Situated-Cost Alignment

Central idea — An idea becomes a business not because it's good, but when, for each actor who has to say yes, the cost of changing drops below the cost of staying.

The short version: 3 min

Full article: 22 min

The short version#

We talk about business ideas as if their quality belonged to them: this idea is desirable, that one feasible, another viable. This article argues the opposite. Desirability, feasibility, and viability are not properties of the idea — they are relations between the idea and the concrete conditions of the people who must use it, pay for it, and produce it. Each of those people keeps their own ledger, and says yes only if the solution costs them less than the problem they already live with. Seeing a business this way — as an alignment of situated costs — changes what you look for before building: no longer "is this a good idea?" but "who already pays what, and for whom can I make the trade-off favorable?" It's a materialist, and surprisingly practical, way to know what to work on.

In three ideas#

1. What it observes. Classical frameworks treat desirability, feasibility, and viability as attributes of the idea, and so stay abstract: an idea "desirable" to a theoretical user can be too costly to adopt for a real one.
2. The shift it proposes. Those three words describe not the idea but costs paid by three different actors — user, buyer, producer. And the only reliable evidence isn't what they say, but the real cost they accept to pay now, consistent with the role they'll later play.
3. What follows. A business doesn't exist where one actor says yes, but at the intersection of three costly yeses on the same slice. And the framework itself has blind spots — it poorly lights up value that creates a desire rather than reducing a cost.

Takeaway#

Don't ask whether your idea is good. Ask who already pays a cost, what your solution would ask them to pay in return, and at what moment someone accepts a real cost that proves the trade-off has tipped. A business is born at the exact place where three people each find, for their own reasons, that changing has become cheaper than staying.

Full article#

A materialist, systemic approach to building a business#

A note on what this document is#

This text does not offer a theory that predicts which businesses will succeed. It offers a method: a framework for thinking and for advancing — for building a business and deciding what to do next. The distinction matters, and it is held to throughout.

A theory claims to predict the world; you judge it by what it can be refuted by. A method claims to organize your action; you judge it by what it keeps you from missing. We don't ask double-entry bookkeeping to be falsifiable — we ask it to drop nothing. We don't ask a surgical checklist to predict the outcome of an operation — we ask it to force attention to the right places at the right moment.

The vocabulary of "cost" used here is therefore a language, not a measurable model. Like any good language, it is universal: almost anything can be said in it. That is its strength — nothing falls outside its field — and also its danger, addressed explicitly at the end. A language is not right or wrong. It illuminates some things well and others badly. This document says which.

Central intuition#

An idea does not become a business because it is good in itself. It becomes a business when it slots into a concrete system of conditions where several actors — users, buyers, producers — each have a material reason to say yes.

A business opportunity is not a promising idea. It is an alignment of situated costs.

The pre-MVP therefore does not merely serve to "validate an idea." It serves to understand whether that idea can become a favorable trade-off for the actors its existence requires.

1. The problem with classical approaches#

Classical innovation frameworks often speak of three dimensions.

This triptych is useful, but it stays abstract. It treats desirability, feasibility, and viability as if they were properties of the idea. In fact they are relations between an idea and the concrete conditions of the actors who must use it, buy it, produce it, sell it, maintain it, and support it.

An idea can be desirable to a theoretical user yet too costly to adopt for a real one. It can be technically feasible yet too costly to produce for this team, at this moment, with the skills and energy it actually has. It can create value and still not be viable — if the one who receives the value isn't the one who pays, if the buying cycle is too long, or if the cost of selling is too high.

The materialist, systemic approach therefore drops one layer deeper:

For whom? Under what conditions? At what costs? At what risks? With what resources? Against what alternatives? On what behavioral evidence?

2. The proposed shift: from idea to system of trade-offs#

A business idea never exists alone. It runs through a system of actors. At minimum:

1. The user must accept changing something in their practice.
2. The buyer must accept paying, deciding, justifying, and taking a risk.
3. The producer must accept building, selling, maintaining, and supporting the solution.

Each of these actors has conditions of their own.

The central question becomes:

Does this idea remove more cost than it adds, for each of the actors who must say yes?

