Presentation Transcript

1. Demand Forecasting and Managing Variability in a Supply Chain

2. Learning Objectives • Role of forecasting in a supply chain • Components of a demand forecast • Demand forecasting using historical data • Analysing forecast errors • Managing demand/supply in a supply chain • Forecasting with Excel

3. Role of forecasting • Demand forecasts form the basis of all planning decisions in a supply chain • Push: produce to anticipated demand levels • Pull: set capacity and component availability levels • Forecast time horizons • Short term (days, weeks): shift scheduling • Medium Term (weeks, months): workforce planning, materials purchasing, promotions • Long term (months, years): capacity expansion, capital/financial budget

4. Characteristics of Forecasts • Forecasts are always wrong! • Expected value • Error/variability from the expected value • Long-term forecasts are usually less accurate than short-term forecasts • Aggregate forecasts are usually more accurate than disaggregate forecasts • Mature products with stable demand are easier to forecast than seasonal goods or “fashion” items with short product-life

5. Influences on Customer Demand “Predictions are usually difficult, especially about the future” – Yogi Berra • Historical patterns • Past demand -> future demand • Seasonality? Trend? • Externalities • Weather • State of the economy • Internal factors • Planned promotional/discount campaigns • Display position and advertising efforts • Competitors’ actions

6. Components of Observed Demand Observed demand (O) = Systematic component (S) + Random component (R) • Forecasting should focus on identifying the systematic component • Systematic component = Expected value of demand • Time series model: • (Basic) demand level • Trend, rate of growth/decline in demand per period • Seasonality, (predictable) seasonal fluctuations

7. Forecasting Methods • Qualitative • Subjective, human judgement and opinions • Little historical data available, e.g. new product, .com • Time Series • Assume past history is good indicator of future demand • Best for stable environments • Easy to implement • Causal • Assume other (measurable) factors that is correlated with demand • Models (e.g. regression) identify factors and quantifies the strength of the correlations • Simulation • Assumes some underlying principles of customer behaviour and develop possible scenarios in the future to predict demand

8. Basic Approach to Demand Forecasting • Understand the objective of forecasting • Forecast horizon; affected by suppliers’ lead times • Integrate demand planning and forecasting • Co-ordination between marketing, production and suppliers • Identify major factors that influence the demand forecast • Growth Trend? Seasonality? • Substitutes and complementary products? • Understand and identify customer segments • Levels of aggregation • Determine the appropriate forecasting technique • Geographical location, product life-cycle, etc. • Establish performance and error measures for forecasts • Assess cost impacts; consider investments in improving forecasts accuracies No forecast is perfect; system must be flexible and have contingency plans to handle forecast errors

9. Common Time Series Patterns • Constant (stationary) • Increasing/decreasing linear trend • Seasonal variations • Non-linear (e.g. exponential) trend • Combinations • Additive • Multiplicative • Mixed

10. Demand Time Common Time Series Patterns Purely Random Error - No Recognizable Pattern Increasing Linear Trend Demand Time Seasonal Pattern plus Linear Growth Seasonal Pattern Demand Demand Time Time

11. Curvilinear Trend (quadratic, exponential) Demand Time Change of variables? Other Time Series Patterns

12. Underlying model and definitions Systematic component = (level + trend) x seasonal factor L = estimate of level for period 0 (de-seasonalised demand) T= estimate of trend (increase/decrease in demand per period) St= Estimate of seasonal factor for period t Dt= Actual demand observed for period t Ft= Forecast of demand for period t Ft+k = [ L+ (t+k)T ]St+k

13. De-seasonalising Demand • De-seaonalised demand is the demand that would have been observed in the absence of seasonal fluctuations • The periodicity p is the number of periods after which the seasonal cycle repeats itself (e.g. if period length = 3 months, p = 4)

14. Estimating Model Parameters • Seasonal factors: • Seasonal factor for a given period (in the future) can be estimated by averaging seasonal factors of periods of corresponding seasons

15. Static Forecasting • Values of L and T estimated based on a set of data • Methods: • Simple averaging • Linear regression • These (static) values of L and T are used for future forecasts

