CHAPTER 5 DEMAND FORECASTING & COLLABORATIVE PLANNING, FORECASTING, & REPLENISHMENT 명지대학교산업시스템공학부
You should be able to: Learning Objectives Explain the role of demand forecasting in a supply chain. Identify the components of a forecast Compare and contrast qualitative and quantitative forecasting techniques Assess the accuracy of forecasts Explain collaborative planning, forecasting, and replenishment 2
Introduction Chapter Five Outline Matching Supply and Demand Forecasting Techniques Qualitative Methods Quantitative Methods Forecast Accuracy Collaborative Planning, Forecasting, and Replenishment Software Solutions 3
Introduction Forecasting provides an estimate of future demand The goal is to minimize forecast error. Factors that influence demand and whether these factors will continue to influence demand must be considered when forecasting. Improved forecasts benefit all trading partners in the supply chain. Better forecasts result in lower inventories, reduced stock-outs, smoother production plans, reduced costs, and improved customer service. 소매업의경우품절발생시 41% 는판매를잃는다. (Proctor & Gamble 사 ) 부정확한수요예측은공급사슬을따라 bullwhip effect 초래 : 품절, 판매기회손실, 과잉재고및진부화, 원료부족, 미흡한시장대응, 낮은수익률 Nike 사 2001 년과잉예측으로과도한재고를서둘러처분. 수익률 33% 감소 4
Matching Supply and Demand 시장은고객이주도권을쥔 pull 환경으로변화, 원하는제품과인도시간을고객이결정. Suppliers must accurately forecast demand so they can produce & deliver the right quantities at the right time at the right cost. Suppliers must find ways to better match supply and demand to achieve optimal levels of cost, quality, and customer service to enable them to compete with other supply chains. 방법 -1: 공급자가충분한재고유지 방법 -2: 유연한가격정책 방법 -3: 잔업, 하청, 임시노동력활용 스포츠오버마이어사 : 고급스키복디자인및머천다이징전문회사 판매시기 : 9 월 ~ 다음해 1 월, 12~1 월피크, 수요예측, QR(Quick Response) 시스템을통해수요와공급불일치완화 5
Forecasting Techniques Qualitative forecasting is based on opinion and intuition. Quantitative forecasting uses mathematical models and historical data to make forecasts. Time series models are the most frequently used among all the forecasting models. ( 응답자의 60% 사용 ) 6
Qualitative Forecasting Methods Generally used when data are limited, unavailable, or not currently relevant. Forecast depends on skill & experience of forecaster(s) & available information. - 자료가별소용없는상황에서장기전망하는경우 - 수요관련자료가없는신제품을도입할경우 7
Four qualitative models used are: 1. Jury of executive opinion: 경험을가진사람들이협력하는장점. 한사람이토론을압도하는경우신뢰도가떨어질수있다. 장기계획, 신제품도입에적합 2. Delphi method : 내외부전문가에게여러번의의견조사실시. 매회의견을집계요약하여참여자에게제공하고참여자들은다시자신의의견을수정가능. 합의에이르렀다고판단될때까지이과정을반복. 고위험기술예측, 대형프로젝트, 신제품도입시적합 3. Sales force composite: 영업사원들의지식을근거로수요예측. 실제판매치가목표치를초과할수록보너스를많이받을경우, 비관적인예측치를제시하는단점. 4. Consumer survey : 소비자를대상으로구매습관, 신제품아이디어, 제품에대한의견을설문. 표본집단선정하고일정수준이상응답을받아야하는단점. 8
Quantitative Methods Time series forecasting- based on the assumption that the future is an extension of the past. Historical data is used to predict future demand. Associative forecasting- assumes that one or more factors (independent variables) predict future demand. Since quantitative methods rely solely on past demand data, they become less accurate as the forecast s time horizon increases. For long time horizon forecasts, It is generally recommended to use a combination of quantitative and qualitative techniques. 9