Each actor brings their own cost function and returns their own answer. A business does not exist where one actor says yes, but where all three answers meet. In other words:

A business is the place where the three costly "yeses" land on the same slice, at the same moment. Situated-cost alignment is literally that intersection.

This is why an enthusiastic yes from the user is never enough: if the buyer can't justify the spend, or the producer can't deliver without destroying itself, the intersection is empty and there is no business — only a good idea on a single ledger.

3. Definition: situated-cost alignment#

Working name: situated-cost alignment. Alternative name: Actor-Cost Fit.

An idea becomes an opportunity when, under the real conditions of the actors its adoption requires, the cost of changing becomes lower than the cost of staying in the current situation, while the cost of producing and commercializing the solution remains sustainable for the team that carries it.

This grid does not replace the classical frames. It makes them more fundamental.

On the commensurability of costs#

An obvious objection: time, risk, reputation, habit, and money are not in the same unit. How can one claim that one cost "becomes lower" than another if they don't compare?

The answer does not run through an analytical conversion of these costs into a common currency. It runs through revealed preference. You don't compute for yourself the equivalence between an hour lost and a political risk: you observe what the actor actually accepts to pay, today, for a slice of the solution. The actor does the conversion themselves, through their behavior. The price paid crushes the units.

This deserves to be stated as a principle, not merely as a measurement convenience. Reducing incommensurable dimensions to a single number is an operation no one has the right to perform on the actor's behalf. If you invent a common unit to compare their time, their risk, and their reputation, you're force-flattening quantities that aren't yours — arbitrary, and probably wrong. The actor does have that right: it's their life, their budget, their risk; for them the dimensions genuinely are commensurable, because they live them inside a single existence. Revealed preference is therefore not a shortcut approximating a true measurement out of reach. It is the one place where the reduction is legitimate. You don't delegate the computation because you can't do it; you delegate it because you're not the entity authorized to do it. Their entire cost matrix projects, in the act of paying, onto a single observable answer: yes, or no.

The cost function isn't something you compute. It's something each actor brings.

This elegance has a price that must be named: the evidence obtained this way is local. What an actor pays for a slice now does not guarantee they will adopt the whole later — the chasm between pilot and full rollout remains a hypothesis in its own right. The measurement is clean; the extrapolation is not automatically so.

A note on the temptation to formalize. This mechanism — inferring a hidden cost function from an actor's repeated behavior — has an established mathematical cousin: inverse reinforcement learning, and more broadly discrete-choice models. The temptation is strong to model the actor as an agent minimizing a cost across a graph of possibilities. But formalizing it that way silently reimports the very assumption this method gains by avoiding: that a single scalar weight summarizes the cost. The graph demands one number per edge; you must therefore pick a unit — and at that precise instant you force-settle the commensurability you had rightly just delegated to the actor. Using this kinship as an image ("we sell the least costly path, not the shortest") is illuminating. Using it as a computation engine buries the framework's blind spots under an appearance of proof. So: keep the intuition, refuse the mechanization.

4. What the pre-MVP must actually discover#

The pre-MVP does not first serve to build a minimal product. It serves to discover the conditions of existence of the future product. It must answer five questions.

4.1. What are the actors' real conditions? For the user: real day, routines, existing tools, constraints, fatigue, fear of change, saturation level. For the buyer: budget, objectives, hierarchy, decision cycle, accountability, risk tolerance, who they answer to. For the producer: skills, available energy, cash, network, sales capacity, time horizon, health, strategy.

4.2. What costs are the actors already paying? A problem is not merely something someone declares important. It is something that already imposes a cost: time, money, attention, risk, fatigue, coordination, reputation, forgone opportunity. The right question is not "Does this problem interest you?" but "What is this problem already costing you?"

4.3. What costs does the solution add? A solution never only removes a cost. It adds them too: learning, migration, habit change, trust, vendor lock-in, integration, price, political risk. A solution can be objectively better and still go unadopted, because it demands too much change.

4.4. Which trade-offs become favorable? For the user: does the relief outweigh the effort? For the buyer: does the gain justify the purchase, the risk, and the decision? For the producer: is the cost of production, sales, and support sustainable?

4.5. What behavioral evidence confirms the trade-off? Opinions are weak. Behaviors are stronger.

A hypothesis is strengthened when an actor accepts paying a real cost consistent with the role they will later have to play.

