A Pre-Decision Research Checklist for an Investment Idea

A pre-decision research checklist should establish whether you understand an idea well enough to evaluate it. It should not promise that a decision can be completed in a fixed number of minutes. Some questions take one document; others remain unresolved after substantial work. Start by naming the decision under consideration and the information it depends on. Investigating a company's revenue durability is different from deciding whether a particular security fits an existing portfolio.

This article offers an original evidence-gating framework, not a validated investment strategy or a substitute for individual advice. Investor.gov notes that the SEC does not vouch for the accuracy of company filings. The linked guide provides context for reading them. Your checklist therefore needs interpretation and cross-checking, not merely a box confirming that a filing exists.

Primary-source context: Investor.gov: How to Read a 10-K/10-Q.

Write the central claim in plain language without promotional adjectives. Replace a world-class growth opportunity with an explicit proposition such as existing customers will continue buying at prices that cover the company's operating costs. Next to the claim, list the evidence you have actually inspected and the most important missing evidence. Record the relevant period, units, and source location so a later reader can reconstruct your reasoning.

Separate an unavailable answer from a negative answer. If a company does not disclose the customer measure you wanted, the conclusion is that this particular question remains unresolved. It is not automatically evidence of misconduct or business weakness. Conversely, an attractive narrative does not fill the gap. Give unresolved assumptions their own place in the checklist, with a note explaining how heavily the overall idea depends on them.

Ask what other explanation could produce the same observation. Revenue growth might reflect additional customers, higher prices, an acquisition, or a combination. Your checklist should identify which explanation matters to the claim and whether the available evidence distinguishes among them. A useful challenge question is: What would I expect to observe if my preferred explanation were wrong?

Then define a research boundary. Some gaps may justify continued study; others prevent meaningful evaluation. For example, uncertainty about a minor expense category may be tolerable for an exploratory comparison, while uncertainty about which legal entity issued the security may make the comparison unusable. Do not mechanically total check marks. A single foundational gap can matter more than ten completed administrative checks. Record the reason for that judgment so the boundary cannot quietly move when enthusiasm rises.

Consider fictional Meridian Services. Its reported revenue rises from 80 to 100 currency units, a 25% increase. Your initial idea attributes this to stronger demand from existing customers. You have also noted a recent acquisition, but you have not determined its contribution. The checklist exposes the missing bridge between reported growth and the specific business explanation you want to test.

For an illustrative scenario, suppose additional disclosure attributes 18 units to the acquired business. The remaining increase is 2 units, or 2.5% of the original 80. That arithmetic does not establish an organic growth measure because pricing, currency, disposals, and reporting definitions might still matter. It does show why the original 25% figure cannot independently support your existing-customer claim.

The recorded outcome is research incomplete: reconcile comparable revenue and inspect the definition of the acquisition contribution. No transaction follows from passing or failing this example. The checklist has clarified what must be understood next.

Check the identity of the issuer and instrument; the exact proposition under review; the reporting dates and units; the source behind each important number; and the distinction between company statements and your own assumptions. For every central assumption, name at least one observation that would weaken it. Include an alternative explanation and state which document or data field could help distinguish the two.

Record any unresolved financial obligations, cash-flow questions, valuation assumptions, and implementation constraints relevant to the proposed decision. These headings are prompts for investigation, not a universal definition of completeness. Mark items as supported, contradicted, or unresolved, and include a short reason for each status.

Finish with a written outcome: continue research, obtain specialist input, defer, or proceed to a separate decision assessment. Include the date and the circumstances that would reopen the checklist. Preserve the completed version rather than overwriting it later.

A pre-decision checklist works better as a sequence of gates than a collection of equally weighted boxes. First establish what is being evaluated. Then establish whether the evidence is comparable. Only then assess what the evidence suggests about the central claim. A missing definition near the beginning can invalidate a later calculation even when every arithmetic step is correct. The order should reflect dependencies in the argument, rather than the order in which documents happened to arrive.

For a hypothetical equipment company, a claim about recurring service revenue may depend on separating maintenance contracts from installation projects. If the disclosed service category combines both, a polished comparison of service growth cannot answer the recurring revenue question. The checklist should stop that particular inference while allowing other questions to proceed. It need not suspend all research on the company. Write the boundary precisely: aggregate service revenue is observable; recurring maintenance revenue is unresolved.

Distinguish essential gates from useful context. An industry overview might enrich the investigation, but it cannot compensate for uncertainty about the numerator in the main calculation. Conversely, a missing contextual detail should not necessarily block a narrow comparison that does not depend on it. For each unchecked item, finish the sentence: this prevents me from concluding blank. If you cannot identify a consequential missing conclusion, the item may belong in optional context rather than the evidence gate.

Extend the invented Meridian Services example with explicitly assumed, nonoverlapping revenue contributions. Starting revenue is 80 units. An acquisition adds 18, changes in prices add 5, lower comparable sales volume subtracts 2, and currency translation subtracts 1. The bridge reaches 100: 80 plus 18 plus 5 minus 2 minus 1. This construction explains the entire 20 unit increase under the stipulated categories. It does not describe a real issuer or establish a standard organic growth definition.

Now compare that bridge with the original claim that growth reflected stronger demand from existing customers. The total increase is 25%, but the assumed volume contribution is negative. Pricing might still reflect a valuable commercial capability, yet it is a different explanation from selling more to existing customers. Even the volume contribution would need a customer scope before it could answer that precise question. A businesswide volume decline can coexist with an increase within a particular customer cohort.

