Reverse / Goal Seek: solve for a result

Methods and settings

Define a solvable question

Select the output with its scenario, period and operator. Choose the target mode and value or range, then select supported variables in the workspace. Enter finite lower and upper bounds for each variable and add result constraints where needed. Solve first, inspect the status and changes, then use Preview before deciding whether to Apply. Solving does not write financial inputs.

A supported operating question

For a forecast EBITDA target, select an eligible owning price or operating assumption and ask for a stated EBITDA within defensible input limits. Downstream calculations follow the model’s actual relationships. If the chosen method, source ownership or target period makes that assumption ineligible, select the supported source; do not assume every displayed subtotal can be directly solved as an input.

Search settings and interpretation

The current defaults are 600 evaluations, 200 iterations and five starts. Settings permit 10–10,000 evaluations, 1–2,000 iterations and 1–16 starts. More search work can help but does not prove that a global optimum has been found. For an explicit target, the search can prefer smaller weighted changes; for maximize/minimize, it reports the best feasible result it found.

Goal tolerance and improvement tolerance

Advanced shows the goal tolerance in the target’s display unit; for percentage targets the input is in percentage points. An equality goal is satisfied when the absolute difference from the requested value is within tolerance. For ≥, ≤ and range goals, tolerance extends the permitted boundary; the default is zero. Equality uses a small positive default based on the target scale. Review the shown value rather than treating a rounded display as an exact match.

For Maximize or Minimize, this field becomes Improvement tolerance and governs when objective differences are meaningful. It is not a permission to violate secondary constraints. Variable bounds and each constraint remain separate checks.

A candidate is not an applied decision

Review goal achievement, constraints, variable bounds and financial plausibility separately. A closest feasible candidate may satisfy constraints while missing the requested goal. No feasible candidate found is not a mathematical proof that none exists. Preview is temporary; Apply revalidates eligible sources. See statuses and unavailable inputs and Preview and Apply.

Public models do not store this analysis session

Reverse / Goal Seek solver setup, state, previews and candidates are not persisted in public model snapshots. Publication preserves operator inputs and deterministic operator outputs. After Fork, run this analysis again locally against the restored editable inputs. A reviewed change explicitly applied to an underlying operator input can be published as that input; the analysis session that produced it is still excluded.