Abstract

Early-trading ventures cannot obtain all relevant knowledge before acting, because customer interaction and delivery frequently generate the evidence on which subsequent decisions depend. This paper develops Diagnosis Before Commitment as a principle of proportionality between the scrutiny of a decision and the exposure it creates, rather than a requirement to postpone action until uncertainty is eliminated. It distinguishes learning activity from commitment by their consequences, not by their labels or expenditure alone. Drawing selectively on effectuation, real-options reasoning and research on scientific entrepreneurial decision-making, the paper specifies exposure as a multidimensional condition involving irreversibility, obligations, resource coverage and downstream dependence. It identifies an important qualification: greater exposure does not imply that unlimited research is worthwhile, because investigation is costly, evidence may be unobtainable and delay can itself create loss. Three propositions address the timing of commitment, the accumulation of individually modest dependencies and the possibility of diagnostic overreach. The paper proposes decision-level evaluation using contemporaneous records, equivalent-attention comparators and explicit measures of both premature commitment and unnecessary delay. Its contribution is a bounded framework for examining the proportionality of early commercial decisions. It does not establish a universal optimal rule, a prediction of success or authority to displace founder judgement.

1. The problem of acting under uncertainty

A startup may need to demonstrate an offer, attempt delivery or ask a customer to choose between concrete terms before a commercially useful question can be answered. An approach that makes all action conditional on prior certainty would prohibit some of the activities that create evidence. The historical title Diagnosis Before Commitment must therefore be interpreted functionally: examination should precede or accompany a consequential commitment to the extent warranted by its exposure. It does not establish a rigid sequence in which analysis ends before action begins.

The unit of analysis is a consequential decision within the formation and early trading of a specific business configuration. The relevant horizon is the commercial obligation or cycle affected by that decision. Long-range corporate strategy, permanent founder-development programmes and prediction of eventual venture success are outside the present argument. The immediate question is whether the venture understands enough about the contemplated commitment to make its consequences inspectable.

The distinction also changes the meaning of progress. Completing an artefact, making a purchase and learning something decision-relevant are different events. One activity can produce all three, but the connection must be demonstrated. A more explicit decision can legitimately result in proceeding, changing scope, postponing or stopping. None of these outcomes, considered alone, proves that the decision process was good.

2. Relationship to existing research

Sarasvathy (2001) contrasts prediction-oriented reasoning with an effectual logic concerned with shaping what can be controlled. That distinction is relevant because early ventures may act without possessing reliable forecasts. McGrath (1999) uses real-options reasoning to examine how conditional investment can preserve opportunity while containing losses. Neither source establishes the effectiveness of the AMS method; each identifies an existing theoretical foundation that limits any claim to conceptual novelty.

Research on scientific entrepreneurial decision-making supplies a closer empirical comparator. Camuffo and colleagues’ original study (2020) and larger replication (2024) examine structured hypothesis-based enquiry. The latter combines four trials involving 759 firms, including the original trial, and reports changes in idea termination and the pattern of radical pivots. These findings concern a specified intervention in particular settings. They do not show that every additional check improves a decision, that stopping is always beneficial, or that a separate AMS intervention has been tested. The present paper contributes a different emphasis: the relation between the evidence sought and the obligations that a decision creates across a commercial configuration. It does not introduce experimentation, staged commitment or uncertainty as new ideas. Its proposed value is to make their application to near-term commercial commitments more explicit, particularly where a small visible expenditure creates a larger set of less visible dependencies.

3. Method and construct definitions

The method is targeted conceptual synthesis of the canonical practitioner doctrine and directly relevant research, followed by analysis of boundary cases and propositions. The review is selective, not systematic. Its design is consistent with treating conceptual development as an explicit research activity rather than an uncited assertion of good practice (Jaakkola, 2020). The paper reports no primary data and estimates no treatment effect.

Commitment is an action that creates an obligation, reduces feasible alternatives or makes revision materially costly. It can involve money, founder time, a customer promise, ownership, a contract, a technical dependency or the assignment of scarce capacity. A commitment need not be irrevocable; its relevance depends on what reversal would require and who would bear the consequences.

Exposure is the set of material consequences that could follow if a decision’s supporting assumptions do not hold. Exposure is not synonymous with a failure probability. A venture can have a clear view of the consequences of delayed payment while lacking a defensible probability estimate of that delay. Cash loss, time loss, liability, customer harm and dependence on a supplier should not be combined into an unexplained universal score.

Evidence is information with an identifiable source and a defensible relationship to a particular claim. Its relevance is decision-specific. A report about a market may inform context without establishing that this customer will buy this offer. A customer’s objection is evidence of that response, not automatically proof that the whole market rejects the proposition. Evidence can increase understanding while making the commercial case less favourable.