The worksheet therefore needs separate rows for reported arithmetic and interpretation. The arithmetic row says the stipulated bridge reconciles without a residual. The interpretation row says the existing customer demand claim remains unsupported and may conflict with the available volume explanation. A further source would need to identify the relevant cohort and measurement basis. This example shows why completing a bridge can make a story less persuasive without establishing a complete opposing investment thesis.

Sometimes you can investigate the importance of missing information without pretending to know its value. Construct a sensitivity exercise and label every assumed input. Imagine a fictional project with annual contribution before a particular service cost of 30 units. The unresolved cost might be 12, 22, or 32 in three hypothetical scenarios. The resulting balances are 18, 8, and negative 2. These are conditional arithmetic results, not probabilities, forecasts, or a distribution of likely outcomes.

The exercise identifies which missing information deserves priority. If the research claim requires the balance to remain positive, the highest cost scenario challenges that claim. The next question is whether the cost definition and evidence can narrow the range. Do not choose the middle number because it feels balanced. A midpoint is another assumption unless the evidence supports it. Likewise, do not average scenarios with equal weights simply because you have presented three of them.

Record the boundary at which the conclusion changes: in this simplified construction, cost of 30 consumes the entire contribution. That is an algebraic threshold created by the example, not a recommended investment cutoff. Add the omitted factors alongside it, such as other expenses or timing differences if they matter to the question. Sensitivity work is useful when it tells you which uncertainty is decisive. It becomes misleading when a tidy table substitutes for investigating that uncertainty.

An alternative explanation deserves more than a sentence saying that conditions could worsen. It should account for the observation that initially attracted you. If reported revenue increased, propose another mechanism that could produce the same increase. Then specify an observation that would distinguish the mechanisms. For Meridian, acquisition contribution is an alternative to stronger comparable demand. A revenue bridge can test the distinction more directly than collecting another favorable description of the enlarged business.

Use a hypothetical challenge session with three roles, even if one person performs them at different times. The claimant states the narrow proposition. The challenger identifies its weakest necessary assumption. The recorder writes what evidence would resolve the dispute. These roles organize the reasoning; they do not imply an independent review if the same researcher performs all three. A second person who reads only your summary is also not independently confirming numbers they have never inspected.

End the challenge with an update to the checklist. An objection may be answered, remain open, or reveal that the claim was too broad. Do not treat every objection as equally consequential. A remote possibility with no bearing on the proposed mechanism should not displace a known gap in the main evidence. Explain why the unresolved objection matters and which conclusion it constrains. The goal is a testable disagreement, not a long list of generic reasons to be cautious.

Prepare a one page summary with five short fields: decision under consideration, central claim, evidence inspected, unresolved dependencies, and permitted next step. Attach the supporting calculations separately. The summary should allow a reader to identify what the research establishes without rereading every note. It should not hide uncertainty behind a single overall label such as approved. A packet can support continued analysis of a business while remaining insufficient for an assessment of a specific instrument.

For each important number, add a compact provenance line: document, location, period, units, and whether the value is reported or calculated. For each judgment, add the assumption that connects it to the evidence. In the Meridian packet, 25% revenue growth belongs under calculation; stronger customer demand belongs under interpretation. Keeping those fields distinct prevents a later reviewer from treating an inference as though it appeared directly in a company document.

Adapt the remaining prompts to the decision. Comparing operating models may require different information from examining a security's cash flows or an implementation constraint. Do not copy a checklist from another instrument and mark irrelevant fields complete merely to satisfy the template. Replace irrelevant prompts with a sentence explaining the scope. Include an expiry condition tied to material changes in the evidence, definitions, or decision itself. The completed packet should describe one assessment at one time, with explicit limits.

A research gate can fail because information is missing, because the claim is poorly formed, or because the question exceeds the available method. These failures require different responses. Missing information calls for a named source or a waiting condition. A vague claim calls for rewriting. A question that depends on inaccessible contract details may require accepting that the public record cannot support the intended conclusion. More pages of loosely related reading do not repair those distinctions.

Set a stopping instruction before an investigation becomes open ended. For example, a hypothetical researcher may inspect the relevant filing, its referenced supplement, and the company's metric definitions, then stop if the necessary cohort breakdown remains unavailable. This is a scoped search plan, not a universal document limit. If a new primary document later supplies the breakdown, the gate can reopen. Until then, record the unanswered question rather than repeatedly restating the same favorable narrative.

The final packet can say: reported growth reconciled under disclosed categories; attribution to existing customer demand unresolved; further decision assessment would depend on that assumption. That is a substantive result even though the original idea remains incomplete. Preserve the unsuccessful research path briefly so another reader does not repeat it unknowingly. A checklist earns its place by identifying where a conclusion must stop and what specific information would allow it to move further.

Passing an evidence gate does not make an investment safe, suitable, or attractively priced. A coherent explanation can still be wrong, and a well-understood business can still present an unsuitable investment. The checklist also cannot eliminate uncertainty about future behavior. Its purpose is to identify the assumptions you are carrying into a decision, not turn those assumptions into facts.

Avoid grading the research solely by the subsequent share price. An unresolved acquisition adjustment remains a research gap even if the price rises afterward. A careful investigation remains careful if an unpredictable event later changes the outcome. Judge this framework by whether a second reader can understand the claim, inspect its support, and see why the stated next step follows. That is a reviewable research result, not a promise of financial performance.