Experienced operators do not need either of these terms defined. What is worth writing down is when each of them lies, because both are one-number summaries, and one-number summaries mislead in predictable, describable situations.
The going-in cap rate lies whenever the numerator is not the truth about the property. Buy a building with a tenant paying thirty percent under market on a lease expiring in eighteen months and the cap rate reports the past, not the deal. Buy one with an over-market lease held up by a struggling tenant and the cap rate reports a fiction with an expiration date. Buy meaningful vacancy and the cap rate reports almost nothing at all. In each case the number is computed correctly and still wrong as a description of what you are buying.
Yield on cost lies in the other direction. It is a statement about the future: stabilized income over everything it took to get there. Its failure modes are aspiration dressed as arithmetic. A stabilized NOI built on rents the submarket will not actually pay. A capex budget scoped before anyone opened a wall. A timeline in which lease-up takes twelve months because the model said twelve months. Yield on cost is only as honest as the business plan underneath it, and it inherits every soft assumption without flagging any of them.
The number that carries real information is the spread between the two. That spread is the market's price for executing your business plan. A wide spread says you are being paid well for the risk and work of stabilization, if your stabilized number is real. A thin spread says you are underwriting effort for little reward, and past a point it says the market has already priced in the upside you thought you found. Development and reposition investors live on this spread; it deserves the same attention on an eight-figure strip center as on a ground-up deal.
Run your own deal through the widget below. Then apply the two-question test to the output. Question one: what would have to be true for the going-in cap to be a fair description of this asset? Question two: which single assumption inside the stabilized NOI, if it slipped, would collapse the spread? If you cannot answer the second question, the spread is not information yet. It is a hope with a basis-point label.
This is also where a one-number screen hands off to real underwriting. The spread tells you whether a deal deserves hours. It cannot tell you where the deal breaks, because that lives in the schedule of leases, the cost of each rollover, and the structure of the debt. PropCalc™ exists for that second stage: the same inputs that produce these two summary numbers feed a full lease-level model, so the screen and the underwriting stop being separate documents.