Components of Time Series- Data should be plotted to detect for the following components: Trend variations: either increasing or decreasing Cyclical variations: wavelike movements that are longer than a year. (ex) business cycle(recession or expansion) Seasonal variations: show peaks and valleys that repeat over a consistent interval such as hours, days, weeks, months, years, or seasons. (ex) 빙과류, 동계스포츠, Random variations: due to unexpected or unpredictable events such as natural disasters. 10
Time Series Forecasting Models Simple Moving Average Forecasting Model. Simple moving average forecasting method uses historical data to generate a forecast. Works well when demand is fairly stable over time. 11
Ex 5-1 12
특징 : 자료의수가적을수록실제수요의변동에민감 우발변동에도민감하게반응 사용하기쉬우나추세변동에빨리반응할수없다. 13
Time Series Forecasting Models Weighted Moving Average Forecasting Model- based on an n-period weighted moving average, follows: 14
Ex 5-2 15
특징 : 최근실적치를더강조할수있다. 단순이동평균법보다수요의변동에더민감하게반응 평균을취하므로여전히변동에시차가있다. 16
Time Series Forecasting Models Exponential Smoothing Forecasting Model- a weighted moving average in which the forecast for the next period s demand is the current period s forecast adjusted by a fraction of the difference between the current period s actual demand and its forecast. Only two data points are needed. Ft+1 = Ft+ (At-Ft) or Ft+1 = At + (1 ) Ft Where Ft+1 = forecast for Period t + 1 Ft = forecast for Period t At = actual demand for Period t = a smoothing constant (0 1). 17
Ex 5-3 18
특징 : 두개의자료만필요 계산과정이간단하므로매우널리사용됨 자료가추세나계절성을거의보이지않을때적합 값이 1 에가까울수록최근자료를많이반영 값이작을수록과거수요를많이반영 최초의예측치는정성적방법을사용하여구하든지당기의실제수요로대치할수있다. 19
Time Series Forecasting Models Trend-Adjusted Exponential Smoothing forecasting Model. a trend component in the time series shows a systematic upward or downward trend in the data over time. F t = A t + (1 - )(F t-1 + T t-1 ), T t = ß(F t -F t-1 ) + (1 ß)T t-1, and the trend-adjusted forecast, TAF t+m = F t + mt t where F t = exponentially smoothed average in Period t A t = actual demand in Period t T t = exponentially smoothed trend in Period t = smoothing constant (0 1) ß = smoothing constant for trend (0 ß 1) 20
추세평활계수 B 가클수록최근추세의변화에비중을많이둠. 추세평활계수 B 가작을수록최근추세의변화에비중을적게둠. Example 5-4 21
Time Series Forecasting Models Linear Trend Forecasting Model. The trend can be estimated using simple linear regression to fit a line to a time series. Ŷ = b 0 + b 1 x where Ŷ = forecast or dependent variable x = time variable B 0 = intercept of the line b 1 = slope of the line 22
Ex 5-5 23
Associative Forecasting Models- One or several external variables are identified that are related to demand Simple regression. Only one explanatory variable is used and is similar to the previous trend model. The difference is that the x variable is no longer a time but an explanatory variable. Ŷ = b 0 + b 1 x where Ŷ = forecast or dependent variable x = explanatory or independent variable b 0 = intercept of the line b 1 = slope of the line Example 5-6 24
Associative Forecasting Models- Multiple regression. Where several explanatory variables are used to make the forecast. Ŷ = b 0 + b 1 x 1 + b 2 x 2 +... b k x k where Ŷ = forecast or dependent variable x k = kth explanatory or independent variable b 0 = intercept of the line b k = regression coefficient of the independent variable x k 25
Forecast Accuracy The formula for forecast error, defined as the difference between actual quantity and the forecast, follows: Forecast error, e t = A t - F t where e t = forecast error for Period t A t = actual demand for Period t F t = forecast for Period t 26
