How a Startup Can Fight and Not Lose to Big Incumbents and Frontier Models: Choosing a Market You Can Corner with 7 Powers, Dominance, and Lanchester

Hello, this is Ryuta Hamamoto from TIMEWELL.

Lately I have been hearing a particular worry a lot, from people about to found a company or trying to launch a new business inside a big organisation. It goes like this: "Even if I hit on a good service, what if a well-capitalised incumbent, or one of those enormous AI model providers, starts doing the same thing and crushes me in an instant?"

Honestly, I think that worry is a sound instinct. When you trade blows on the same ground against an opponent who is ahead of you on capital, headcount, brand, and distribution all at once, the side with more stamina tends to win. This is not something willpower can overturn; it is a matter of dynamics. But that does not lead to the conclusion that a startup has no chance. Look back through history and there are any number of examples of a small company outmanoeuvring a giant. What they had in common was that they did not fight head-on, and that they had carefully chosen a narrow place where they could be number one.

In this article I first want to confirm, through real cases, why fighting head-on tends to end in defeat, and then to explain, using classic frameworks of business strategy in terms clear enough for a beginner, how to fight so that you do not lose. What I take up is Peter Thiel's case for monopoly, Hamilton Helmer's 7 Powers, the dominance strategy that developed in Japan, and Lanchester strategy. The names sound difficult, but what they are saying is very simple, so I will break each one down in turn. At the end I fuse them into a single winning line and bring it right down to the first step you can take tomorrow.

Let me note one thing at the outset. Every company I mention here is treated neutrally, as a case for understanding the dynamics of competitive strategy. I have no intention of pronouncing anyone a winner or a loser, and on matters still in dispute I take care not to assert as settled what has not been settled.

Why does "fight head-on and you lose" hold?

First, let me share the premise that is the starting point of this article: that a startup and a big incumbent are, in the first place, playing by different rules of battle.

A big incumbent has capital built up over a long stretch of time. It has the stamina to keep fighting for years even while bleeding red ink, and it can cut prices and blanket the market with advertising deeper and longer than you can. It already has a known brand, and it holds the sales network and the shelf space of distribution. It has many people, and the hours it can pour into building the same feature are on a different order.

In that state, what happens if you try to sell exactly the same product, to the same customers, in the same way? The opponent can take your profit just by cutting the price a little, and can bury you on recognition just by advertising a little more. In other words, the moment the contest is turned into one of total strength, the difference in stamina tends to show up directly in the result. This is the basic risk for a startup.

This dynamic gets backed up a little more rigorously by Lanchester's law, which I will explain later. For now, just hold on to the feeling that "competing on total strength on the same ground is a rule that is unfavourable to the weaker side." On that footing, let us look at what actually happens, through three types of case.

Case 1: The mass battle of capital (PayPay's large-scale rebate)

The first is a mass battle that uses capital itself as a weapon. The emblematic example is the "We'll Give Away 10 Billion Yen" campaign PayPay launched at the end of 2018.

The first round began on 4 December 2018. It rebated the equivalent of 20 percent of the payment amount as a PayPay bonus, with a chance that the full amount would come back, and a cap of 100,000 yen per grant. It was a large-scale cashback. The response far exceeded expectations, and because it reached the 10 billion yen set aside as its funding, it ended early on 13 December 2018, only about ten days after it started.¹ The second round ran from February to May 2019, and in light of the lessons of the first round, it set a two-tier limit: a cap per grant and a monthly total cap (50,000 yen a month).¹

What I want to note here is not the flashiness of the sums themselves. It is that behind this measure lay a clear intent: to take recognition and share all at once, in a short window, on the back of large capital. A payment service is a business whose value emerges only once both the shops that accept it and the people who use it increase. As I will explain in detail later, in this kind of field "the side that gets more people using it first, and gets the snowball rolling, is stronger." So one can read it as buying the snowball's initial momentum with money, knowing full well it meant some red ink.

What happens if a startup, thin on stamina, takes on exactly the same rebate war with exactly the same kind of payment service? It stops the moment its funding runs dry. The opponent still has several more turns to spin. This is the fearsome thing about a mass battle. Of course, this is a legitimate way for a well-capitalised company to fight, and it is not something to be condemned. What we should learn is one single point: do not go volume-for-volume in a field where the opponent can push with volume.

