A two-stage DCF projects free cash flow (or owner earnings) year by year for N years at an early growth rate, then sums up everything after year N as a single forever-value that grows at a slow steady rate, and discounts both pieces back to today using the cost of capital.
The pieces: • Starting cash flow: the last twelve months of free cash flow or owner earnings • Early growth (g₁): the rate you project for the first N years • Long-run growth (g_T): the slow forever growth after year N. It must be well below the discount rate, usually 2% to 3%, and never above about 4%, because no business outgrows the whole economy forever. • Discount rate (r): the cost of capital. See the quality-band approach for choosing it without faking precision. • Projection horizon (N): usually 5 or 10 years • Net-debt bridge: add cash and subtract total debt to go from the value of the whole business to the value for shareholders
The answer depends entirely on the assumptions. Change any one and the answer moves. A grid that nudges the discount rate and the early growth rate each up and down by 2 percentage points shows the range of believable answers.
Limits: • Garbage in, garbage out. A DCF on a business with no track record of free cash flow is fiction. • The forever-value usually dominates, often 60% to 80% of the total. That means the answer rests mostly on the long-run growth assumption, not the year-by-year projections. • Cyclical businesses break it. Starting from a boom year gives a crazily high value, and starting from a bust year gives the opposite. Smooth the inputs across the cycle. • A single DCF number gives a false sense of certainty. Always show it as a range, next to the no-growth floor (EPV) and the growth the market is assuming (the reverse DCF).