Every investment thesis rests on a set of conditions that must remain true for the expected outcome to occur. Most investors know this at some level, but the typical response is to produce a list: a set of things that could go wrong, organized in no particular order. A list is not an assumption hierarchy. The distinction is not cosmetic. A list treats all assumptions as roughly equivalent and provides no basis for prioritizing monitoring attention, setting break conditions, or deciding what constitutes a material change. A hierarchy forces an explicit ranking, which forces an explicit answer to the question: which assumption matters most?
What is an assumption hierarchy?
An assumption hierarchy is the complete, ranked set of conditions that must remain true for a thesis to produce its expected outcome, ordered from most consequential to least. It is not a list of risks or a catalogue of things that could go wrong. It is a map of the analytical bets that underlie the thesis, organized by how much each bet matters to the conclusion if it fails to hold.
The distinction from a risk list is important. A risk list is externally oriented: it describes things that might happen in the world that could hurt the investment. An assumption hierarchy is analytically oriented: it describes the conditions the thesis itself asserts must remain true for the conclusion to follow from the evidence. Some of the items in a risk list will correspond to assumptions in the hierarchy, but not all of them will, and the hierarchy will also contain assumptions that do not appear in a risk list at all because they are conditions that the thesis treats as given.
For example, a thesis on a SaaS business might assume that enterprise software gross margins will remain above 70%. This is not typically listed as a risk. No one says "the risk is that gross margins might fall." But it is a load-bearing assumption of the valuation thesis: the entire terminal value calculation depends on margin structure. If gross margins fell to 50%, the business might still be growing and the management might still be excellent, but the thesis would be wrong. The assumption hierarchy captures this bet explicitly, where a risk list might not.
The hierarchy is also different from a checklist of due diligence items. Due diligence items are things the investor needs to verify before entry. Assumptions in the hierarchy are conditions the investor expects to remain true during the hold, which need to be monitored on an ongoing basis. Checking a due diligence item is a one-time action. Monitoring an assumption is a recurring responsibility.
Why ranking matters more than listing
Any thoughtful investor can generate a list of assumptions underlying a thesis. The discipline of ranking them forces a different and more demanding question: if I had to monitor only one assumption because of limited time and information, which one would I choose? The answer to that question identifies the load-bearing assumption, which deserves the most monitoring attention, the most carefully designed break condition, and the largest weight in any position-sizing decision that relies on conviction in the thesis.
Ranking also reveals hidden complexity. When an investor lists assumptions in no particular order and is then asked to rank them, the exercise often reveals that two assumptions the investor considered independent are actually related. If assumption A is true, assumption B is more likely to be true. If assumption A fails, assumption B becomes uncertain. This conditional relationship was invisible in the list but becomes visible in the ranking process.
Ranking also reveals over-confidence. An investor who lists five assumptions and believes all five are highly probable is implicitly claiming that the probability of the thesis working is approximately the product of those five probabilities. Even at 80% confidence in each assumption, five independent assumptions produce a combined confidence of approximately 33%. Most investors who enumerate five assumptions do not consciously believe they have a one-in-three thesis. The ranking process, by forcing the investor to evaluate each assumption individually and then consider them together, tends to produce more calibrated overall confidence estimates.
Finally, ranking provides a monitoring priority ordering. When information arrives that is relevant to multiple assumptions, the investor should route it to the highest-ranked assumption first. When limited time forces a choice between investigating developments that are relevant to different assumptions, the hierarchy says which investigation is more important. Without ranking, all assumptions compete for equal attention, which means the most important assumptions may receive no more attention than the least important.
How to build an assumption hierarchy
Start by listing every condition that must remain true for the thesis to produce its expected outcome. Do not filter at this stage. The goal is completeness, not concision. Consider conditions related to the business model, the competitive environment, the macroeconomic context, management's execution capacity, the regulatory environment, and the financial structure of the company.
Once the list is complete, apply a single sorting question to each item: if this assumption fails while all other assumptions remain intact, does the thesis fail completely, weaken significantly, or survive mostly unchanged? Sort the results into three groups based on the answer. The first group (thesis fails completely) contains the load-bearing assumptions. The second group (thesis weakens significantly) contains the supporting assumptions. The third group (thesis survives) contains the incidental assumptions.
Review the first group. There should be no more than three items in it. If there are more, either some of the items are not truly load-bearing (they weaken the thesis rather than destroying it) or the thesis is genuinely too complex and should be simplified before capital is committed. A thesis that can only work if ten independent conditions remain true is not a thesis with a clear market-gap argument. It is a series of hopes dressed as analysis.
For each item in the first group, write the break condition: the specific observable evidence that would tell you the assumption has failed. The break condition must be specific enough to be checked. "The business deteriorates" is not a break condition. "Enterprise segment gross margin falls below 65% in two consecutive quarters" is a break condition. The specificity forces the investor to think carefully about what the assumption actually predicts and what would falsify it.
