TimeSeriesAnalysis(英文版)

1、Time Series AnalysisBenefits and Uses of Time Series Benefits of time series Monitor sales performance over time remove variation in monthly sales caused by calendar differences and seasonality that can conceal potential problems with sales Accurately determine the direction and rate of growth/decli

2、ne in sales Quickly identify changes in sales trends and correlate them to factors affecting sales industry, company, competition Improve decision making regarding sales and marketing actions Uses of time series Assess current sales performance and evaluate the effectiveness of sales programs Determ

3、ine underlying sales trend and project year end sales Establish appropriate budgets for next year and estimate monthly budget spreadsTime series analysis is the primary sales analysis technique at A-B2021-12-12Time Series Analysis What is Time Series Analysis? How are Time Series plots developed? Wh

4、at are the advantages of Time Series Analysis? What are Time Series used for?2021-12-13What is Time Series Analysis ? Time series analysis is a statistical technique used to analyze and monitor sales volume over time.2021-12-14Why Time Series ?Beer Sales050100150200250199819992000200120022003Thousan

5、d Barrels Beer sales are highly seasonal It is very difficult to evaluate monthly sales over time.2021-12-15How do time series work? Monthly variation in sales is caused by two major factors Seasonality Selling Days (calendar effects) Time Series technique statistically removes the effects of these

6、two factors Time Series technique uses the X-11 procedure for seasonal adjustments The X-11 procedure was developed by the U.S. Bureau of Census in the 1950s. It was brought to A-B in the early 1960s and has become the standard for reporting sales.2021-12-16How do time series adjust sales ? A sellin

7、g day adjustment factor for each month is computed and applied to the raw sales This factor allows you to compare months as if they had the same number of selling days e.g. accurately compare the June this year vs. June last year A seasonal factor is computed and applied to the selling day adjusted

8、sales This factor, when applied, gives you monthly data directly comparable to any other month e.g. accurately compare June this year with May this year2021-12-17Selling Days All other things being equal, sales in Aug-03 would decrease 4.8% because of one less selling day. In order to compare the tw

9、o months Aug-03 sales will have to be adjusted up +4.8%.SMTWTFSSMTWTFS121231.00.01.01.00.03456789456789100.01.01.01.01.01.00.00.01.01.01.01.01.00.010111213141516111213141516170.01.01.01.01.01.00.00.01.01.01.01.01.00.017181920212223181920212223240.01.01.01.01.01.00.00.01.01.01.01.01.00.02425262728293

10、0252627282930310.01.01.01.01.01.00.00.01.01.01.01.01.00.0310.0August 2003August 2002Aug-2003 has 21 selling daysAug-2002 has 22 selling days2021-12-18Seasonality Seasonality is expressed as an index for a month compared to an average month. A month where sales were 20% higher than average would have

11、 a seasonal factor of 120. A month which was 10% lower than average would have a seasonal factor of 90.JanFebMarApr MayJunNo Seasonality100100100100100100Strong Seasonality607580120140120Jul Aug SepOct Nov DecNo Seasonality100100100100100100Strong Seasonality118807562120150020406080100120140160JanFe

12、bMarAprMayJunJulAugSepOctNovDecStrong SeasonalityNo Seasonality2021-12-19Adjusting Sales Raw Sales X Selling Day Factor Seasonal FactorSeasonally Adjusted Sales = MonthActual Sales (M bbls)Selling Day FactorSeasonal FactorAdjusted SalesJun-03211X1.0041.210=175 Jul-03221X0.9581.212=175 Aug-03196X1.05

13、41.190=174 Sep-03160X1.0040.948=169 2021-12-110How do time series work?Raw SalesSelling Day AdjustedSeasonally Adjusted2021-12-111Dissecting a Time Series Plot 0200400600800100012001400160018002000199419951996199719981999Annualized Sales in M bblsAnnualized Sales tells us how big the market is.Trend

14、 Line tells us the direction of sales based on past & present performanceIrregular variations shows us the impact of market place actionsSTRs; Ontario STCsData Description tells us the type of data plotted2021-12-112Advantages of Time Series Advantages of time series: Removes variation in monthl

15、y sales caused by calendar differences and seasonality Help us to accurately estimate the direction and rate of sales growth/decline They are an improvement over other methods such as year-over-year growth or moving averages because they show us what is happening sooner an early warning of changing