To go a little deeper, a mass battle works most easily in a field where it is hard to open up much of a difference in the product itself, and where victory is decided by "how many people you reached, and how fast." A service with the property that you want to use what everyone else is using, like payment, is the classic example. Because network effects, which I will explain later, work strongly in such places, the side that takes a certain scale first gains an advantage, and the side with capital goes to buy the initial momentum with money. Put the other way round, what a startup should aim for is a field where initial scale does not decide everything, or a field so narrow that the opponent hardly feels it is worth investing capital. This connects directly to Lanchester's strategy of the weak, which I explain later.

Case 2: The platform ships a feature and your footing disappears

The second is the type where, the moment the side providing the foundation ships a new feature, the reason for existence of the small players doing business on top of it thins out. This has happened again and again in the world of IT and AI.

A name known for a long time is "Sherlocking." It comes from the fact that the third version of Apple's search tool, Sherlock (in Mac OS X 10.2, 2002), built in a standard feature much like the popular third-party app Watson.² The side that made Watson took it as being "copied without permission, payment, or credit," while there was also the view that it was "nothing more than the natural evolution of Sherlock."² Which of these is right aside, from here a universal risk came to be widely spoken of among developers: that when you do business riding on a platform, the core itself can absorb your feature and turn into a rival.

The same structure came sharply into focus all at once in the age of generative AI. On 6 November 2023, at an event called DevDay, OpenAI announced features such as custom GPTs and the Assistants API.³ This in itself was a step forward for developers, but at the same time, a dynamic was widely discussed: the distinctive value of so-called "wrapper"-like services, which merely lay a thin layer of UI and prompts over the underlying model, was thinning out. The read is that each time the provider of the foundational model adds one feature, the place occupied by a small service that had been living off that feature grows narrower. That said, the very premise that "the frontier models take everything and you lose" may not stay true forever. I lay out why the rules can change in the companion piece Is Losing to Frontier Models Inevitable? Kimi K3's Comeback and Layer Strategy.

The example often cited is Jasper, a copywriting-support service. It raised a large round before ChatGPT came into wide use, but afterwards it is said to have gone through an internal downward revision of its valuation and staff cuts.⁴ I avoid asserting the fine figures, since low-confidence information is mixed in there, but as a lesson it has been drawn on repeatedly: "a thin layer that holds neither proprietary data nor deep operational integration is hard to defend against the evolution of the foundation." What matters here is not to pronounce Jasper a loser. It is far more useful to take it as a structural risk that can happen to anyone.

This point, that "what AI can easily build cannot be defended," is dug into from other angles in the companion pieces In the Age of AI, Do Not Turn "An App the Customer Could Build Themselves" Into a Product and Hand Your Proprietary Data to AI and Your "Moat" Disappears. Read them together and the contours of the moat come into sharper relief, I think.

Is your own service, or your company's, turning into this kind of "thin layer"? For those who first want to check where they stand, starting from the free AI literacy check is one good option.

Case 3: A partner becomes a rival (platform risk)

The third takes Case 2 one step further: the risk that "the very party you joined hands with becomes, before you know it, a rival." As an example that drew attention around a recent period, there is the matter of Figma and Anthropic. Since this is a topic where dispute and assessment are not settled, I write carefully, only within the range of what has been reported.

According to reporting, around February 2026 the design tool Figma and the AI company Anthropic were reported to have partnered on AI features.⁵ It was a tie-up along the lines of converting generated code into editable design. Afterwards, Anthropic is said to have shipped features overlapping with those of the design-oriented services that were its integration partners, and it has been reported that, in response, partner companies including Figma publicly voiced the concern that "Anthropic is becoming a major rival of theirs."⁵

Let me be precise here. What can be confirmed goes only as far as this: that the Figma side (and other partners) has publicly voiced concern as a party to the platform risk of a partner turning into a rival. There is no substantiation, at this point, of any trademark suit or formal legal filing. Descriptions such as "shipped a competing feature without prior notice" rest on some secondary media, and the claims of both parties and the facts of the matter remain unsettled. Please read it on the premise that no wrongdoing or breach of contract by Anthropic has been established.

Even so, this matter contains an important lesson for a startup. Joining hands with a powerful platform can be a shortcut to growth, but it is inseparable from the risk of losing your escape route when the owner of that foundation comes down into your field. Separately from whether you blame the other party, the design philosophy of "do not entrust the lifeline of your business entirely to one company's platform" makes sense as self-defence.