Specification is an explicit statement of what a contemplated arrangement must accomplish, under which conditions and with which material dependencies. A document can specify an inappropriate objective with great precision. Specification therefore supports scrutiny and accountability; the existence of a specification is not sufficient evidence of its commercial adequacy.

Learning activity is activity intended to change a decision-relevant understanding. Its description as a test does not make its exposure small. A pilot involving significant customer reliance or substantial unrecoverable expenditure can be a major commitment. Conversely, an apparently operational activity can generate useful learning if its result is observed and correctly interpreted.

4. A proportionality account

The principle is that the requirement to examine a decision should rise with the material exposure it creates. This is a normative governance position, not a demonstrated empirical law that more analysis monotonically improves every outcome. Its defensibility rests on distinguishing the seriousness of being wrong from the value of each additional investigation.

Exposure can be examined through five non-additive dimensions. Resource commitment concerns the cash and time made unavailable for other purposes. Reversibility concerns the cost and delay involved in changing direction. Obligation concerns what the venture owes to customers, suppliers or other parties. Dependence concerns which subsequent decisions rely on the present one. Coverage concerns whether the venture has the resources or alternatives to absorb a plausible adverse result. These are research distinctions for interpreting the existing principle, not a new AMS scoring instrument.

A decision with a small purchase price may create substantial exposure when it determines a data structure, commits a customer promise or necessitates several later purchases. By contrast, a relatively expensive but readily cancellable arrangement might preserve important alternatives. Upfront price is therefore an incomplete proxy. The analytical mechanism proposed is dependency amplification: an initial commitment becomes more consequential because other commitments become contingent upon it.

This argument must include the cost of examination itself. Investigation can consume the cash or time required to trade, allow a genuine opportunity to expire, or create false confidence through increasingly elaborate but irrelevant analysis. The decision question is not whether more evidence would be desirable in the abstract, but whether a feasible enquiry could change the choice, scope or conditions of the commitment enough to justify its burden.

Where credible probabilities and pay-offs exist, established decision analysis can compare the expected value of additional information with its cost. Where they do not, the paper does not recommend inventing numerical probabilities. A transparent comparison of plausible conditions, resource consequences and reversibility is preferable to unsupported precision. An unresolved uncertainty can remain visible while a founder decides to act.

5. The anatomy of a decision episode

The canonical sequence begins with founder conviction, separates its assumptions, identifies the decision and exposure, considers evidence and specification, and retains the observed result when the position is revised. Here that sequence is interpreted as an account of what should be made inspectable in a decision episode. It is not a command that all decisions pass through a new form or identical number of questions.

Conviction can establish the founder’s intent and objectives. It does not establish customer behaviour or delivery economics. Advice can contribute specialised knowledge, identify alternatives or suggest relevant tests. Its evidential weight depends on the claim and source, not simply the status of the adviser. Expert judgement is not discarded, but neither is it silently converted into an observation of an untested commercial fact.

The decision baseline should identify what is currently believed, why it matters, what would change the position and which uncertainty remains. Later results should be interpreted against that contemporaneous baseline. This supports accountability without assigning personality labels or retrospectively attributing motives to the founder. A changed decision may reflect new information rather than inconsistency; an unchanged decision may be justified where the new information is not relevant to its assumptions.

The framework does not authorise tasks, spending, supplier choice or progression. Those are separate operating decisions. It can make an unresolved dependency visible without directing the founder to purchase a particular remedy. That separation is especially important in a practitioner-sponsored framework where the organisation providing the examination may also sell further involvement.

6. Boundary cases in early trading

A bounded test of the window-display offer

In the fictional case introduced in The Trinity Model®, three local retailers paid £240 for a window display installed within five working days. The next decision is whether to promise that turnaround across a wider area. A capped test can reserve only a handful of delivery slots, state the actual area and terms, and record who buys, the rush and travel costs, and when payment clears. The commitment has a defined cost and endpoint. An area-wide campaign with prepaid production capacity and an unconditional five-day promise puts materially more cash, supplier capacity and customer obligations at risk.

At the decision date the evidence supports a local referral transaction. The wider route, peak delivery cost and working-capital requirement remain open. A useful record sets out the assumption behind the wider commitment, the maximum exposure in the test, what result would change the decision, and a point at which the founder reviews the evidence. The founder can preserve the option to narrow the area or revise the price. A future report can then compare the outcome with the recorded reasons rather than reconstructing those reasons from hindsight.

Other boundary cases

A reversible offer test. A fictional service founder considers two limited offers for a small number of independent customers, using available capacity and explicit delivery terms. The exercise can reveal purchasing objections and actual delivery effort. Its legitimacy does not depend on eliminating uncertainty first. The material issue is whether the commitments made to customers and the resources at risk remain understood and manageable.

