Capital Budgeting for a Five-Warehouse Expansion.
A 10-year incremental cash-flow model that tested whether the project still created value after operating assumptions, downside scenarios, sensitivity, and financing capacity were made explicit.
My role
I served as the primary analytical lead, completing most of the underlying research and financial modeling while teammates reviewed and double-checked outputs.
How the project economics were built
The final model separated merchandise and membership revenue, used a 60%/85%/100% sales ramp, modeled cannibalization on lost contribution margin, and included operating costs, CapEx/depreciation, maintenance CapEx, operating NWC recovery, and after-tax salvage.
| Driver | Final treatment |
|---|---|
| Revenue | Mature merchandise sales modeled separately from membership revenue; warehouse ramp applied over Years 1–3. |
| Cannibalization | Applied to lost contribution margin rather than gross sales. |
| Reinvestment | Initial CapEx, maintenance CapEx, straight-line depreciation, NWC build/recovery, and after-tax salvage. |
| Discount rate | 8.18% WACC using a forward-looking 21% statutory tax rate for the debt tax shield. |
The base case created value—but the downside was thin
Pessimistic/base/optimistic NPVs were approximately $3.99M / $140.96M / $326.13M.
Stress the capital-budgeting decision
Switch between the final combined scenarios, then inspect the three largest one-way NPV drivers. Sensitivity points are the tested values from the authoritative August 19 model.
The browser does not interpolate untested assumptions. It shows the low / base / high points actually evaluated in the final model.
Sensitivity turned the model into a control map
The most useful part of the project was identifying what actually drove the result. Mature warehouse revenue was the largest one-way NPV driver, followed by variable warehouse operating cost and the initial CapEx multiplier. Tested two-way combinations remained positive.
Mature sales mattered more than the headline NPV
The final memo’s one-way sensitivity put mature warehouse revenue first: moving mature sales from 80% to 120% of base shifted NPV from roughly $94.27M to $187.66M. That reframed the recommendation: site selection, demand validation, and ramp execution were the operating controls that mattered most.
Financing: avoid solving a liquidity problem that did not exist
I supported a 100% internal-financing recommendation by comparing the modeled project with Costco’s FY2025 liquidity and cash generation: $15.284B cash + short-term investments, $13.335B operating cash flow, and about $7.837B simple free cash flow after CapEx.
The $128.91M project represented about 0.84% of cash/ST investments and 1.64% of annual simple FCF in that academic analysis.
Model corrections mattered as much as model outputs
Before the authoritative final version, I documented and corrected membership-revenue overlap and tax-rate treatment. Those changes were not cosmetic: double-counting revenue or using the wrong tax treatment can create a clean-looking NPV that answers the wrong question.