[ " Easiest and naivest way is linear extrapolation: gathering the historic financial data and deriving a growth rate (CAGR) for any line on the income statement (sales, gross profit, operating profit, net income etc). If historically profit ratios have been constant you can just forecast sales and derive the rest as a percentage of sales. Check competitors to confirm profit margins are close to industrial average.

You can check management reports and see if there are any projects that can boost future sales. Maybe a successful R&D is coming through.

A more comprehensive way is understanding the product, assess the macro environment (GDP growth) and the prospects of that market. Can the company increase market share? Do a Porter analysis.

Break down the income and expense components of the earnings and see how each item can change in the future (e.g. a cheese factory needs to buy and store milk, what volumes can it support using existing assets?). In other words create a model of what drives the company.

Although all forecasts suck anyway, still forecasting each moving part of the company with forward looking quantitative and qualitative justifications is more reliable than just applying a flat rate based on historic performance.

In summary company valuation should include meeting with management and visiting the company unless everything is available and can be researched from your computer. " ]

What are your thoughts on this author's take on forecasting earnings? Why would we want to forecast a company's earnings? How would managers and investors use this information? Which of these approaches do you agree with? Which ones do you think would be of little value and why?