Sanity check
Valuation & CompsA test of an output that reaches the answer by a route independent of the calculation that produced it.
Also written: cross check, reasonableness check
The defining condition is independence. Recomputing the same formula a second time confirms that you can do arithmetic consistently, which was never the question. A check earns its name only when it comes from history, from the peer set, from a rough figure you can produce in your head, or from reading the output backwards to see what it implies.
The balance check is the internal exception, and its limits show why the others are needed. It proves the three statements are linked correctly and says nothing whatever about whether the assumptions are plausible.
Three external checks do most of the work on a valuation or a deal model. Compare the forecast margin against what the business earns today and what its peers earn, and require a named cause for any expansion. Divide the enterprise value the model produces by EBITDA to see what multiple your own work is asserting, then judge it against comparable companies and recent transactions. And compute how much of a discounted cash flow value sits beyond the explicit forecast, which is normally most of it.
Where a check fails, it has not proved the model wrong. It has told you where to look, which is the only thing a check has ever done.
Worked example
Revenue of 800 with operating profit of 100 and 50 of depreciation gives EBITDA of 150 and a margin of 18.8%. A model showing 24.0% in year five is claiming more than five points of expansion and needs a reason named.
The same business at an enterprise value of 1,350 is being valued at 9.0x that EBITDA, which is a claim to check against comparable companies rather than a number to accept.
Five years of free cash flow at 80, discounted at 9% with 2% terminal growth, gives 311 for the explicit period and 758 for the terminal value, so 71% of the 1,069 total sits beyond the forecast.