16. Example: Natural Gas

17. Adaptive forecasting • The estimates of level, trend and seasonality are updated after each demand observations

18. Moving Average • Assumes no trend and no seasonality • Level estimate is the average demand over most recent N periods • Update: add latest demand observation and and drop oldest • Forecast for all future periods is the same • Each period’s demand equally weighted in the forecast • How to choose the value of N? • N large => • N small =>

19. Simple Exponential Smoothing (No trend, no seasonality) • Rationale: recent past more indicative of future demand • Update: level estimate is weighted average of latest demand observation and previous estimate • a is called the smoothing constant (0 < a < 1) • Forecast for all future periods is the same • Assume systematic component of demand is the same for all periods (L) • Lt is the best guess at period t of what the systematic demand level is

20. Simple Exponential Smoothing - Example L0= 22083 a = 0.1 F1 = L0 D1=8000 E1 = F1 – D1 = 22083 – 8000 = 14083 L1 =a D1 + (1 - a ) L0 = (0.1)(8000) + (0.9)(22083) = 20675 F2 = L1= 20675, F10 = L1 = 20675

21. Simple Exponential Smoothing • Update: new level estimate is previous estimate adjusted by weighted forecast error • How to choose the value of the smoothing constant a? • Large a => responsive to change, forecast subject to random fluctuations • Small a => may lag behind demand if trend develops • Incorporates more information but keeps less data than moving averages • Average age of data in exponential smoothing is 1/a • Average age of data in moving average is (N+1)/2

22. Understanding the exponential smoothing formula • Demand of k-th previous period carry a weight of hence the name exponential smoothing • Demand of more recent periods carry more weight

23. Exponential Smoothing with Seasonality (no trend) • De-seasonalise demand data • Apply exponential smoothing update • Seasonalise forecast

24. Trend corrected exponential smoothing (Holt’s model) • b is the smoothing constant for trend updating • If bis large, there is a tendency for the trend term to “flip-flop” in sign • Typical b is a2

25. Holt’s model - Example L0= 12015 T0=1549 a = 0.1 b = 0.2 F1 = L0 + T0 = 12015 +1549 = 13564 D1=8000 E1 = F1 – D1 = 13564 – 8000 = 5564 L1 =a D1 + (1 - a )(L0+ T0) = (0.1)(8000) + (0.9)(13564) = 13008 T1 =b (L1- L0) + (1 - b )T0 = (0.2)(13008 - 12015) + (0.8)(1549) = 1438 F2 = L1+T1= 13008+1438 = 14446, F10 = L1 + 9 T1= 13008 + 9(1438) = 25950

26. Trend and seasonality corrected exponential smoothing (Winter’s model)

27. Winter’s model - Example L0= 18439 T0=524 S1 = 0.47, S2 =0.68, S3 =1.17, S4 =1.67, a = 0.1, b = 0.2, g = 0.1, F1 = (L0 + T0) S1 = (18439 + 524)(0.47) = 8913 D1=8000, E1 = F1 – D1 = 8913 – 8000 = 913 L1 =a (D1/S1) + (1 - a )(L0+ T0) = (0.1)(8000/0.47) + (0.9)(18439+524) = 18769 T1 =b (L1- L0) + (1 - b )T0 = (0.2)(18769 - 18439) + (0.8)(524) = 485 S5 =g (D1/L1) + (1 - g)S1 = (0.1)(8000/18769) + (0.9)(0.47) = 0.465 F2 = (L1+T1)S2 = (18769+485)(0.68) = 13093, F11 = (L1 + 10 T1)S11= (18769 + 10(485))(1.17) = 27634

28. Analysing Forecast Errors • Monitor if current forecasting methods accurate • Consistently under-predicting? Over-predicting? • When should we adjust forecasting procedures? • Understand magnitude of forecast error • In order to make appropriate contingency plans • Assume we have data for n historical periods

29. Measures of Forecast Error • Mean Square Error (MSE) • Estimate of variance (s2) of random component • Mean Absolute Deviation (MAD) • If random component normally distributed, s=1.25 MAD • Mean Absolute Percent Error (MAPE)

30. Tracking Errors • Errors due to: • Random component • Bias (wrong trend, shifting seasonality, etc.) • Monitor quality of forecast with a tracking signal • Alert if signal value exceeds threshold • Indicates underlying environment changed and model becomes inappropriate • Tracking signal sometimes used as smoothing constant • Reactive, but often unstable in practice

