"Forecasting is difficult, especially if it's about the future."
In general, two forecasting methods exist. The first, tops down, seeks to identify an addressable market and revenue becomes a function of market share. The second, bottoms up, is often best captured by detailed assumptions regarding how to get to revenue. For example, I am partial to the productive sales rep model: W# of sales reps * X rep maturity ramp * Y quota per rep * Z productivity per rep.
In my experience, both models are straightforward to develop. The art lies, however, in the skillful blending of the two approaches to satisfy a board, prospective investor, sales team, etc.
Forecast errors compound across time. Therefore, it pays to focus on three to six month bottoms up plans. A near-term bottoms up plan requires thoughtful detail and projections based on specific variables - how many sales heads, quota, ASP, etc. Operating history then turns the assumed variables into actuals and the model begins to solidify.
Beyond six months, the bottoms up approach is simply an exercise in guess work. How then should a management team think about the long-term? I have found two useful models. The first is to examine historical comparables - for analog companies, what was the revenue ramp rate? A model can be focus on projections that assume either the average, median, or top decile growth rates for comprable companies. Ie - it happened before and history looked like $xm, $ym, and $zm over three years.
Another model is to set strategic market share goals. In essence, the long term plan focuses more on clearly defining an addressable market and a strategic goal of capturing x% of the market. Again, the art lies in defining the market, the market's growth rates, and a strategy for competing for 20-30% of the market at large.
It pays to invest in both near-term detail and longer term aspirations. The former assuages board members and investors and provides them confidence in your operating bona fides. The latter fires the imagination and helps paint the picture of how a very large, venture return company can be built.