Title: Northcutt Bikes Case Answers
1Northcutt Bikes CaseAnswers
2Q1 Demand Data Plot
3Q1 Plot Shows
- There is seasonality
- There is a trend
- Forecast should take into account both
4Construction of base indices
Year 2008 2009 2010 2011 Mean Base
January 0.53 0.72 0.59 0.59 0.61
February 0.74 0.74 0.95 1.09 0.88
March 0.88 0.84 0.79 0.98 0.87
April 1.09 1.00 1.18 0.92 1.05
May 1.10 1.16 1.15 1.27 1.17
June 1.60 1.57 1.39 1.51 1.52
July 1.29 0.94 1.35 1.56 1.28
August 1.19 1.30 1.43 0.71 1.16
September 1.00 1.13 0.91 1.08 1.03
October 1.09 0.74 0.96 0.77 0.89
November 0.73 0.99 0.78 0.84 0.84
December 0.74 0.88 0.51 0.67 0.70
Mean Demand 818.42 990.50 1032.08 1181.25
5 Multiple Regression ResultsX is Period and
Base
Regression Statistics Regression Statistics
Multiple R 0.982917071
R Square 0.966125969
Adjusted R Square 0.964620456
Standard Error 59.82147676
Observations 48
ANOVA
df SS MS F
Regression 2 4592970.404 2296485.202 641.7256395
Residual 45 161037.4087 3578.609082
Total 47 4754007.813
Coefficients Standard Error t Stat P-value
Intercept -219.4209094 35.31667659 -6.212954633 1.50687E-07
Period 8.730540524 0.623285303 14.00729407 5.12015E-18
Base 1011.295853 30.74315604 32.89499139 4.07081E-33
6Q2 Forecasting Methods
- Multiple regression or MR (Y is forecast, Xs are
period and base) MAD 45.096 - Simple regression or SR (deseasonalize demand,
seasonalize forecast, X is period) MAD 32.403 - Exponential Smoothing or ES (adjusted for trend
and seasonality) MAD 13.258
7Q2 Forecast for January April 2012
Month Mean Base Period MR SR ES
January 0.61 49 825.27 745.12 720.56
February 0.88 50 1107.05 1082.68 1039.50
March 0.87 51 1105.66 1078.04 1027.69
April 1.05 52 1296.43 1310.32 1240.31
8Q3 Best Forecast
- Exponential smoothing forecast has lowest MAD
- Disadvantages the exponential smoothing forecast
should be updated frequently (say once a month).
9 Q4 Additional Information
- Jans knowledge of market could be used to
- - Add additional independent variable to
multiple regression - - Be used to adjust other forecasts (caution
should be used, however) - Monthly increments best as forecast can react to
latest information, provided this is not costly -
10Q5 Ways to Improve Operations
- Quicker response reduce manufacturing lead
times possibly implement online ordering - Suppliers reduce lead times set contracts
- Improve information systems
- Work force increase flexibility temps
11Q6 Recommendations
- Operation is likely not too large - Jan can
control operation effectively if she - delegates
- improves information system
- reduces lead times
- implements lean (to be discussed)
- uses different modes of operation for different
style bikes - Information needed on costs of above
12Questions ?