31. Summary so far • Importance of forecasting in a supply chain • Forecasting models and methods • Exponential smoothing • Stationary model • Trend • Seasonality • Measures of forecast errors • Tracking signals • Regression models Forecasting

32. Regression Analysis • Statistical technique to determine the degree of association between set of variables and demand. • Given values of the predictor (independent) variables, the regression equation provides a forecast of demand. Forecasting

33. Simple Linear Regression • Only one independent variable • Time series: • how demand (dependent variable) changes over time (independent variable) • Scatterplots Forecasting

34. (Pearson’s) Sample Correlation Coefficient X = (S xi)/n Y = (S yi)/n SX = sqrt(S (xi – X)^2/(n-1)) SY = sqrt(S (yi – Y)^2/(n-1)) SXY = S (xi – X) (yi – Y)/(n-1) r = SXY / SX SY -1 < r < 1 r measures how good the linear relationship is between the variables Watch out for induced correlation trap! Forecasting

35. Linear Regression Find the relationship: Y = a + b X For each data point (xi , yi), the residual error is ei = yi – (a + b xi) Choose a and b to minimise S ei^2 b = SXY / SX^2, a = Y – b X Forecast: a newsituation yields x’ and the predicted value for Y is a+bx’ Forecasting

36. Least Squares Regression Forecasting

37. Nonrandom Errors Forecasting

38. Special Forecasting Difficulties for Supply Chains • New products and service introductions • No past history • Use qualitative methods until sufficient data collected • Examine correlation with similar products • Use a large exponential smoothing constant • Lumpy derived demand • Large but infrequent orders • Random variations “swamps” trend and seasonality • Identify reason for lumpiness and modify forecasts • Spatial variations in demand • Separate forecast vs. allocation of total forecasts

39. Managing (Predictable) Variability How should we plan production to meet forecasted demand?

40. Managing Supply • Chase strategy: production matches demand • No inventory • Capacity under-utilised during low demand periods • Level (stable) production • High capacity utilisation • Personnel management/training simpler • Inventory build up in anticipation of seasonal demand variations; obsolescence risk • Order backlog; loss of goodwill • Capacity vs. Inventory trade-off

41. Managing Capacity • Time flexibility from workforce • More shifts during busy season • Overlapping shifts • Seasonal/Temporary workforce • Subcontracting (Use of dual facilities) • Internal or main facility: focus on efficiency (low cost), level production • Peak demands subcontracted out or produced on more flexible facilities • Designing product flexibility into the production process • Production lines re-configurable for different production rates • Complementary products produced in same facility (e.g. snowblowers and lawnmowers)

42. Managing Inventory • Use common components across multiple products • Aggregate demand more stable (forecast more accurate) • Less obsolescence risk • Build inventory of high demand or predictable demand items • Delay production of “fashion” items until closer to selling season

43. Managing Demand • Promotion/Discount pricing • Impact on demand • Market growth (attract new customers) • Increase overall demand • Stealing market shares (attract competitors’ customers) • Increase overall demand • Forward buying (own customers buy earlier) • Demand “smoothing”? • Timing of promotion: peak vs. non-peak • Change in revenue vs. change in costs

44. Factors affecting promotion timing

45. Demand management • Changes in demand impact production scheduling and inventory levels • Promotion at peak demand periods increases demand variability • Promotion at non-peak “smoothes” demand • Pricing and production planning must be done jointly • Preempt,don’t just react, to predictable variability • Actively managing predictable variability can be a strategic competitive advantage

46. Flexibility and Quick Response No forecast is perfect … There is no better forecast than the actual order! • If supply chain is responsive, then effect of forecast errors is minimised • Especially important when demand is unpredictable • Example: National bicycle • Difficult to predict styles in sports bikes • Re-engineer supply chain • Customers design their own bike on Internet • Bike produced in Kashiwara and delivered in 2 weeks!

47. Summary • Importance of forecasting • Exponential smoothing models • Level, Trend, Seasonality • Measuring forecast error; tracking • Predictable variability • Managing supply • Managing demand • Unpredictable variability? • Co-ordinated management of supply and demand optimises profit • Forecasting with Excel