A low-price technical dependency. A low-cost software subscription becomes embedded in the purchasing and delivery process. The immediate payment is modest, but moving away could require recreating records, retraining a supplier and changing customer commitments. Examination of these dependencies is relevant even though a conventional monetary threshold might not trigger review. The conclusion is not that the tool should be rejected; it is that price alone understates the decision. A perishable opportunity. A customer requires a decision before a limited delivery window closes. Additional research could reduce uncertainty but would also cause the opportunity to expire. A proportionate examination must include the consequence of delay. It may be reasonable to proceed under constrained terms with residual uncertainty. Diagnosis should not be constructed so that only inaction can satisfy it.

A genuine need for capital. Trading can expose a resource gap that cannot be covered by available cash, timing changes or a narrower configuration. The canonical requirement to establish the job for capital before considering a funding instrument does not imply that external finance is undesirable. It distinguishes an evidenced requirement from activity invented because funding is available. This paper does not assess funding products or make a borrowing recommendation.

These are analytical cases, not observations of AMS clients. Together they show why the proposed rule must respond to the configuration and the decision rather than prescribe the same sequence to every founder.

7. Propositions and rival explanations

D1. Exposure-sensitive scrutiny

When a consequential commitment has an identifiable, material uncertainty that can be investigated before the decision without destroying its value, exposure-sensitive examination will improve recognition of the conditions under which proceeding is justified. Improvement should be assessed against decision-relevant evidence, not by the volume of research or frequency of refusal.

D2. Cumulative dependence

Review of connected obligations will identify material exposure overlooked by a review that considers each expenditure independently. The mechanism is the disclosure of consequences created jointly by commitments. The proposition would be weakened where independent expenditure review identifies the same issues at lower cost, or where the added dependencies are immaterial to the decision.

D3. Diagnostic burden

Beyond the point at which feasible additional enquiry can change the current decision, further examination will increase cost or delay without a commensurate diagnostic benefit. The location of that point will depend on decision urgency, information accessibility and complexity. This proposition directly permits evidence against excessive use of the method.

Important rivals remain. Benefits attributed to the framework may reflect reviewer experience, additional founder attention or better access to specialist advice. A cautious intervention might improve immediate loss avoidance while suppressing valuable experimentation. Participants with more resources may both document decisions better and tolerate mistakes more easily. None of these explanations can be removed by reporting favourable testimonials.

8. Evaluation at the decision level

A proportionate initial study would compare decisions made from matched case materials under equivalent time and information conditions. Assessors would examine whether participants recognised the relevant obligations, distinguished source evidence from assertion, and selected scrutiny appropriate to the decision. Cases should include situations where action is warranted despite uncertainty and where delay is itself costly. Otherwise the evaluation would reward refusal by construction.

A prospective field study could retain pre-decision records, actual commitments, evidence obtained and observed early-trading results. Outcomes might include avoidable specification changes, costs of reversing commitments, missed decision deadlines and informed reduction of scope. Opportunity losses require special care: the performance of an abandoned option is usually unobserved. They cannot be inferred simply from a founder later regretting the decision.

The design should specify its comparison condition, timing, attrition handling, reviewer effects and primary outcomes before analysis. Random assignment, where feasible and ethical, would strengthen attribution. Where it is unavailable, process evidence can support a bounded contribution account but should not be relabelled as a clean causal estimate. Current government evaluation guidance provides a relevant distinction between mechanisms, implementation and impact (HM Treasury & Evaluation Task Force, 2026).

This is a proposed research design, not a demand for a continuing founder dossier. Participation, lawful access, confidentiality, retention and research governance require separate arrangements. It does not authorise use of existing private founder records for research or alter current AMS retention rules.

9. Implications and conclusion

For business administration, the paper places the quality of a near-term decision between two inadequate extremes: confidence-led commitment and analysis without a decision limit. Its contribution is not a new aspiration to “reduce risk”, but a more explicit account of what makes a commitment consequential and when further examination is useful.

For enterprise-support policy, evaluation should distinguish expenditure on support from value created for the participant. More completed assessments, more documents or fewer venture continuations are not sufficient indicators. The distribution of burden matters: an approach requiring extensive documentation may disadvantage precisely the small or resource-constrained businesses it is intended to assist. This is a potential implementation effect to examine, not an established finding about AMS.

Diagnosis Before Commitment is most defensible as a principle of proportionality applied to the present commercial decision. It permits action to create evidence and permits the founder to act with residual uncertainty. It requires greater scrutiny when consequences become more material, while recognising that scrutiny has a cost. Its effectiveness remains an empirical question; its legitimacy depends on remaining bounded, transparent and distinct from the founder’s authority to decide.

Research and disclosure

Prepared for Active Mentor® Studio as a conceptual white paper for editorial review. The Studio has a commercial interest in the framework discussed. AI assistance was used in drafting and literature discovery. This paper draws on the Studio’s v3.2 Canon and the 5 November 2026 research manuscript. It uses conceptual analysis and fictional illustrations. Human authorship and final responsibility are pending editorial sign-off.

References