Case 2 and Case 3 share the same root. In both, if you provide only value that overlaps with someone else on a foundation you cannot control, your footing disappears the moment that foundation moves one step. Put the other way round, if you can hold on your own side something the owner of the foundation cannot make, such as deep know-how specific to a particular industry, proprietary data that piles up through use, or a direct relationship of trust with customers, you become harder to topple even when the foundation moves. The essence of defence can be rephrased as "pile up an asset that is yours alone, outside the foundation."

The principle of not losing lies in "monopoly and oligopoly"

What do the three cases have in common? I think it is that "if you choose the wrong ground to compete on, the difference in strength becomes the result as it is." Taking on with volume an opponent who can push with volume, competing on the same feature as the owner of the foundation, living only inside one company's garden. In each, you are fighting in the place where the opponent can bring its full force to bear.

So where should you fight? Here the thinking of Peter Thiel, an entrepreneur and investor, is worth consulting. He puts it a little provocatively: "Competition is for losers."⁶ The point is that the more fiercely competitive a market, the thinner the profit, and it tends to become mere attrition. He also says this in his book: "All happy companies are different: each one earns a monopoly by solving a unique problem. All failed companies are the same: they failed to escape competition."⁷ It is a play on the famous opening line of Tolstoy.

The "monopoly" he means here is not hogging in the bad sense. It refers to "a state where, by solving a unique problem no one else can solve, you all but occupy that field yourselves." It means choosing a place with no competition, or making a place only you can make. I often reread this, a little more softly and for real-world startups, as "oligopoly" (a state where a small number occupy most of a market). Even short of a complete monopoly, if you can occupy first or second place in a narrow field, you can escape the war of attrition.

Why is being able to reach oligopoly a source of strength? The reason is simple: you get chosen without having to force the price down. In a fiercely competitive market you keep wearing yourself out trying to be a little cheaper and a little more feature-rich, and profit thins. With thin profit you cannot make the next investment, hire good people, or improve the product at your own pace. But if you can become an irreplaceable presence in that field, you can sell at a fair price, reinvest the profit that arises into the product and the organisation, and that widens the gap further. In other words, oligopoly is less an end than "the entrance to turning a good cycle." So my view is that a startup does better to take first a small market it can occupy densely than to enter thinly into a large but crowded one. The size of the market is something you widen after you have won.

That is the principle. So how do you spot, and how do you make, that "place you can monopolise or corner"? From here I look, in turn, at three practical frameworks.

7 Powers: the seven types of durable strength

The first framework is Hamilton Helmer's "7 Powers." They are the seven types of durable competitive advantage, that is, the "moat," set out in his book 7 Powers: The Foundations of Business Strategy (2016).⁸ A moat is the ring of water that surrounds a castle; the deeper and wider it is, the harder it is for an enemy to storm in. In business it refers to elements a competitor cannot easily copy.

Let me explain the seven powers one by one for a beginner, with an example attached to each. Read with a mind to which of these you can make tell for your own idea.

Scale economies are the strength whereby the larger your scale, the lower the cost per unit. The side that makes more and sells more carries a lighter burden of unit cost and fixed cost, and a latecomer cannot turn a profit fighting at the same price. For a startup, though, winning with this from the start is hard. The knack is rather to consider whether you can bring it into a shape where "if you limit it to a narrow field, you can hold the largest scale."

Network economies are the strength whereby the more people use it, the higher the value for each individual. The telephone and messaging apps are the classic cases; the more people you can reach, the more the value leaps up. The PayPay rebate mentioned earlier can be read as a move to buy the initial momentum of this effect with money. Once the virtuous cycle of more users, higher value, still more users begins to turn, a latecomer finds it hard to catch up. The detailed typology of network effects is handled carefully in the companion piece the article on "an app the customer could build themselves".

Counter-positioning is especially important for a startup, so let me explain it a little carefully. This is the strength of adopting a new business model the incumbent cannot copy however much it wants to. Why can it not copy it? Because if the incumbent copied it, it would cannibalise its own existing breadwinner. Suppose, for instance, that against an incumbent profiting on high fees, a newcomer challenges with a free model that removes those very fees. The incumbent knows perfectly well it could "just go free too," but doing so would erase its large existing revenue, so it cannot move. This structure of "knowing yet being unable to move" is one of the few breakthroughs by which a small company can outmanoeuvre a giant. The idea is to look for a place where the opponent's strength turns, as it is, into an inability to move.

