One approach uses the subjective judgments of the decision maker, such as direct or lost-horse forecasts. A direct forecast involves estimating the value to be forecast without any intervening steps. A lost-horse forecast starts with the last known value of the item being forecast and then lists the factors that could affect the forecast, assesses whether they have a positive or negative impact, and makes the final forecast. Statistical methods, a second approach, involve extending a pattern observed in past data into the future. The best-known statistical method is linear trend extrapolation.

Surveys of knowledgeable groups are a third approach, involve obtaining information such as the intentions of potential buyers or estimates provided by the salesforce.