Turn a financial decision into a bounded Solver problem

Specify the economics and objective

The native regression fixture has an unchanged EUR 120 annual revenue forecast and EUR 48 EBITDA, with the 2026 actual period and 2027–2029 forecast. The LBO starts at 9× entry, 8× exit and a 4× debt tranche with 5% cash interest. Monetary inputs use EUR in the underlying fixture; multiples and IRR retain their own units.

The objective is maximize IRR, not reach 25%. Entry Multiple may range from 5× to 9×; debt leverage from 0× to 5×. The separate output constraint is native LBO entry gross debt / entry metric ≤4×. It is not the Credit operator’s Net Debt / EBITDA metric.

Inspect the candidate and its residuals

The native run returned status “solved” after 290 evaluations within a 300-evaluation budget (200 iterations, 3 starts). Entry reaches the lower bound of 5×. Debt leverage is 4.000000009996×: the tiny excess over 4× is within the explicit 0.00000001 tolerance. The constraint is binding.

The IRR increases from about 7.602705% to 42.345739%. This candidate improves the specified objective within the tested bounds. It is not a certificate of global optimality, an achievable negotiated entry price or a financing recommendation. An optimization has no 25% target-gap claim to report.

Preview before an explicit Apply

Solving itself leaves the canonical workspace untouched. Preview recalculates the full native candidate and must agree with the Solver output. Here Preview reproduces the candidate IRR. Only supported, still-valid writable changes can be applied.

1. Import the example, select Deals → LBO IRR and open Analyze → Reverse.
2. Check Maximize, both variables, bounds, the leverage constraint and its tolerance.
3. Run, review status and Base versus candidate, then open Preview.
4. Inspect the LBO cash flows and changed inputs. Apply only after review.
5. One Apply is one reversible batch. Native tests verify Undo restores the exact workspace and Redo reapplies the candidate.

A best attempt is not a feasible solution

A separate variant adds gross leverage ≥4.5× while retaining ≤4×. The run ends with status “budget” at 80 evaluations and no solution. Do not promote an objective-improving best attempt that fails the constraints.

These two inequalities are directly contradictory, but a search status in a general model is not a mathematical proof of infeasibility. Review bounds, constraints, unavailable outputs and the evaluation budget. The complete synthetic failure case is available separately.

Download the synthetic example, explicitly import it under Model, select Advanced and open Deals → LBO → Analyze → Reverse. Import creates a separate model; it does not overwrite an existing model.

Related decisions and model reference

For adjacent questions: Financial Goal Seek: Solve a Target, Review and Apply; Goal Seek vs Solver: When One Changing Input Is Not Enough; LBO Model & Sponsor Returns; Financial Model Analysis — Drivers, Margins and Cash.

For the exact controls and source behavior, use Reverse / Goal Seek: solve for a result and Reverse with multiple variables.