A valuation can look robust across familiar optimistic and pessimistic cases while remaining fragile near the decision that matters. Reverse stress testing changes the direction of inquiry. First define the valuation threshold at which an investment, financing plan or strategic choice would no longer be acceptable. Then work backwards through the model to find the smallest coherent changes in assumptions that cross it. The result is not a prediction. It is a visible decision boundary. Valuator, Viotus’s desktop valuation program, publicly centres structured assumptions, scenarios, ranges and a traceable relationship between beliefs and calculated outcomes.
Define failure before changing inputs
Write the decision and an unacceptable valuation threshold in the same unit, currency, date and ownership basis as the base case. A useful statement is operational: “Below 70 million euros of equity value, this decision fails our stated condition.” Avoid vague labels such as bad case because they provide no boundary to solve for.
Record why the threshold matters without pretending the number is universal. It may reflect a price, covenant, funding need or internal decision rule. Reverse stress testing does not determine whether that rule is wise; it shows which assumptions would bring the model to it. Keep the base value, threshold and gap visible together.
Map assumptions to the threshold
List the assumptions that can materially move value: operating outcomes, timing, margins, reinvestment, risk or other inputs actually present in the chosen model. Change one assumption at a time just enough to reach the threshold, then record its distance from the base case. This first pass exposes sensitivity but not yet a realistic failure story.
Use an assumption-distance table: assumption, base value, threshold-reaching value, absolute change, relative change and reason the change could cohere with other inputs. Do not compare unlike percentages mechanically. A two-point margin change and a two-point rate change have different meanings. The table organizes evidence; it does not turn every variable into the same kind of risk.
Find the nearest coherent combination
Single-variable breaks may require implausible movement. Test small combinations next. For example, a modest revenue reduction paired with a modest margin reduction may cross the threshold sooner than either change alone. Preserve relationships: lower growth may alter reinvestment needs, and delayed outcomes may change the relevant period.
Search for the smallest coherent assumption distance, not the most dramatic catastrophe. Document each step and stop once the threshold is crossed. The nearest combination is often more useful than an extreme scenario because it identifies the amount of adverse movement the decision can absorb before its stated condition fails.
Work a small example
Suppose a base case gives 100 million euros and the unacceptable boundary is 70 million. Trial A requires revenue 25 percent below base while other assumptions remain fixed. Trial B reaches 70 with revenue 10 percent lower and margin three points lower. Trial C reaches it after a one-year delay plus a smaller margin change.
Record these as alternative paths, not probabilities. The table might show distances of 25 revenue points; 10 revenue points plus three margin points; or one year plus two margin points. The model alone cannot say which path is credible. The review question is which path best matches available evidence and whether its distance is tolerable.
Place the boundary beside the base case
Report the base valuation, unacceptable threshold, nearest coherent path, changed assumptions and unresolved evidence together. Never present the breaking point without its definition or imply investment advice. Valuator is analytical software and offers no investment advice or guarantees.
Valuator’s public workflow—frame the question, organize assumptions, compare cases and review ranges—provides the components for a reviewable reverse test. Explore its public product page for current context. The portable method is threshold first, assumptions second, smallest coherent path third, and a recorded decision boundary beside the base case.