Since the whole method rests on this gesture — the actor who pays, and thereby projects their entire matrix onto an observable "yes" — three guards must frame it. Without them, the same mechanism produces false evidence with exactly the same confidence as true evidence.

• The proof-cost must be homologous to the future role. An hour of curiosity does not prove a workflow change; a yes to a free pilot does not project the same matrix as a yes to a paid one. The cost paid now is evidence only if it hurts in the same place as real adoption.
• Local is not global. The answer they bring you is evaluated at the point where you probed — this slice, today. It tells you nothing about the full-switch weights (migration, dependency, irreversibility) that weren't in the trade-off of the small slice.
• One actor is not the system. Each actor brings their function, but a business requires all three to return yes on the same slice (see section 2). Strong evidence on the user side never compensates for a structural "no" on the buyer or producer side.

This is where the framework is strongest, and it's worth underlining: in measurement, it is agnostic to motivation. Whether the actor acts out of calculation, imitation, identity, or internal politics, you don't test their reasoning — you observe a costly behavior. The "why" is irrelevant at the moment of proof. This agnosticism is a strength and it is robust. Its limit appears elsewhere, when the framework stops measuring and starts to predict where to look: see section 9.

5. The books as practical toolboxes#

The classical books are not to be discarded. They become specialized tools within a more fundamental method.

6. Central matrix#

7. Practical sequence in eight steps#

1. Frame the idea as a cost hypothesis: which current cost do we think we reduce?
2. Identify the actors: user, beneficiary, buyer, decision-maker, blocker, prescriber, internal support.
3. Collect the conditions: interviews, observation, real documents, workflows.
4. Map current costs: time, money, risk, attention, coordination, reputation, opportunity.
5. Map solution costs: learning, migration, price, decision, integration, support.
6. Compare the trade-offs: is the cost of changing lower than the cost of staying?
7. Seek behavioral evidence: time given, data shared, quote requested, pilot paid, repeated usage.
8. Decide: continue, simplify, reposition, pivot, abandon.

8. Discovery as a chain of disqualification#

Discovery should not be designed as a collection of confirmations. It should be designed as a chain of questions and experiments that give reality the chance to say no as early as possible.

A good discovery question is not one that confirms the idea. It is one that tells you whether the next question is worth asking.

Broad question about priorities
→ Gate 1: does the topic come up spontaneously or strongly?
→ Question about the real workflow
→ Gate 2: does the problem exist in actual practice?
→ Question about current costs
→ Gate 3: does the problem cost enough?
→ Question about alternatives
→ Gate 4: are the alternatives insufficient?
→ Question about the buyer
→ Gate 5: can and will someone pay?
→ Proof experiment
→ Gate 6: does the actor accept a real cost?
→ Pilot / limited build

Each gate can halt the inquiry or redirect it. That is not a failure — it is the very point of the method.

A good method does not try to protect the idea. It tries to make it meet, early enough, the reasons it should not be pursued.

This keeps you from zooming in too fast on the idea's topic. If you want to test an inventory-management tool, it may be more honest to start from general operational priorities. If inventory never appears among the top three lived problems, the hypothesis is already weakened.

The product as a costly question put to reality#

The question is not "Should we build or not build?" but "Which hypothesis are we trying to test, and what minimal experiment produces reliable evidence?"

A prototype, a spreadsheet, a landing page, a manual service, a fake button, a Figma mockup, a paid pilot, or a mini-product are not mandatory steps. They are instruments of proof.

A prototype is not a first version of the product. It is a costly question put to reality.

You therefore don't build because you want to "see if it works." You build when an important piece of evidence can no longer be obtained by a cheaper question. Before building anything, seven questions:

1. Which critical hypothesis do we want to test?
2. What answer would invalidate this hypothesis?
3. What behavioral evidence would be strong enough?
4. What experiment can produce that evidence?
5. What is the minimal cost of that experiment?
6. At what point does this experiment become premature?
7. What decision will we make depending on the result?

Don't build a product to see whether there's a market. Build, if necessary, the minimal experiment that tells you whether the next hypothesis is still worth pursuing.

