Five engagements. Five different problems.

A sponsor under LOI with 21 days to firm up equity. The broker pro forma assumed a $250 premium on every renovated unit and 3% expense growth against a tax bill guaranteed to be reassessed at sale price.

Rebuilt the deal from the T-12 up: reassessed taxes at the county's actual millage, staged renovations across 30 months, and built a unit-by-unit ROI matrix that flagged 22 units where the premium didn't clear the renovation cost hurdle.

The LP committed on the revised numbers and the deal closed in 54 days. Trimming the 22 non-performing renovations cut roughly $410k of capex without moving stabilized NOI more than 1%.

A developer with entitled land, a GC budget, and two construction term sheets that looked identical on rate but weren't — different recourse burn-offs, different interest reserve mechanics.

Built the full monthly development model: S-curve draws, capitalized interest, construction-to-perm takeout, and a lease-up ramp tied to a micro-study of retail rents within a half-mile. Ran both term sheets through the same draw schedule to expose the true cost gap.

The 'cheaper' loan proved ~$180k more expensive once the reserve mechanics ran through the draws. The lender package built from the model sized at 62% LTC, and the equity waterfall closed at an 8% pref over a 20% promote.

A family office weighing a stabilized NNN industrial purchase against simply buying bonds — and needing to understand what happens in year nine when the lease expires.

Modeled the credit-tenant cash flows with three rollover scenarios: renewal at market, 12-month re-tenanting with TI/LC load, and a dark-value floor based on replacement cost and local vacancy. Layered a refinance analysis at year five against the hold.

The dark-value scenario reframed the negotiation — the buyer secured a larger reserve holdback and closed with conviction about the year-nine exposure instead of hoping about it. Refi-and-hold beat sale in every scenario above a 7.4% exit cap.

An operator buying a tired independent hotel with a plan to renovate and reposition — and an equity partner skeptical that the post-renovation ADR was anything more than hope.

Built the hotel model on a full P&L basis: occupancy/ADR build-up by segment, departmental costs on POR/PAR logic, a 4% FF&E reserve, and a competitive-set penetration analysis showing where the RevPAR index gains actually come from. Ran branded vs. independent scenarios including franchise fee load.

The penetration math held: the equity partner funded, and the independent path beat the soft-brand conversion by roughly 90 bps of IRR after fees. The PIP budget got a 12% contingency the GC initially resisted — and used.

A sponsor mid-raise on a recapitalization, using an inherited model. An anchor LP's advisor asked one question about the promote math that the sponsor couldn't answer. The raise stalled.

Full forensic audit: precedent mapping of every output cell, hard-code hunt, and re-derivation of the waterfall by hand. Found a double-counted tax reassessment, a broken hurdle reference paying the GP promote 190 bps early, and a circular refinance assumption quietly inflating year-3 cash flow.

Corrected LP IRR moved from 15.9% to 13.6% — before the money came in, not after. The sponsor restructured the pref, re-presented with a clean model and an audit memo attached, and closed the raise without re-trading a single committed investor.