Switching costs are the strength of creating a state where, once a customer starts using you, moving to another company is a hassle and expensive. Data has piled up, it is built into the operational workflow, employees are used to it. The more these elements stack up, the harder it is for a customer to move even if a competitor ships a slightly better product.

Branding is the strength of being thought, for the same feature, "because it is from this company, I can trust it and be satisfied." It takes time to build, but once established, it lets you keep a distance from price competition.

Cornered resource is the strength of holding, yourselves alone, a scarce and favourable resource. A special patent, an exclusive contract, people or a team hard to replace, or proprietary data no other company holds all qualify. In TIMEWELL's case, for example, the accumulation of knowledge and data hard to gather easily, such as the specialist know-how we have built up in the field of economic security, is what corresponds to this resource.

Process power is the strength whereby, however hard someone tries to imitate you, they cannot copy the way of doing things soaked into the organisation. The way you produce or develop has been uniquely optimised through years of accumulated improvement, and handing over the procedure manual alone from outside does not make it run the same way. The Toyota production system, known the world over yet not so easily reproduced, is the easy-to-grasp example.

Any one of these seven is strong if you can make it tell, but when several overlap, the moat deepens sharply. What is realistic for a startup, in my view, is to aim from the design stage at counter-positioning, network effects, cornered resource, and switching costs, rather than at the time-consuming ones like scale and brand.

Dominance strategy: become the overwhelming number one in a narrow field

The second framework is dominance strategy. Originally it refers to a store-opening strategy in which a chain narrows to a region, opens stores at high density there, raises its share of that region, and builds an advantage. It is also called area dominance strategy.

In Japan it is widely known from the example of convenience-store openings. When you narrow to a region and intensively increase your stores there, several benefits arise. First, logistics and delivery become efficient. At one major chain, delivery that ran at roughly 70 vehicles a day per store in the 1970s is said to have been cut to about 9 by 2016; the more densely stores cluster in a narrow range, the more of them a single delivery run can cover.⁹ On top of that, recognition in that region rises in a short time, and the brand takes root as it repeatedly meets residents' eyes. When your share within the region rises, the gap through which a competitor could newly break in narrows too.⁹ On the other hand, there are weaknesses, such as cannibalisation, where your own stores compete for the same customers, and taking concentrated damage at once in a disaster.⁹

Now, how do you apply this to a startup's strategy? Please read this part as my view. You replace the story of physical store openings, as it is, with "narrowing the market and the customer segment." Rather than aiming thinly and widely at the whole country and every industry, you narrow the target to something like "a particular pain point, of a particular job type, in a particular industry," and pour your resources into that in concentration. Precisely because it is narrow, you can become the overwhelming number one even with few resources.

When you become number one in a narrow field, several good things happen. Word of mouth among the customers of that field circulates densely, and a reputation builds that "for this area, it is that company." Because your case studies cluster in the same industry, your persuasiveness to the next customer grows. Because the requests resemble one another, the direction of product improvement is easy to fix. In terms of 7 Powers, scale economies within a narrow range, branding, and the cornered resource of accumulated cases and data all begin to tell at once. And once you have fully secured that narrow castle, you widen out into adjacent fields one at a time. Rather than going for a wide surface from the start and losing thinly, the order of winning narrow and then widening is, I believe, the crux.

Let me make it concrete. Say you are building an AI tool for some kind of work. If you present it as "for every department in every industry," it sounds good, but to no customer does it look like "something made for us." Advertising spend and sales both scatter widely and thinly, and you get buried under a big company's comprehensive tool. On the other hand, if you narrow it down to something like "specialised in a particular kind of paperwork in a particular industry," it lands with the people in charge in that industry, the jargon and the workflow pass through as they are, and the barrier to adoption drops all at once. Because you have narrowed, your understanding of each company deepens, and the roll-out to the next company speeds up too. This cycle, where narrowness produces density and density calls the next customer, is, in my view, the crux of applying dominance strategy to a startup. You deliberately choose a narrowness so narrow that a big company thinks "there is no way we come down in earnest into a market that small."

Lanchester strategy: how the weak should fight

The third framework is Lanchester strategy. Because this teaches, most concretely of all, "how should the weak fight?", let me look at it slowly.

The basis is the mathematical models of combat that F.W. Lanchester of Britain formalised in 1914. Called "Lanchester's law," it consists of two laws.¹⁰

The first law is the "single-combat type." In close-quarters, local battles where you fight one against one in turn, losses are roughly proportional to the number of troops. That is, if 5 fight 3, only the difference of 2 tells, is the image. Even a minority can, depending on conditions, hang in.