For items in the second group, write a monitoring description: what evidence would confirm the assumption is holding and what evidence would challenge it. A full break condition is not required, but a monitoring description ensures that evidence relevant to supporting assumptions is not ignored during reviews.
Items in the third group can be listed for completeness but require no monitoring protocol. They are the background conditions of the thesis that are unlikely to change and that would not materially affect the thesis if they did change.
Load-bearing assumptions: the core of the hierarchy
A load-bearing assumption is one where failure means the thesis conclusion must be abandoned. The investor should exit the position if this assumption fails, regardless of price, regardless of recency of entry, and regardless of how the rest of the thesis is performing. It is the analytical equivalent of a foundational structural element: if it fails, the structure fails.
Examples of load-bearing assumptions help clarify the concept. In a SaaS valuation thesis, the assumption that software gross margins will remain above 70% is typically load-bearing because the entire terminal value calculation depends on the margin structure. In a biotech catalyst thesis, the assumption that a specific drug candidate will receive regulatory approval is load-bearing because the entire thesis is built around that approval. In a management turnaround thesis, the assumption that the new management team will successfully execute the operational changes required is load-bearing because the expected value creation depends entirely on that execution.
The limit of three load-bearing assumptions is not arbitrary. It reflects the practical limit of analytical clarity in a single investment decision. A thesis that genuinely requires more than three foundational conditions to remain true is asking the investor to hold simultaneously high confidence in a large number of independent uncertain conditions. The expected value of the thesis may still be positive, but the range of outcomes is wide enough that position sizing should reflect that uncertainty, and the thesis complexity is high enough that monitoring becomes difficult to do well.
Load-bearing assumptions should be stated as falsifiable claims, not as expressions of optimism. "Management will continue to execute at a high level" is not a falsifiable claim. It is a hope. "Enterprise revenue growth will remain above 15% annually for the next three years, as evidenced by quarterly enterprise segment bookings" is a falsifiable claim with a specified measurement. The distinction is important because break conditions can only be designed for falsifiable claims.
Supporting assumptions: the rest of the hierarchy
Supporting assumptions are the conditions whose failure would reduce the expected outcome but not eliminate it. If a supporting assumption failed, the investor might reduce position size to reflect the reduced expected value but would not necessarily exit. The thesis is weakened, not destroyed.
Examples of supporting assumptions in a compounding quality thesis might include the assumption that the company will return excess capital to shareholders rather than making value-destructive acquisitions, that the management team will remain stable, or that the company will expand into adjacent markets at acceptable economics. Each of these assumptions, if it failed, would reduce the expected return but would not change the foundational thesis that this is a high-quality compounder generating superior returns on capital in its core business.
Supporting assumptions still need to be monitored. A supporting assumption that fails often signals a developing problem that will eventually reach a load-bearing assumption. If the capital allocation assumption fails first (the company makes a large acquisition at a poor price), this might be an early signal that the management judgment assumption, which is a load-bearing assumption, is also at risk. Catching the supporting assumption failure early provides earlier warning than waiting for the load-bearing assumption to break.
Supporting assumptions also affect the appropriate position size. A thesis where all load-bearing assumptions are intact but two supporting assumptions have failed is a different position than a thesis where all assumptions are intact. The investor's monitoring notes should track the status of supporting assumptions explicitly so that position-sizing decisions reflect the actual current state of the thesis, not just whether the load-bearing assumptions are still intact.
How to test each assumption at entry
Before entering a position, each load-bearing assumption should be tested with a two-part check. First: what observable evidence currently exists to support or challenge this assumption? If the answer is "there is no current evidence either way, the assumption is just what the business needs to be true," the investor should pause. An assumption with no supporting evidence is an unsupported bet, not an analytical judgment. The thesis should either be built on evidence that exists today or acknowledge explicitly that this is a high-uncertainty assumption that requires a smaller position size.
Second: what data or event in the next quarter would tell me whether this assumption is still intact? If the investor cannot answer this question, the assumption is not specific enough to be monitored. "The business continues to grow" has no quarterly monitor. "Enterprise segment revenue growth exceeds 15% in Q3 as reported in the quarterly earnings release" has a clear monitor. Rewriting each assumption until it has a clear quarterly monitor is one of the most valuable exercises in assumption hierarchy construction.
This pre-entry test often reveals a distinction that is not visible when assumptions are listed in the abstract: some assumptions are well-established facts about the business that are unlikely to change and that are well-supported by recent historical evidence. Others are highly uncertain bets about future behavior. The hierarchy should treat these differently. Well-established facts with strong supporting evidence deserve less monitoring attention and less weight in position-sizing uncertainty. Highly uncertain bets deserve more monitoring attention and should lead to a smaller initial position with a plan to add as the assumption is confirmed.
The pre-entry test also provides the baseline against which post-entry evidence will be evaluated. If the assumption at entry was that enterprise gross margins were 73% and trending stable, and three quarters later the margins are at 68% and still declining, the investor can evaluate this against the specific claim made at entry rather than against a vague sense of whether the business is still doing well.
