WACC

Start with the right side of a balance sheet, at market values. Somebody financed everything the company operates. Some lent to it, some own it. Each group parted with money it could have deployed elsewhere, and each therefore has a required return: the lenders' alternative is other bonds of the same risk; the shareholders' alternative is other equities of the same risk. The company's cost of capital is simply the value-weighted average of those two opportunity costs:

WACC = (E/V) × Re  +  (D/V) × Rd × (1 − t)

E = market value of equity      D = market value of debt      V = E + D
Re = cost of equity             Rd = pre-tax cost of debt     t = marginal tax rate

If preferred stock exists, it's a third leg, + (P/V) × Kp, where Kp = preferred dividend ÷ preferred price. You'll rarely see it in India, and only occasionally in the US.

What kind of number is WACC?

Three things this number is:

1. An opportunity cost, set in the market, today. It isn't what the company paid historically, and it isn't what management wishes. If ten-year money repriced last year, so did the WACC of every company on earth.
2. A property of the business being valued. Who's buying doesn't come into it. A stable FMCG cash flow is worth the same whether you discount it, a pension fund does, or a conglomerate that "has a 15% hurdle" does. Risk travels with the cash flows, whoever's spreadsheet they end up in. Hurdle rates above WACC are a capital-allocation policy, a topic for the value-drivers work later in the phase, and they aren't a valuation input.
3. A currency-denominated number. A WACC is 11% in rupees or 8% in dollars. A bare percentage means nothing, because each currency carries its own expected inflation, which is Fisher's point from the time-value work. This one idea drives the whole risk-free rate build.

How precise is WACC, and how is it built?

One thing it isn't, though, is precise. Every input below carries error. What you can stand behind is a range. From the terminal-value work onward you'll build ±1-point sensitivity into every model. Craft goes into making each input defensible and the assembly error-free. False precision comes later, from people who skipped the assembly.

Think of the build as an assembly line with six stations, taken in order. First comes the risk-free rate, then beta. Next is the cost of equity, then the cost of debt, then the weights. Last comes assembly, with a sanity check. After that you'll run two complete blank-sheet builds.