The second law is the "mass battle, probability battle." In a wide fight where you fire at each other, losses tell as the square of the number of troops. This is decisively important. When 5 fight 3, it shows up as the difference between 5 squared, which is 25, and 3 squared, which is 9, that is, a large gap of 16. The side with more troops becomes overwhelmingly advantaged, far beyond the mere difference in numbers.¹⁰ This is the true identity of what I said at the start: "when the contest is turned into one of total strength, the difference in stamina becomes the result as it is."

The teaching drawn from here is very clear. The lower-share "weak" side must avoid the wide-area, all-out fight where the second law works. To fight in a way where volume tells as its square is to make defeat inevitable. Instead, bring it into the local, close-quarters battle where the first law works. That is, narrow the field you fight on, concentrate your resources on a single point, and put becoming No.1 in that narrow field first. Differentiate, avoid the frontal clash, and choose a place where you can win locally. This is the "strategy of the weak."¹¹

Conversely, the higher-share "strong" side is said to make it the standard play to bring things into the wide-area, all-out fight where volume tells, and to promptly copy the new moves the weak side makes, the "follow (meet) strategy."¹¹ For the strong, it is more advantageous not to let the weak run wild on a narrow ground, but to drag them into the all-out fight.

In other words, a startup and a big incumbent should, in the first place, take opposite ways of fighting. Copying a big company's way as it is will not win. The weak have the standard plays of the weak. Narrow the field, concentrate on a single point, and reliably become number one in that narrow place. These three are the first thing to be thorough about. Incidentally, it is credited to Nobuo Taoka and others that this law was systematised and spread in Japan as a strategy for management and sales.¹¹

What does this "strategy of the weak" look like when brought down to daily judgment? For instance, rather than scattering a limited advertising budget thinly across the whole country, you concentrate it on the well-read venues of a single targeted industry. In sales too, rather than chasing every possible prospect, you narrow to one industry with a high win rate and build case studies through to the end. The product, too, rather than being feature-rich and all-encompassing, solves one pain point more deeply than anyone. Every one of these is the same idea: "narrow the field, and concentrate your firepower there." Conversely, if you feel your resources are scattered, I think you may take it as a sign that you have stepped, of your own accord, into the second law's wide-area fight. Because the moment the weaker side starts fighting on total strength, its chances thin out.

Drawing strategy on paper is easy, but the moment you apply it to your own idea, it suddenly turns difficult. At TIMEWELL, through our AI consulting service WARP, we work together with you on this design of "where, and how, to fight." For those about to found a company in particular, there is WARP Entre (/en/warp/entre), which specialises in accompanying founders, and we can think through, from the idea stage, "where is the narrow market you can corner?"

Fusing the three frameworks into a single winning line

So far I have looked at the principle of monopoly and oligopoly, and at the three frameworks of 7 Powers, dominance strategy, and Lanchester strategy. They may look scattered, but in fact they connect cleanly along a single line. Please read this part as my view, as I sketch the whole shape of the winning line.

The starting point is the "choice of where to fight" that Lanchester and Thiel teach. So as not to wage an all-out fight against the strong, you narrow the field drastically. And not merely narrow: you choose a field where you can "monopolise or corner" it. A place where competition is thin, or where there is a unique problem only you can solve.

Next, in that narrow field, you become the overwhelming number one with the mindset of dominance strategy. You concentrate resources on a single point and pile up, densely, the word of mouth, the case studies, and the data of that field's customers. Precisely because it is narrow, you can raise your share even with few resources.

And in the process of securing that number-one seat, you deliberately embed the moats of 7 Powers. What matters especially is what I call "choose a field where the learning loop turns." You choose a place where a wheel turns: the more it is used, the more proprietary data accumulates; with that data the product grows smarter; because it grows smarter it is used still more. Once this wheel turns, the gap with latecomers opens automatically as time passes. Because you are number one in a narrow field, this learning loop turns faster still. In terms of 7 Powers, proprietary data as a cornered resource, network effects, and switching costs all grow here at once.

Finally, once you have secured the narrow castle and the learning loop begins to turn, you use it as a foothold and widen out into adjacent fields where your strength tells. What matters here is not to widen into an exclave. You seep, one step at a time, into the neighbour where the data, reputation, and operational understanding you hold now tell as they are. Win narrow, then widen and win. Keep this order and you will find it easier to take number one again in the places you widen into.