Using the hierarchy for ongoing monitoring
The assumption hierarchy provides the organizing structure for every subsequent review of the investment. When new information arrives, the first question should not be "is this good news or bad news for the stock?" It should be "which assumption does this information speak to?"
Routing evidence to specific assumptions, rather than evaluating it as a general sentiment update, changes the quality of the monitoring process substantially. When a quarterly earnings report arrives, the investor with a hierarchy reads it with specific questions: does the enterprise segment growth figure support or challenge the revenue growth assumption? Does the gross margin figure support or challenge the profitability assumption? Does the management commentary on competitive dynamics support or challenge the competitive position assumption? Each piece of evidence is evaluated in the context of the specific assumption it bears on.
Without a hierarchy, the same earnings report tends to be evaluated as a gestalt: strong beat on revenue, slight miss on margins, reaffirmed guidance, net positive. This evaluation does not connect the specific data points to the specific assumptions the thesis depends on. An investor could read a "net positive" earnings report and miss a concerning trend in the load-bearing gross margin assumption that is visible only when the margin figure is evaluated against the specific threshold the thesis requires.
At each review date, the investor should explicitly check the status of each load-bearing assumption: is it still intact, is it trending toward the break condition, or has the break condition been triggered? This check should be recorded in the thesis change log. The record creates accountability: the investor must state explicitly whether the load-bearing assumptions are intact, rather than simply noting that the thesis continues to look interesting.
When a load-bearing assumption is challenged
Evidence arrives that casts doubt on the top-ranked assumption. This is the scenario the hierarchy is most directly designed to handle, and it is also the scenario where investors without a hierarchy tend to perform worst.
Without a hierarchy, the typical response to a challenged core assumption is one of two failure modes. The first is immediate reaction: the evidence feels threatening, the investor exits quickly, and the decision is driven by anxiety rather than by analytical evaluation of how material the evidence is. The second is rationalization: the evidence is acknowledged but minimized ("one quarter is not a trend," "management has explained this," "the market is overreacting"), and the investor stays in a position where the core analytical bet has been materially challenged.
With a hierarchy and a pre-written break condition, the response follows a four-step process. First, verify the evidence: is this a measurement anomaly or a reporting artifact, or does it reflect a real change in the underlying condition the assumption describes? Second, check whether the break condition has been triggered: has the specific observable event that defines assumption failure occurred, or is the evidence moving toward the break condition without reaching it yet? Third, if the break condition has been triggered, follow the pre-written exit protocol without requiring a new decision. Fourth, if the break condition has not yet been triggered but the evidence is trending toward it, reduce position size to reflect the reduced confidence in the load-bearing assumption and increase the monitoring frequency.
The pre-written break condition and the four-step protocol convert what is otherwise a high-emotion, high-uncertainty decision into a structured process. The investor is not deciding in real time whether this new piece of evidence is material. The investor is checking whether a pre-specified condition has been met. That distinction produces more consistent behavior and reduces the influence of recency bias, loss aversion, and the sunk-cost fallacy on the exit decision.
Frequently asked questions
What is an assumption hierarchy in an investment thesis?
An assumption hierarchy is the complete, ranked set of conditions that must remain true for an investment thesis to produce its expected outcome, ordered from most consequential to least. It is not a list of risks. It is a map of the analytical bets underlying the thesis, organized by how much each assumption matters to the conclusion if it fails.
How many load-bearing assumptions should a thesis have?
A well-constructed thesis should have no more than three load-bearing assumptions: conditions whose failure would require the thesis to be abandoned entirely. If a thesis appears to have more than three load-bearing assumptions, it is either too complex or has not been simplified enough. The exercise of reducing the load-bearing assumption count forces the investor to identify which analytical bets are truly foundational and which are secondary.
What is the difference between a load-bearing and a supporting assumption?
A load-bearing assumption is one where failure means the thesis conclusion must be abandoned. If this assumption proves false, the investor should exit the position. A supporting assumption is one where failure would reduce the expected outcome but not eliminate it. If a supporting assumption fails, the investor might reduce position size but would not necessarily exit. Both types should be monitored, but only load-bearing assumptions require formal break conditions.
How do I test an assumption before entering a position?
For each load-bearing assumption, identify: what observable evidence currently exists to support or challenge this assumption, and what data or event in the next quarter would tell you whether this assumption is still intact. This pre-entry test often reveals which assumptions are well-established facts and which are highly uncertain bets, and that distinction should influence how large the initial position is.
What should I do when a load-bearing assumption is challenged by new evidence?
When new evidence challenges a load-bearing assumption, follow a four-step process: verify the evidence to determine whether it is a measurement anomaly or a real change; check whether the pre-written break condition for this assumption has been triggered; if triggered, follow the pre-written exit protocol; if not yet triggered but trending toward it, reduce position size and increase monitoring frequency. The assumption hierarchy provides the framework for an ordered response rather than a panic or an indefinite hold.