The five assumptions that sink a hospitality feasibility study
Ramp-up, ADR benchmarking, operating cost inflation, FF&E reserve and exit yield — where sponsor models and lender models part company on a hotel scheme, and what evidence closes the gap.
A hospitality feasibility study rarely fails because the arithmetic is wrong. It fails because five assumptions were set by optimism rather than evidence, and every one of them compounds. A lender reading the study knows exactly where to look. It is worth knowing before they do.
The pattern is consistent: the sponsor’s model is not dishonest, it is simply built from the best year the market has had rather than an ordinary one.
Ramp-up
Almost every model we are asked to test assumes the asset stabilises faster than comparable properties in the same market actually did. Stabilisation is not a design assumption; it is an observed outcome, and it varies with segment, location, brand strength and how much new supply opens in the same window.
The correction is not to pick a more conservative number out of the air. It is to find the properties that opened most recently in the same sub-market and segment, establish what their first three years looked like, and explain in writing why this asset should differ. If the answer is that this one is better, the study needs to say why in terms a lender can test — a signed operator agreement, a contracted demand source, a genuine gap in supply.
ADR benchmarking
Average daily rate is where sponsor models drift furthest, usually through a comparable set that is chosen rather than derived. The set is often the assets the sponsor admires, not the assets the property will realistically compete with on a Tuesday in low season.
- Build the set from properties a guest would actually substitute — same segment, same catchment, same booking channels.
- Separate rate from occupancy. A model that lifts both simultaneously, year after year, is asserting pricing power the market has not given anyone.
- Check the seasonality profile, not just the annual average. Two assets with identical annual ADR can have very different cash positions in the months when debt service falls due.
- State the rate in real terms and inflate it explicitly, so the reader can see what is volume and what is inflation.
Operating cost inflation
Revenue lines are usually inflated with some care. Cost lines are frequently held flat, or grown at a single blended rate applied to everything. Payroll, utilities, food cost and insurance do not move together, and in hospitality payroll is the largest and least flexible of them.
A study that survives review inflates the major cost categories separately, states the source for each, and shows what happens to the debt service cover ratio if payroll runs ahead of the assumption by a few points. If that single sensitivity breaks the covenant, the capital structure is wrong — and it is far cheaper to learn that before the facility is signed.
FF&E reserve
The furniture, fixtures and equipment reserve is the most commonly under-provisioned line in the whole model, because it costs the sponsor nothing in year one and everything in year seven. Operator agreements typically specify a reserve as a percentage of revenue, escalating over the early years. Models often apply the first-year figure permanently, or omit the reserve from the cash available for debt service.
Two questions settle it. Does the reserve in the model match the reserve in the management agreement, year by year? And is the reserve treated as a genuine cash outflow before the debt service calculation, or quietly added back? A lender will find the answer in the first read.
Exit yield
Exit assumptions carry more of the project return than most sponsors acknowledge. A model that exits at a yield tighter than the entry yield is assuming the market re-rates in the sponsor’s favour, and that assumption often carries the entire equity return on its own.
The disciplined approach is to exit at or slightly wider than entry, then show separately what a re-rating would add. That way the return is defensible without it, and the upside is visible rather than buried. If the project only works on a tighter exit, that is not a study finding — it is a warning.
What a bankable study looks like
The difference between a study that supports a construction facility and one that quietly fails is rarely sophistication. It is traceability. Every one of the five assumptions above should be traceable to a source with a date on it, and every one should have a downside case attached showing what happens to cover ratios if it is wrong by a realistic margin.
We build the model so it can be handed over and interrogated — assumptions on their own tab, sources stated, sensitivities live rather than hard-coded. A study a counterparty can test is worth considerably more than one they have to take on trust.