had copied it.

Asked what individual product managers can do, Helmer says he is not a fan of the idea of a "strategically driven organization" because it blurs how this knowledge is useful at different stages. A PM in a successful business should know the source of power, since it informs what they are working for, and may spot things from the weeds that others need to hear. Ideas for entirely new businesses also often bubble up from below.

He splits a business into origination, takeoff and stability. In takeoff, with rapid growth and other entrants, a technology shift is what made product-market fit possible, and offshoots and complements keep appearing. Winning that market-share battle means incorporating new features or recognizing when a new segment becomes attractive, which he calls meat-and-potatoes work and potentially decisive.

To get better at thinking strategically, he suggests reading the book and then discussing with colleagues whether the company really has a given kind of power. He says Reed Hastings had him train Netflix's top hundred people in strategy through classes inside the company, before the book existed. He no longer has time to do full courses.

Helmer says he currently sees no change to the seven powers from generative AI. The technology does raise the familiar questions in new forms: a billion-dollar fixed cost of building a model could create scale economies, a model that learns across users could create network economies, and one that learns about you could create switching costs. He finds those questions useful for judging which business models might work.

He describes three kinds of plays around a major technology: the companies that are the technology, like Intel for microprocessors; those that would not exist without it, like Microsoft; and those that use it but existed before and after, like carmakers. He is uncertain, but suspects generative AI's biggest impact will be in that third group, comparable to electricity, which let factories be reconfigured but took a lot of redesign, investment and learning.

He compares it to 1990s business process re-engineering, a big consulting opportunity, though he thinks AI feels different from crypto because there seems to be a real ultimate use case. He notes that if a 50% improvement in programming efficiency is not uncommon, that alone is worth a lot of money. Asked about an eighth power, he says he is always looking, since finding one would also be a great investment opportunity, but so far he considers seven an empirical, not theoretical, exhaustive set.

On the common pitch that a startup has the best team and moves fastest, Helmer cites Michael Porter's controversial claim that operational excellence is not strategy. In an end state, those capabilities can be mimicked through consultants with best practices or by hiring from competitors. He adds that in the takeoff phase, when you are trying to attain competitive position, operational excellence is everything, so it is essential to a strategy even if it is not itself power.

To count as power, operational excellence has to be material and also opaque, meaning rivals cannot easily see or copy what is going on. He points to TSMC bringing up a new fab, where know-how lies in trained staff and is not necessarily documented, and to Toyota in his book, but says he does not know whether TSMC has process power. In a stable business, most of your day is spent on these issues and should be, because otherwise a competitor will overtake you, but that is a treadmill and not power.

Lenny puts it to him that the only things that create value are power, market size and operational excellence, and that many people will find that surprising. Helmer replies that both views are right: business involves a multitude of things to pay attention to, but they all fall into those three categories, which he says simply comes out of the math.

Asked about broad trends affecting founders, Helmer says he is very concerned about the debt trajectory of the United States and many other countries. There has been a crisis roughly every ten years, from the dot-com bust to the financial crisis and COVID, and he fears a future one arriving without dry powder, meaning the ability to deficit spend. In that case capital markets could lock up and company formation would suffer.

He argues that the problem is hard to solve because it sits at the crux of the dance between capitalism and democracy, with entitlements as the underlying driver. One camp sees capitalism as producing inequality that government must offset, while the other sees creeping socialism, and the result is a deadlock in which taxes are not raised and spending is not cut. He calls it a rant and apologizes for it.

On the hopeful side, he calls himself an optimist and cites Joseph Schumpeter's view that an economy's vitality depends on entrepreneurs. He believes creativity and action are how society advances, and that a free society and Silicon Valley have huge advantages in that. His final advice is that action is the first principle of business: his book offers guideposts, not instructions, and the real work is doing things.