How to Evaluate Your Thesis Assumption Hierarchy

What an assumption hierarchy is

Every investment thesis rests on claims about the future that are not yet proven. The company will win the contract. The regulatory approval will arrive on schedule. The management team will execute on the cost reduction program. The addressable market is large enough to support a second-place competitor. These claims are assumptions: things the investor believes are true and that the thesis depends on being true.

An assumption hierarchy is the complete set of these claims, ranked by the damage each would cause to the expected investment outcome if it turned out to be wrong. Not all assumptions are equally important. The failure of the most critical assumption may collapse the thesis entirely. The failure of a secondary assumption may reduce the expected return without eliminating the investment case. The hierarchy makes this difference explicit and actionable.

Without a hierarchy, every assumption receives roughly equal attention during monitoring. The investor scans earnings calls and news for any item that touches the thesis, without a systematic way to assess how much any particular piece of information actually matters. The result is either over-reaction to minor developments or under-reaction to major ones, depending on which way the noise is running in any given week.

With a hierarchy, the investor knows exactly which assumptions to watch most closely, what evidence would challenge each one, and what action is required if a load-bearing assumption is violated. The hierarchy converts monitoring from an open-ended attention exercise into a structured checklist.

How to build an assumption hierarchy

Building the hierarchy is a three-step process. The first step is enumeration. The second is damage assessment. The third is ranking.

Step 1: Enumerate every assumption

Write down every claim the thesis requires to be true. Do not pre-filter. The goal at this stage is completeness. Include assumptions about the company (management execution, product development, competitive position), about the market (demand growth, pricing dynamics, market share trajectory), about the external environment (regulatory outcomes, macro conditions, interest rate impact on the business model), and about the investment mechanics (how long the valuation gap will take to close, what catalyst will drive repricing).

Most theses, when this exercise is done honestly, reveal between five and ten distinct assumptions. If fewer than three emerge, the thesis is likely describing the investment in terms too general to function as a monitoring tool. If more than ten emerge, the assumptions may be listed at a level of granularity that is too detailed to track, or the thesis is carrying too many concurrent dependencies.

Step 2: Assess damage for each assumption

For each assumption, ask a single question: if this assumption turns out to be wrong, what happens to the thesis? The answer has three possible forms:

• The thesis fails completely. The expected outcome cannot occur if this assumption is wrong. Exit is required.
• The thesis weakens significantly. The expected outcome still occurs, but at a smaller magnitude or with a longer time frame. Position size reduction may be warranted.
• The impact is marginal. The thesis is directionally intact even if this assumption is wrong. No immediate action is needed, but the assumption should be noted.

Step 3: Rank by damage

Rank assumptions from highest damage to lowest. The top two or three assumptions are the load-bearing ones. The remainder are supporting. The hierarchy is now the monitoring framework. Subsequent information flow gets routed to whichever assumption it speaks to, with the response determined by where that assumption sits in the hierarchy.

Load-bearing vs. supporting assumptions

The distinction between load-bearing and supporting assumptions is not about probability. An assumption can be very likely to be correct and still be load-bearing, because the question is what happens to the thesis if it is wrong, not how likely it is to be wrong.

A load-bearing assumption is one whose failure causes the thesis to fail completely. The expected outcome disappears. The rationale for holding the position disappears with it. When a load-bearing assumption is violated by incoming evidence, the correct response is exit, regardless of current price, sentiment, or recent market performance. The thesis has been invalidated, and holding through an invalidated thesis is not discipline; it is hope.

A supporting assumption is one whose failure weakens the thesis but leaves it directionally intact. The expected outcome still occurs, but perhaps with lower magnitude or over a longer time frame. When a supporting assumption is violated, the correct response is reassessment: does the thesis still reach the return threshold at a smaller magnitude? Is a reduced position size appropriate? Does the time horizon need to be extended? These are decisions, not automatic exits.

The practical value of the load-bearing vs. supporting distinction is that it creates clear pre-committed rules for how to respond to bad news. Without the distinction, every piece of negative news triggers the same ambiguous question: is this a reason to sell or a reason to hold? With the distinction, the investor can ask instead: does this news violate a load-bearing assumption? If yes, exit. If no, reassess the magnitude but stay in the position.

Testing each assumption at entry

Before opening a position, each load-bearing assumption should be tested against available evidence. The goal is not to prove the assumption is correct. It is to understand how much observable evidence already exists to support or challenge it, and to identify what signals over the next quarter will confirm it is still intact.

For each load-bearing assumption, answer three questions at entry:

1. What observable evidence already exists that supports this assumption? If the answer is "none," the assumption is entirely speculative. That does not necessarily mean it is wrong, but it should influence position sizing. An investment built on multiple unobserved load-bearing assumptions carries more uncertainty than one built on assumptions with existing evidence behind them.
2. What would tell you in the next quarter if this assumption is still intact? This defines the monitoring signal. It should be specific: a metric, an event, a management statement, a competitive development. "The company keeps growing" is not a monitoring signal. "Sequential revenue growth accelerates in Q3 as the new distribution channel becomes productive" is a monitoring signal.
3. What would tell you this assumption has been violated? This defines the break condition. It should be as specific as the monitoring signal. The break condition is the pre-committed trigger for exit or reassessment. Writing it before the position is open removes the bias that comes from writing it while the position is running at a loss.

Answering these three questions for each load-bearing assumption before entry takes time. That time is the due diligence the assumption hierarchy makes concrete. A thesis with three load-bearing assumptions where none of these questions can be answered has not been adequately researched for the position size being considered.

Using the hierarchy for ongoing monitoring

Once the hierarchy is built and entry testing is complete, the hierarchy becomes the filter through which all subsequent information is processed. This is the payoff for the upfront work: monitoring becomes focused rather than open-ended.

The practical discipline is to route every significant piece of incoming information to the assumption it speaks to, then assess that assumption's status. An earnings release is not just an earnings release. It is data that speaks to specific assumptions: the revenue assumption, the margin assumption, the cash flow conversion assumption. Route the data to those assumptions first. Does the result confirm the assumption is still intact, weaken it, or violate it? That assessment determines the response before price movement determines the emotion.

The hierarchy also protects against the opposite error: over-weighting information that is relevant to a supporting assumption and treating it as though it speaks to a load-bearing one. A competitor launching a tangentially related product may be relevant to a supporting competitive-position assumption without touching the primary load-bearing assumptions at all. Without the hierarchy, that news might trigger an anxiety-driven exit from a position that was actually intact. With the hierarchy, the investor can locate the news within the assumption structure and assess it proportionally.

Review the hierarchy formally at each scheduled thesis review date. Confirm each assumption's status: confirmed, uncertain, weakened, or violated. A load-bearing assumption that has moved from confirmed to uncertain is a signal to investigate, not necessarily to exit. A load-bearing assumption that has moved to violated is an exit trigger, regardless of the price at the time of review.

Frequently asked questions

References

• Investment Thesis Lab
• Swoopr Research Workbench
• Swoopr Research Methodology