16、sales conditions Time series significantly improve decision making Allows us to take corrective action sooner Allows us to take the right corrective action Helps to establish appropriate sales objectives2021-12-113Advantages of Time Series If the time series shows a relative smooth pattern from one

17、year to the next the trend and the year over year growth would provide roughly the same reading. But, if there was a significant market event or change, the year over year trends will be misleading051015202519981999+50%2021-12-114Misleading Growth Rates024681012141619981999Positive Trend:Flat % Chan

18、ge0%024681012141619981999Trend Flat; Positive% Change+21%2021-12-115More Misleading Growth Rates 024681012141619981999Trend Flat; Negative% Change-21%051015202519981999Trend Negative;Positive % Change+15%2021-12-116What are time series used for? At A-B we use time series to Assess current sales perf

19、ormance Develop current year sales projectionsPYE (projected year-end) Forecast next year sales develop budgets and monthly spreads Other quantitative sales analysis2021-12-117Assessing Sales Performance Beer Sales050100150200250199819992000200120022003Thousand BarrelsHow is our YTD performance?2021

20、-12-118Assessing Sales PerformanceBeer Sales05001,0001,5002,0002,500199819992000200120022003Annualized Sales in US bbls (in 1000s)Budget: 2,160M bblsPYE: 2,060M bbls%Change vs. Year AgoSep-03: +4.1%; SDA -0.9%YTD Sep-03: +1.2%; SDA +1.2%2021-12-119Beer Sales05001,0001,5002,0002,500199819992000200120

21、022003Annualized Sales in US bbls (in 1000s)Estimating PYE If there is no change in the business environment sales will continue on current trend. Points off the trend line can be used to estimate monthly sales.Underlying TrendPredicted2021-12-120Underlying Trend Underlying trend is a trend line tha

22、t best describes the current sales growth rate. It is the collective representation of all underlying factors that are influencing sales industry, competition, and company specific, etc. It is determined using a best-fit line to a set of points on the time series. The points are selected based on in

23、-depth understanding of the underlying factors influencing sales, how they have changed over time, and how they will likely change in the future. Points of inflection on the time series often signal changes in the underlying factors and hence the underlying trend.2021-12-121Beer Sales05001,0001,5002

24、,0002,500199819992000200120022003Annualized Sales in US bbls (in 1000s)Estimating PYEPredictedActual SalesPredictedSeasonally Adj. Sales Selling Day Factor X Seasonal FactorMonthly Sales = ActualTrend (AnnualizedSDAF SeasonalMonthPYE& Deseasonalized)EstimateJ120 120 F118 118 M138 138 A166 166 M1

25、95 195 J211 211 J221 221 A196 196 S160 160 O2,106 0.9590.932 12171 171 N2,110 1.1100.914 12145 145 D2,113 1.0041.239 12217 217 2,059 2003 PYE Estimate (M bbls)2021-12-122Beer Sales05001,0001,5002,0002,5001998199920002001200220032004Annualized Sales in US bbls (in 1000s)%Change vs. Year AgoSep-03: +4

26、.1%; SDA -0.9%YTD Sep-03: +1.2%; SDA +1.2%Establishing Budgets and Spreads Given our YTD Sep-2003 performance what would be an appropriate budget for next year and how should that volume be spread by month?2021-12-123Establishing Budgets and Spreads Beer Sales2,12305001,0001,5002,0002,50019981999200

27、02001200220032004Annualized Sales in US bbls (in 1000s)2003Trend (AnnualizedSDAF SeasonalMonth%SDA& Deseasonalized)Estimate vs. 2003J120 2,117 1.0040.675 12119 +3.4%F118 2,121 1.0540.754 12126 +6.8%M138 2,124 0.9170.852 12165 +10.0%A166 2,128 1.0540.995 12167 +0.7%M195 2,131 1.0541.082 12182 -1.