9. Blind spots and slopes of the framework#

This section is not an appendix. It is the condition of honest use. A tool is judged not only by what it does when used well, but by what it inclines toward when used normally — a bit rushed, a bit tired. The framework has real virtues; its blind spots are precisely the shadows cast by those virtues.

9.1. The bias toward legible cost#

The framework makes measurable frictions easy to see — time, data entry, coordination, documented risk. It illuminates poorly the value that does not read as a reduced cost: desire, status, pleasure, meaning, aspiration.

You can always, after the fact, relabel a desire as the "lack" it fills, and therefore as a cost. But that is a linguistic move, not a prediction. For a status or aspiration good, the lack often does not exist before the product: it is the product that manufactures the lack it then relieves. The causal arrow reverses. Where the framework assumes a preexisting cost to reduce, there is in fact a created desire. The vocabulary of cost can describe it; it cannot make it appear.

Practical consequence: a user disciplined by this framework structurally drifts toward efficiency businesses and misses desire businesses — not because the framework judges them bad, but because it doesn't light them up. The lens is excellent for what releases a blocked value, blind to what produces a new one.

Guardrail to build in: an extra gate — "What if the value here weren't a reduced cost but a created desire?" — to force the gaze where the framework won't go on its own.

9.2. Discipline as refuge#

A chain of disqualifying gates is a magnificent machine for never building. There is always a next gate, a cheaper piece of evidence to seek first, a hypothesis to secure beforehand. The framework rewards caution — and caution can become the intelligent form of fear.

A false theory gets refuted; a method that misdirects does not get refuted: it silently burns the time of whoever follows it. That is the specific risk of a disqualification tool.

Guardrail to build in: an explicit clock. Define in advance the evidence threshold beyond which building becomes the cheapest experiment left. Past that threshold, seeking still-cheaper evidence is no longer rigor — it's avoidance.

9.3. The rational actor: agnostic in measurement, presupposed in prediction#

Two uses of the framework must be distinguished.

In measurement (section 4.5), it is agnostic to motivation: it observes a costly behavior without needing to explain its cause. This use is robust.

In prediction — gates 3 and 4, "where was this cost massive," which serve to know where to look — the framework implicitly reintroduces the rational actor. It reasons: "whoever pays dearly for X will switch to whatever reduces X." But an adoption driven by imitation, identity, or internal politics breaks that prediction, even when the measurement itself stays clean.

In other words: agnostic measurement, yes; search heuristic, no. The framework can register any behavior; it presupposes rationality when it tries to guess where the behavior will occur.

9.4. The infalsifiability risk — and why it is acceptable here#

For every counterexample, the vocabulary of cost can either fold the case into its language (desire → lack → cost; market size → cost of access), or dissolve it through revealed preference. That is the sign of a powerful framework. It is also, exactly, the symptom of a system that explains everything and constrains nothing.

This would be disqualifying for a theory. It is not for a language — of which we expect precisely that it can say anything (see the opening note). A ruler predicts no length; it lets you measure them all. The framework is a graduated ruler, not a thermometer.

But this immunity comes with a strict counterpart: since it cannot be refuted, the only test left is use. The question is no longer "is the framework right?" — it will always be right, verbally. The question is: applied to a real business, does it make you see something you wouldn't have seen without it, or does it merely help you put clean words on what you already saw? The framework will only have paid its own cost the day a gate makes you abandon something your intuition wanted to keep.

10. Market size is not the number of actors#

A business is the sum, over consenting payers, of the margin per actor. That sum is reached with a few whales as with a crowd: few actors can pay enormously, many can pay almost nothing. Raw count therefore has no virtue in itself, and this is an elegant consequence of the framework: what matters is the favorable trade-off, not demographics.

Two residues remain, both expressible as costs and therefore compatible with the framework:

• Reachability. Viability established for the tested actors does not prove there exists a reachable population sharing the same cost profile. This is the "cost of market access" line — present in the framework, but not to be under-developed.
• Concentration. A viability resting on a single whale is a fragile viability. It isn't size that's at stake, it's the risk of rupture.

So the right phrasing is not "size doesn't matter," but: raw count doesn't matter; the number of reachable actors sharing the same trade-off does.