Forecast Accuracy Several measures of forecasting accuracy follow: (p137-138) Mean absolute deviation (MAD) MAD = 0 : the forecast exactly predicted actual demand. MAD > 0 : 과대또는과소예측여러예측기법중에서 MAD 가가장작은기법을선호하게된다. Mean absolute percentage error (MAPE)- provides perspective of the true magnitude of the forecast error. Mean squared error (MSE)- analogous to variance, large forecast errors are heavily penalized Running sum of forecast errors (RSFE)- an indicator of bias in the forecasts RSFE > 0 : 예측치가너무낮음 RSFE < 0 : 예측치가너무높음 Tracking Signal (TS)- is checked to determine if it is within the acceptable control limits. 일반적으로고수요품목은 ±4 값을, 저수요품목은 ±8 을사용. TS 가작아질수록예외를감지할확률이높아진다. 27
Forecast Accuracy Example 5-7 28
Collaborative Planning, Forecasting, & Replenishment Collaborative Planning, Forecasting, & Replenishment (CPFR) Collaboration process whereby supply chain trading partners can jointly plan key supply chain activities from production and delivery of raw materials to production and delivery of final products to end customers American Production and Inventory Control Society (APICS). Objective of CPFR- optimize supply chain through improved demand forecasts, with the right product delivered at right time to the right location, with reduced inventories, avoidance of stock-outs, & improved customer service. Value of CPFR- broad and open exchange of forecasting information to improve forecasting accuracy when both the buyer and seller collaborate through joint knowledge of base sales, promotions, store openings or closings, & new product introductions. 29
Collaborative Planning, Forecasting, & Replenishment Most firms implement CPFR based on the Voluntary Inter-industry Commerce Standards (VICS) Process Model. GCI(Global Commerce Initiative) 는 VICS 의 CPFR 모델을보완하여 CPFR 표준을제시하였다. 30
Collaborative Planning, Forecasting, & Replenishment VICS s CPFR process model 31
Collaborative Planning, Forecasting, & Replenishment Step 1: Develop Collaboration Arrangement 협력의목적, 의견충돌해소를위한기본적규칙, 공유정보의기밀보장, 재무인센티브, 자원배정, 성공평가척도, 판매예측관련한예외기준과조정간격, 예측기간, 허용한계를포함한동결기간 (frozen time period) 등 Step 2: Create Joint Business Plan 제품카테고리전략 ( 상호대체가능한유사제품군 ). 재고정책 ( 최소주문량, 리드타임, 주문간격, 동결기간, 안전재고지침등 ). 판촉활동, 가격정책등 Step 3: Create Sales Forecast 소매점의 POS 정보, 물류센터출고현황, 제조과정소요량, 점포개폐, 신제품도입등의자료공유수요예측시여러거래당사자들의예측치를고려하여예측. Step 4: Identify Exceptions for Sales Forecast 예 ) 소매점재고가 95% 이하인경우, 판매예측오차가 20% 이상인경우 Step 5: Resolve/Collaborate on Exception Items 32
Collaborative Planning, Forecasting, & Replenishment Step 6: Create Order Forecast POS 자료, 과거수요, 선적자료, 현재생산능력, 최소주문량, 리드타임, 주문간격, 불변기간, 안전재고, 현재상황, 신제품도입, 점포개폐등분석 주문예측은제품별, 수납장소별, 시기별로상세하게주문요구를기술 Step 7: Identify Exceptions for Order Forecast 협력계약에명시된고객서비스지표, 주문충족률 (order fill rate), 예측오차등의이상식별 Step 8: Resolve/Collaborate on Exception Items Step 9: Order Generation 주문예측을확정주문으로전환 33
Collaborative Planning, Forecasting, & Replenishment CPFR 소프트웨어가 96 년에소개되었으나기대만큼광범위하게수용되지않고있다. CPFR 적용의장애물 : - 소매업자들이자신의민감한고유자료를공유하기를꺼린다. 예 ) 월마트납품업자는자신의자료를월마트와기꺼이공유하고자하나다른납품업자와는공유하기를꺼린다. - 자료의신뢰성부족 - 소매업자와제조업자의내부통합결여 - 기존정보시스템의노후 - 최고경영층의지원부족등 CPFR 소프트웨어 - Manugistics 의 Manugistics CPFR 솔루션 - i2 Technologies 의 i2 Demand Collaboration - Syncra Systems 의 SyncraXt 34
Homework 델파이기법에대해조사하시오. CPFR 에대해조사하시오 (www.cpfr.org, scm.gs1kr.org, www.i2.com, www.manugistcs.com 등참조 ) 스프레드시트를사용하여문제 2 (p153) 를푸시오. 35