28、9%J211 2,135 0.9591.210 12225 +1.6%J221 2,138 1.0041.214 12215 +1.8%A196 2,142 1.0041.185 12211 +2.4%S160 2,146 1.0040.952 12170 +6.2%O171 2,149 1.0540.932 12158 +1.9%N145 2,153 1.0040.914 12163 +2.2%D217 2,156 1.0041.239 12222 +2.0%2,059 2,123 +3.1%2004 Budget (M bbls)2021-12-124Another Example Usi

29、ng Time Series0.06.412.819.225.632.038.444.851.257.664.0Price IncreaseAnnualized STRs in M BBLS199819992000200120022003Elasticity CalculationP1: 18.99; P2: 20.45 i.e. +7%V1: 44.5; V2: 38.5 i.e. -14%Elasticity = -2.0Estimating the price elasticityPrice = P1Volume = V1Price = P2Volume = V22021-12-125C

30、onclusions Time Series technique is a very useful sales analysis tool it is the standard for reporting and analyzing sales at A-B It is a powerful decision making tool for assessing sales performance, making accurate forecasts, and establishing appropriate budgets and spreads.2021-12-126Time Series

31、Thru September3,552.005001,0001,5002,0002,5003,0003,5004,0004,5001998199920002001200220032004Annualized Sales Metric Tons in Thousands% Change:+61.2%+20.9%Change vs. Year AgoSep-03: -2.5%; SDA -7.2%YTD Sep-03: +9.0%; SDA +9.0%2004 Forecast: 3,552.0M Met. Tons; +7.9% vs. 032003 PYE: 3,290.8M Met. Ton

32、s; +10.2% vs. 02Trend +8.2%Shipments2021-12-127Time Series Thru OctoberShipments3,471.805001,0001,5002,0002,5003,0003,5004,0004,5001998199920002001200220032004Annualized Sales Metric Tons in Thousands% Change:+61.2%+20.9%Change vs. Year AgoOct-03: -6.6%; SDA -6.6%YTD Oct-03: +7.9%; SDA +7.9%2004 For

33、ecast: 3,471.8M Met. Tons; +8.0% vs. 032003 PYE: 3,215.7M Met. Tons; +7.6% vs. 02Trend +6.7%2021-12-128Qingdao RegionShipments772.501002003004005006007008009001998199920002001200220032004Annualized Sales Metric Tons in Thousands% Change:+1.9%+17.8%Change vs. Year AgoOct-03: -15.7%; SDA -15.7%YTD Oct

34、-03: +2.5%; SDA +2.5%2004 Forecast: 772.5M Met. Tons; +8.3% vs. 032003 PYE: 713.5M Met. Tons; +2.0% vs. 02Trend +3.9%2021-12-129South RegionShipments630.201002003004005006007008001998199920002001200220032004Annualized Sales Metric Tons in Thousands% Change:+102.3%+47.8%Change vs. Year AgoOct-03: +6.

35、9%; SDA +6.9%YTD Oct-03: +1.9%; SDA +1.9%2004 Forecast: 630.2M Met. Tons; +3.9% vs. 032003 PYE: 606.7M Met. Tons; +2.4% vs. 02Trend +3.1%2021-12-130North RegionShipments741.601002003004005006007008009001998199920002001200220032004Annualized Sales Metric Tons in Thousands% Change:+174.3%+11.0%Change

36、vs. Year AgoOct-03: -5.7%; SDA -5.7%YTD Oct-03: +15.2%; SDA +15.2%2004 Forecast: 741.6M Met. Tons; +7.8% vs. 032003 PYE: 687.9M Met. Tons; +14.0% vs. 02Trend +8.8%2021-12-131Luzhong RegionShipments322.201002003004005006001998199920002001200220032004Annualized Sales Metric Tons in Thousands% Change:-

37、3.4%-9.2%Change vs. Year AgoOct-03: +10.1%; SDA +10.1%YTD Oct-03: +49.5%; SDA +49.5%2004 Forecast: 322.2M Met. Tons; +48.4% vs. 032003 PYE: 217.1M Met. Tons; +49.8% vs. 02Trend +42.9%2021-12-132Huaihai RegionShipments210.90501001502002503001998199920002001200220032004Annualized Sales Metric Tons in Thousands% Change:+82.3%+14.6%Change vs. Year AgoOct-03: -28.6%; SDA -28.6%YTD Oct-03: -8.8%; SDA -8.8%2004 Forecast: 210.9M Met. Tons; +2.4% vs. 032003 PYE: 206.0M Met. Tons; -8.4% vs. 02Trend -4.1%2021-12-133East RegionShi