11. Disqualifying conditions you don't set yourself#

The framework tests actors' conditions as given. An objection: entrepreneurship often consists of shifting conditions, not accepting them. How can you test as fixed the very thing you're about to disrupt?

The resolution rests on an asymmetry. The entrepreneur changes others' conditions, not their own. Since they want a monetary exchange, the value they produce is for someone else — therefore external to them. They can thus legitimately test as given the conditions they will themselves modify, because those aren't their own conditions. The pre-MVP measures the state of the world before their product acts on it.

This resolution holds in the standard case. It has two precise limits, best kept in view rather than masked:

• The reflexive case. Two-sided markets, network goods, standards: the value to one actor depends on the number of other actors already adopting — which the entrepreneur's success brings about. The condition being tested is then modified by their own success; it is no longer exogenous. (Handled in section 12.)
• The moving case. A condition can shift without the entrepreneur being its cause (new technology, regulation, a rising standard). A validated gate can therefore expire. The condition is external, but not fixed.

12. Network effects: stillborn, or a bootstrapping cost to place elsewhere#

A special case of the previous one, important enough to isolate.

A business whose first unit of value already requires global density cannot be born as a business. The value to the first entrant is nil as long as the network doesn't exist, and no actor can afford to carry alone the bootstrapping cost of a global network while seeking to capture it from the start. The framework doesn't merely note this: it gives its structure — no favorable trade-off exists for the first entrant as long as they carry the network's cost alone.

This does not condemn network effects in general. Two paths remain open, and the framework names them:

• Local scale first. You don't test "does the actor switch alone" (nobody does), but "does the actor pay for the value at current density": a single side of the market, an already-dense pocket, a case where the slice is worth the cost even without the network. You grow local networks, then connect them. (A single team before the network; a single city before the marketplace.)
• Bootstrapping cost paid off-market. The great universal networks (protocols, standards) were almost never born of a business. Their bootstrapping was paid by an actor not yet seeking capture — subsidy, mandate, standard, donation — then the business settled on top. The global network isn't a business you start; it's a condition someone else funds, on which a business later becomes possible.

In the framework's vocabulary: what is stillborn isn't "the network effect," it's the attempt to capture its value while carrying its bootstrapping cost alone.

13. Four paths for exploring opportunities#

The approach does not always start from an idea. There are at least four points of departure.

The most powerful case appears when several paths cross: a general shift in conditions meets a team particularly well placed to explore it in a field where the old costs are already high.

The team as a condition of the business (capability-led)#

The team itself is a condition of the business: its skills, network, health, capital, credibility, territory, and field access determine which problems it can explore at lower cost.

An opportunity is not just a problem that costs the customer a lot. It is a problem that costs the customer a lot and costs this team abnormally little to understand, prove, produce, or distribute.

An early competitive advantage can therefore be defined as a cost asymmetry in the team's favor.

This approach must not become product-led. It must stay cost-led. The wrong question would be: "We know how to do this — where can we sell it?" The right one: "In which fields do our capabilities strongly reduce the cost of understanding, proving, and producing a genuinely costly problem?"

Exploration by condition shift (condition-shift-led)#

This path starts from an innovation now available that modifies the general conditions of action. It must not be understood as pure technology push. The question is not "A new technology exists — where can we sell it?" but: "A new capability enters the conditions of existence. Which old costs does it make compressible? Which new trade-offs does it become able to make favorable?"

Available innovation
→ new capability
→ old cost reduced
→ actors who were paying that cost
→ contexts where that cost was massive
→ possible use cases
→ trade-off hypotheses
→ field evidence
→ offer / product / service

Its own gates:

• Gate 1 — Which new capability is actually available? Don't say "AI exists." Say precisely what has become faster, cheaper, more reliable, or more accessible.
• Gate 2 — Which old cost becomes compressible? Writing, entry, structuring, documentation, search, coordination, updating a system, transmission, traceability.
• Gate 3 — Who was already paying that cost? Identify those who actually bore the friction, not those who find the innovation interesting.
• Gate 4 — Where is that cost massive, frequent, and poorly solved? Not all sectors are worth the same.
• Gate 5 — What new costs does the solution add? Trust, validation, confidentiality, error, integration, habit change, vendor lock-in.
• Gate 6 — What evidence shows the trade-off has changed? Not that people are impressed, but that they hand over a real case, change a flow, pay for a pilot, use the result, or integrate it into their work.

