There are several methods and they disagree. What was assumed is everything.
Valuation estimates what a company is worth. The fair value that comes out is not an objective fact but a figure derived from the inputs. Different people producing different numbers for the same company is not a calculation error; their assumptions differ.
The most widely used approach compares similar companies. You find what multiple of earnings or assets peers in the same industry trade at, and apply that multiple to the target. It is simple and reflects the current market mood, with the limitation that when the whole market is heated or depressed, that state becomes your benchmark.
The other approach forecasts the cash a company will generate and converts it to present value. It is more fundamental in theory, but it requires assuming cash flows years ahead, a discount rate, and a perpetual growth rate.
Adjusting a discount rate or perpetual growth rate by a single percentage point often shifts the final number by tens of percent. Presenting a result as one figure therefore projects more confidence than exists. Varying assumptions to build a range, and working backwards to see what assumptions justify today's price, is often more useful.
When looking at someone's target price, what matters is not the conclusion but the assumptions behind it. Seeing what growth rate was used and which companies were compared tells you how much to trust the figure. This explains the structure of the methods; it does not judge whether any stock is cheap or expensive.
Each button runs this keyword on that search engine