And into this way of widening, I would want to weave in the guard against the platform risk we saw in Case 3. Joining hands with a big platform for the sake of growth is, in many cases, a rational choice. But when you do, check that "is our moat outside that platform?" The accumulation of proprietary data, the direct relationship with customers, the deep embedding into operations, hold all of these on your own side as assets the owner of the foundation cannot easily copy. Then, even if the owner of the foundation comes down into the same field, the reasons your customers will not switch remain in your hands. It is a two-tier stance: grow while borrowing one company's garden, yet keep the lifeline itself on your own land.

Just to be safe, let me add: this winning line is not "magic that never loses." Against events such as the market itself vanishing, or the premises being uprooted wholesale, no moat is all-powerful. Even so, choosing where to fight and designing so that you make time your ally raises, distinctly, the probability of surviving, compared with charging head-on without thinking. That is my read.

To sum up, the winning line is this. Do not fight the strong head-on; choose a narrow field you can corner. There, build a moat where the learning loop turns, and become the overwhelming number one. Then widen out into the neighbour where your strength tells. This, I believe, is the realistic design philosophy for a startup that trails on both capital and technology to survive, and to win, even amid the threat of big incumbents and frontier models.

For those who want to bring this thinking further down to how an individual entrepreneur should move, do also read How to Raise Entrepreneurs Who Command AI.

A checklist for the first step

Abstraction alone will not get you moving, so let me bring it down to a checklist you can use from tomorrow. When you hit on a new idea, throw the following questions at yourself in turn.

• Is it narrowed down? Not "the whole country, every industry," but have you narrowed the target down to "which pain point, of which job type, in which industry"? If it is not narrowed, narrow it to a single point first
• Can you aim for number one there? In that narrow field, could you become first or second in share even with limited resources? Is it a place the strong cannot come down into in earnest?
• Does counter-positioning tell? Are you striking at a structure where the incumbent, however much it wants to copy, cannot move because it would cannibalise its existing business?
• Does the learning loop turn? Is it a field where the more it is used, the more proprietary data accumulates, the product grows smarter, and it is used still more? Can you make time your ally?
• Have you entrusted your lifeline to one company? Do you not have the life or death of your business held in the hands of a particular platform or trading partner? Can you endure even if the owner of the foundation turns into a rival?
• Is there a neighbour to widen into? After you secure the narrow castle, can you widen without strain into an adjacent field where your current strength tells as it is?

It is fine even if you cannot say "yes" to all of them with confidence. Rather, that it becomes clear where you are weak is itself the harvest. You can then work out how to make up for the weak parts at the design stage. Weaving them in from the start tells far more than tacking them on later.

To sum up

It ran long, so let me organise the key points.

• Wage an all-out fight on the same ground against an opponent ahead of you on capital, headcount, brand, and distribution, and the difference in stamina tends to become the result as it is. Lanchester's second law (in a mass battle, strength tells as its square) backs up that dynamic
• What the cases show are a mass battle (a move to buy share with a large-scale rebate), the risk of your footing vanishing when a platform ships a new feature, and the risk of a partner turning into a rival. All share the point that "fight in a place where the opponent can bring its full force to bear, and you are at a disadvantage"
• The principle for escaping is monopoly and oligopoly. Thiel's "solve a unique problem and escape competition" is, in reality, reread as "oligopoly in a narrow field," and that is what you aim for
• 7 Powers are the seven types of durable moat. A startup should aim, from the design stage, especially at counter-positioning, network effects, cornered resource, and switching costs
• Dominance strategy is "become the overwhelming number one in a narrow field, then widen out." The weak side's standard play in Lanchester strategy is "differentiation, local battle, single point of concentration, becoming No.1"
• The fused winning line is the order of choosing a narrow field you can corner, building a moat where the learning loop turns and becoming number one, then widening into the neighbour where your strength tells (my view)

We are already in an age where being able to build, being able to think something up, is hard to turn into differentiation. Precisely because of that, ask again at the business-design stage: "where can I fight so as to make time my ally and keep on winning?" Which narrow market can you corner, and what kind of moat can you build there? For those who want to talk through this kind of business design in concrete terms, please consult the WARP team. From the idea stage, we will work alongside you in growing it into a shape that avoids the frontal clash and can win.

