Exit Multiples Expressed as Uncertainties

by a searcher in Portland, OR, USA

We know that Exit Multiples are highly correlated with Revenue, EBITDA and Margin Expansion. Most spreadsheet athletes will have no issue handling this by setting up 2x2 sensitivity grids of the various parings. The grids would map visually well to show possible outcomes on contour plots. To the extent that we all know that such models are wrong, but, to some degree, they can be useful on the other hand, what could be misleading about all this?

In a way, these discrete outcomes are presumably expressed as almost equally likely. without much more information. For obvious reasons, these equally likely outcomes seem to be more than a bit counter intuitive. But, what could we do to better represent true uncertainties found about Exit Multiples, Revenue, EBITDA and Margin Expansion. It may sound as if I'm calling for a 4x4 covariance matrix of these four business levers. That would be a much more rigorous multivariate regression technique.

However, let's say you're meeting your targets with high confidence per your value creation plan for Revenue, EBITDA and Margin Expansion after several few years working vigilantly as an operator . Yet, Exit Multiples would be the last unknowns to sort out from the market. In this case, we can express our uncertainties about what the market may do with more realistic or empirical distributions rather than treating Exit Multiples as uniform distributions as was previously described.

An empirical distribution can be built following these steps.
1.) Get BVR's 90th, 50th, 5th percentiles on historical Exit Multiples that best represent your business controlling for Revenue, EBITDA and Margin Expansion in time

2.) Create a three-point Metalog Distribution and get randomly generated values for Exit Multiples

3.) Calculate the new purchase price for your business using the randomly generated values for Exit Multiples

I used one variable in this exercise. This treatment can be done with one or more variables. Correlated variables would be a bit more tricky. But, if variables can be treated as fairly independent of each other, then, this would be a great way to animate your spreadsheets.