You *could* try to estimate a company's growth rate and future margins in order to figure out what a company is worth. But in Expectations Investing, Michael Mauboussin and his co-author suggest reverse engineering the market's expectations instead.
They push it as an approach less prone to error than the typical discounted cash flow process. They may be right about that, though I can't say for sure. But it is still an approach fraught with danger, in my opinion, as it too relies on a bunch of assumptions that, when tweaked, can change a valuation by orders of magnitude.
The process involves looking at a current stock price and current analyst expectations in order to determine an estimate of for how long the market anticipates the current growth expectations to continue. This gives an investor an idea of how long a company must grow at what rate in order for the current market price to be correct.
The investor then considers some qualitative factors to determine whether the company can outperform (or underperform) these expectations, and then uses a probability-based method to triangulate an intrinsic value estimate.
In theory, I think this process is great. I feel the same way about discounted cash flows. In practice, however, I think these procedures prioritize false precision over rough accuracy to their detriment. I don't think one can predict sales growth in the double-digits for several years in a row, but that's essentially what these processes seem to boil down to, to my understanding at least.
I don't think I learned much from this book, but if this process seems to make more sense to you than it does to me, you might like it.