References and primary sources

Footnotes

1. Overview of "PayPay" and the "We'll Give Away 10 Billion Yen" campaign (the first round began on 4 December 2018, with a 20 percent rebate plus a chance of full rebate, a cap of 100,000 yen per grant, and an early end on 13 December 2018 upon reaching the 10 billion yen funding; the second round ran February to May 2019 with a two-tier cap). Cited as established information entered via Wikipedia. https://ja.wikipedia.org/wiki/PayPay ↩ ↩²

2. Origin and concept of "Sherlock (software)" and "Sherlocking" (the circumstances in which Sherlock 3 built in a feature similar to Karelia Software's Watson as standard, and both sides of the argument, the claim of being "imitated" and the counter that it was "natural evolution"). Cited as established information entered via Wikipedia. https://en.wikipedia.org/wiki/Sherlock_(software) ↩ ↩²

3. New features such as custom GPTs and the Assistants API announced at OpenAI DevDay (6 November 2023). The source is OpenAI's official announcement. The framing that a foundational model's feature additions thin out the distinctive value of thin, wrapper-type services is treated as analysis and view by multiple commentators. https://openai.com/blog/new-models-and-developer-products-announced-at-devday ↩

4. The matter of Jasper, after a large raise, being said to have gone through an internal revision of its valuation and staff cuts in the wake of ChatGPT's spread. It is widely reported, but the fine details such as the specific valuation figure, revenue decline rate, and timing derive largely from low-confidence secondary media, so they are held in reserve as "said to be" and "according to some reports." I do not pronounce any specific company a loser, and treat it neutrally as a matter of competitive dynamics, namely the presence or absence of a defensible moat. ↩

5. The AI-feature partnership of Figma and Anthropic (around February 2026, reported by CNBC and others), and the subsequent reporting that Anthropic is said to have shipped features overlapping with those of its partners, and that partners including Figma publicly voiced the concern that "Anthropic is becoming a major rival." The primary article on the partnership reporting could not be opened directly via WebFetch in this environment (403/paywall), so it was confirmed via the headlines and summaries of search results. Some of the media reporting the rivalry are secondary sources that cannot be called major or primary. Descriptions such as "shipped a competing feature without prior notice" are from some secondary media, and the claims of both parties and the facts remain unsettled. The existence of any formal suit or legal filing is unconfirmed at this point, and I do not assert as an established fact any wrongdoing or breach of contract by Anthropic. ↩ ↩²

6. Peter Thiel, "Competition Is for Losers," The Wall Street Journal, 12 September 2014 (an excerpt/adaptation from the book Zero to One). Because the WSJ text could not be reached via WebFetch, the title and source are cited as widely known, established information. ↩

7. Peter Thiel with Blake Masters, Zero to One: Notes on Startups, or How to Build the Future (published September 2014). A passage from the book making the case for the importance of monopoly (the play on Tolstoy). The quotation is kept to the range of a gist. ↩

8. Hamilton Helmer, 7 Powers: The Foundations of Business Strategy (2016). The seven types of durable competitive advantage (scale economies / network economies / counter-positioning / switching costs / branding / cornered resource / process power). The title and author are confirmed at 7powers.com. The name of each power is based on the book's structure. https://7powers.com/ ↩

9. Overview of "dominance strategy (area dominance strategy)" (improvement of logistics and delivery efficiency through clustered store openings, higher recognition and share, and disadvantages such as cannibalisation; the example that per-store deliveries at convenience stores are said to have been cut from roughly 70 vehicles in the 1970s to about 9 in 2016). Cited as established information entered via Wikipedia. https://ja.wikipedia.org/wiki/ドミナント戦略 ↩ ↩² ↩³

10. "Lanchester's law" (formalised by F.W. Lanchester in 1914; the first law = single-combat type, where losses are proportional to the number of troops; the second law = mass battle, probability battle, where losses tell as the square of the number of troops). Cited as established information entered via Wikipedia. https://ja.wikipedia.org/wiki/ランチェスターの法則 ↩ ↩²

11. "Lanchester strategy" (the division of tactics in which the weak = differentiation, local battle, single point of concentration, No.1 principle, and the strong = wide-area battle, all-out fight, meet strategy; originally from F.W. Lanchester of Britain, systematised and popularised in Japan as a strategy for management and sales by Nobuo Taoka and others). Cited as established information entered via Wikipedia. https://ja.wikipedia.org/wiki/ランチェスター戦略 ↩ ↩² ↩³