An innovation does not directly create businesses. It modifies the conditions of existence. Businesses then appear where that modification turns an old tolerated cost into a newly favorable trade-off.

14. Illustration: voice + AI as a condition shift#

The transformation is not merely voice → text. It progressively becomes:

voice
→ transcription
→ structured data
→ correlation with existing data
→ new state of the world
→ decision / action / automation

The first break turns speech into usable information. The second turns it into a structured update of reality. Voice no longer just produces a document: it becomes an interface for updating the operational world.

"There are 12 bottles of white left, 4 cases of Coke, 3 kilos of tomatoes."
→ transcription
→ structured extraction: product / quantity / unit / location / date
→ correlation with product catalog, purchase price, theoretical stock, history
→ new state of the stock
→ stock value, anomalies, alerts, recommended order

The final result is not a clean text. It's a system that knows something new about the world.

The fourth level opens more vertical opportunities, because it requires understanding the world-objects you're updating: stock, worksite, plot, patient, intervention, customer relationship, incident, project, order.

Speech already existed, writing already existed, documentation already existed. What changes is the cost of conversion between spoken observation, structured data, and a usable state of the world.

15. The method does not stop at the pre-MVP#

The notion of a pre-MVP can mislead: it suggests a "pre-product" phase, then a "product" phase, then classical execution. In reality, the pre-MVP is the first occurrence of a discipline the company must keep for its whole life: checking that its solution remains a favorable trade-off under the real conditions of the actors who keep it alive.

A company stays alive as long as it maintains the alignment of its situated costs despite the transformation of conditions.

Conditions change constantly: users change routines; buyers change priorities; budgets move; alternatives get cheaper; market standards rise; regulation evolves; the internal cost of production or support climbs; a technology makes a new way of doing things possible; the product becomes too complex; the organization loses contact with real usage.

The risk is therefore not only mis-validating an idea at the start. The risk is believing that a validated alignment stays valid — which connects directly to the moving case of section 11: a validated gate can expire.

Permanent alignment loop#

Observe → Map → Disqualify → Prove → Realign

Continuous innovation is therefore not just adding new features. It is the capacity to perceive cost displacements within the system.

To innovate is to recompose the costs of a system so that a new trade-off becomes obvious.

16. Synthetic formulas#

An idea does not become a business because it is good. It becomes a business because it becomes a favorable trade-off for several situated actors.

The pre-MVP does not serve to validate an idea. It serves to map the conditions that could make that idea necessary, buyable, and producible.

Desirability is an accepted cost of use. Feasibility is a sustainable cost of production. Viability is a justifiable cost of purchase and a capturable value.

An opportunity appears when the cost of staying becomes greater than the cost of changing.

The strongest evidence is not what people say. It is the cost they accept to pay: time, attention, money, risk, reputation, or habit change.

A business is born from an alignment of situated costs. Innovation consists of maintaining or recreating that alignment when conditions change.

And, without forgetting the framework's own slopes:

This framework admirably illuminates value that releases a cost, and poorly illuminates value that creates a desire. It rewards disqualification, at the risk of rewarding avoidance. It is a language, not an oracle: it cannot be refuted, it is judged by use — the day a gate makes you abandon what your intuition wanted to keep.

Conclusion#

A materialist, systemic pre-MVP approach does not seek to know whether an idea is appealing in theory. It seeks to know whether that idea can inscribe itself in the real conditions of the actors who will have to make it exist.

A business is born when a solution becomes the path of least cost for several actors at once.

This approach does not reject the classical tools of the pre-MVP, lean startup, customer development, design thinking, positioning, or B2B selling. It reorganizes them around a deeper question: What costs do the actors pay today, what costs does our solution ask them to pay, and what evidence shows the trade-off becoming favorable?

There remains the one thing a framework this coherent cannot grant itself: proof of its own usefulness. It cannot be refuted; it is earned in use. You'll know it was worth its cost the day it makes you see a trade-off no intuition had seen — or give up a business that desire alone would have pursued.