Sales Forecasting

Qualitative Methods

• Expert Opinion: Relies on the intuition and experience of senior managers or sales staff. This is most useful when historical data is unavailable, such as during a new product launch.
• Delphi Method: A structured communication technique where a panel of experts answers questionnaires in multiple rounds. After each round, a facilitator provides an anonymous summary, encouraging the experts to revise their earlier answers to reach a consensus.

Quantitative Methods

• Moving Averages: Calculates the average of the most recent $n$ periods to predict the next period. This method effectively smooths out short-term fluctuations but can lag behind significant trends.
• Exponential Smoothing: A weighted moving average technique where more weight is given to recent data. The formula is $F_{t+1} = F_t + \alpha(A_t - F_t)$, where $\alpha$ is the smoothing constant between 0 and 1.
• Regression Analysis: Uses mathematical equations to model the relationship between sales (dependent variable) and one or more independent variables (like advertising spend or GDP).

• Simple vs. Weighted Moving Average: A simple moving average treats all historical periods equally, while a weighted moving average assigns specific importance (weights) to certain periods, usually favoring the most recent data to capture current trends.

• Seasonality vs. Cyclicality: Seasonality occurs within a fixed period (e.g., monthly or quarterly), whereas cyclicality refers to long-term economic cycles (e.g., recessions) that do not have a fixed duration.

• Check the Smoothing Constant: In exponential smoothing problems, remember that a high $\alpha$ (close to 1) makes the forecast highly responsive to recent changes, while a low $\alpha$ (close to 0) produces a very stable, slow-moving forecast.

• Verify the Lag: When using moving averages in a trending market, the forecast will always 'lag' behind the actual data. If sales are rising, the moving average will consistently underestimate the next period.

• Error Calculation: Always use absolute values when calculating Mean Absolute Deviation (MAD). If you don't use absolute values, positive and negative errors will cancel each other out, giving a false impression of perfect accuracy.

• Sanity Check: After calculating a forecast, ask if the number makes sense given the trend. If your calculated forecast is lower than the last three periods of a clearly rising trend, you likely made a calculation error.

• Over-fitting the Model: A common mistake is creating a complex model that perfectly matches historical noise but fails to predict future reality. Simpler models often perform better in volatile environments.

• Ignoring External Shocks: Relying solely on internal historical data can be dangerous. A forecast that ignores a new competitor entering the market or a sudden change in regulation will likely be highly inaccurate.

• Confusing Correlation with Causation: Just because two variables move together (e.g., ice cream sales and shark attacks) does not mean one causes the other. Regression models must be built on logical, causal relationships.