forecast
Download as PPTX, PDF1 like95 views
This document outlines a sales forecasting methodology that uses time series analysis and exponential smoothing. It identifies trends, seasonality, and irregular components in historical sales data. It derives seasonal adjustment factors and uses exponential smoothing to forecast, adjusting for trends. The forecast is compared to actual sales and further refined through joint review with marketing, operations, and finance considering factors like promotions, new products, and production capacity.
forecast
- 1. Sales Forecast Mohammed Jawed Khan Business Analyst FSO Business Solution
- 2. Trend Models Linear VSMovingAveragesmoothened plot 0 2000 4000 6000 8000 10000 12000 14000 16000 0 12 24 36 48 60 72 No.ofCarsSold Months since 2007 Series1 Linear (Series1) 2 per. Mov. Avg. (Series1) 2008 2009 2010 2011 2012 2013
- 3. Quantitative Forecast Method Time Series has four components: Trend(T), Cycles(C), Seasonal(S), Irregular(I) Y= T x C x S x I Identifying factors that cause variation. Isolating, Analyzing and measuring the effect of these factors independently.
- 4. Approach Derive same month sale factor (SEASONALITY FACTOR) Removing seasonality factor from time series. Forecasting with Exponential Smoothening. Adjust the forecasted value with the seasonality factor. Apply “trend smoothening factor” to the forecasted value.
- 5. Seasonality Adjustment:- Computing the average demand for each month. Computing the seasonal index for each season. Adjust the average forecast for next year by the seasonal index.
- 6. ExponentialSmootheningmethod:- Smoothening: remove the effect of random variation due to the irregular effect of the time series without losing the pattern. Exponential smoothening method incorporates a factor (α) which includes the effect of history ( recent or past depending on value of α ) to forecast demand. Ft=Ft-1 + α ( Dt-1 - Ft-1) Weight Assigned to Smoothening Constant Most Recent Period (α) 2nd Most Recent Period α (1- α ) 3rd Most Recent Period α (1- α )2 4 th Most Recent Period α (1- α )
- 7. Trend Adjustment Factor The Exponential Smoothening model work only for a level pattern. If there is a trend in the data the models need to be modified, otherwise the forecasts will "lag" the trend. The model takes the basic exponential smoothing equation and just adds a trend component to compensate for the additional pattern using a smoothing constant ß for trend adjustment fact FITt+1 = Ft+1 + Tt+1 Tt+1 = ß(Ft+1 - Ft ) + (1 – ß) Tt
- 8. Comparison of the results 0 2000 4000 6000 8000 10000 12000 14000 16000 Jun'12 Jul'12 Aug'12 Sep'12 Oct'12 Nov'12 Dec'12 Jan'13 Feb'13 Mar'13 Apr'13 May'13 Achieved Sales Set Target Calculated
- 9. Additional Consideration Marketing adjusts the forecast. These adjustments include promotions of existing products, introduction of new products, or the elimination of products. Operations checks forecast against existing capability. Marketing, operations, and finance jointly review forecast and resource issues. Executives meet to finalize forecast and capacity decisions.
- 10. Sales and Operations Planning
Editor's Notes
- #3: How our sales history since 2008 looks like. A linear increasing trend but considerable fluctuation.
- #4: How do we see the components of time series
- #5: With the limited data we couldn’t account for erratic cycle effect which didn’t repeat
- #9: More complex analysis would give better result, but would require data like effect of economic cycles. Sales are mostly target driven. We see if 8 times of the month the actual was more than the target, it means capacity was not fully utilised.