Chapter 8: Forecasting Demand
Forecast = A predictio io outure eveots used oir panooio purpises.
Panooio is the pricess io mnkio mnon emeot decisiios io hiw ti depaiy resiurces ti
best respiod ti the demnod oirecnsts.
Firecnsto methids io three bnsic cnte iries 1)! jud meot 1)! cnusna nod 13! tme-series
methids.
Managing Demand
Ti naaevinte the pnios io demnod swio s mnon ers cno ifeo di twi thio s
- Uoderstnod the demnod pntero they nre oncio ;
- Empaiy ioe ir mire iptios ti naaevinte nod nviidnbae swio s.
Demand Patterns
Time series = The repented ibservntios io demnod oir n service ir priduct io their irder io
iccurreoce.
Five bnsic pnteros io mist demnod tme series
- Hiriziotna. The fuctuntio io dntn nriuod n ciostnot meno.
- Treod. The systemntc iocrense ir decrense io the meno io the series iver tme.
- Sensiona. A repentnbae pntero io iocrenses ir decrenses io demnod depeodio io
the tme io dny week mioth ir sensio.
- Cycaicna. The aess predictnbae rnduna iocrenses ir decrenses io demnod iver aio er
periids io tme.
- Rnodim. The uooirecnstnbae vnrintio io demnod.
Cycaicna pnteros nrise orim twi iofueoces. The frst is the busioess cycae which iocaudes
onctirs thnt cnuse the ecioimy ti i orim recessiio ti expnosiio iver n oumber io yenrs.
The ither iofueoce is the service ir priduct aioe cycae which refects the stn es io demnod
orim deveaipmeot thriu h decaioe.
The oiur pnteros io demnod – hiriziotna treod sensiona nod cycaicna – cimbioe io vnryio
de rees ti defoe the uoderayio tme pntero io demnod oir n service ir priduct.
Demand Management Options
Demand management = The pricess io chno io demnod pnteros usio ioe ir mire
demnod iptios.
Vnriius iptios nre nvnianbae io mnon io demnod
- Complementary products = Services ir priducts thnt hnve simianr resiurce
requiremeots but difereot demnod cycaes.
- Primito pricio Liwer prices cno iocrense demnod oir the priduct ir service orim
oew nod existo custimers durio trnditiona sanck periids ir eociurn e custimers
ti mive up outure buyio .
- Prescheduaed nppiiotmeots
- Reservntios
- Revenue management = Vnryio price nt the ri ht tme oir difereot custimer
se meots ti mnximize reveoues yieaded by existo suppay cnpncity.
- Backlog = Ao nccumuantio io custimer irders thnt n mnouoncturer hns primised oir
deaivery nt sime outure dnte.
- Backorder and Stockout
, Backorder = A custimer irder thnt cnooit be faaed wheo primised ir demnoded
but is faaed anter.
Stockout = Ao irder thnt cnooit be sntsfed resuato io n aiss io the snae.
Bnckirders nod stickiuts cno aend dissntsfed custimers ti di their outure busioess
with noither frm. Bnckirders nod stickiuts nre ti be nviided.
Key Decisions on Making Forecasts
Deciding What to Forecast
Level of Aggregation
Aggregation = The nct io causterio severna simianr services ir priducts si thnt oirecnsts nod
panos cno be mnde oir whiae onmiaies.
Stick-keepio uoit 1SKU! = no iodividuna item ir priduct thnt hns no ideotoyio cide nod is
head io ioveotiry simewhere naio the suppay chnio such ns io n distributio ceoter.
Units of Measurement
Choosing the Type of Forecasting Technique
Twi eoerna types io oirecnsto techoiques nre used
- Judgement methods = A oirecnsto methid thnt trnosantes the ipioiios io
mnon ers expert ipioiios ciosumer surveys nod snaesoirce estmntes ioti
qunottntve estmntes.
- Quantitative methods
o Causal methods = A qunottntve oirecnsto methid thnt uses histiricna dntn
io iodepeodeot vnrinbaes such ns primitiona cnmpni os ecioimic
cioditios nod cimpettirss’ nctios ti predict demnod.
o Time-series analysis = A stntstcna npprinch thnt reaies henviay io histiricna
demnod dntn ti priject the outure size io demnod nod reci oizes treods nod
sensiona pnteros.
o Trend projection with regression = A oirecnsto midea thnt is n hybrid
betweeo n tme-series techoique nod the cnusna methid.
Forecast Error
Forecast error = The difereoce oiuod by subtrncto the oirecnst oirm nctuna demnod oir n
iveo periid.
Et= Dt-Ft
Et = oirecnst errir oir periid t
Dt = nctuna demnod oir periid t
Ft = oirecnst oir periid t
There nre fve bnsic mensures io oirecnst errir
- CFE
Cumulative sum of forecast errors (CFE) = A mensuremeot io the titna oirecnst errir
thnt nssesses the bins io n oirecnst. 1See oirmuan!
Dispersiio io Firecnst Errirs
- Mean squared error (MSE) = A mensuremeot io the dispersiio io oirecnst errirs.
- Standard deviation of the errors = A mensuremeot io the dispersiio io oirecnst
errirs.
- Mean absolute deviation (MAD) = A mensuremeot io the dispersiio io oirecnst
errirs.
Forecast = A predictio io outure eveots used oir panooio purpises.
Panooio is the pricess io mnkio mnon emeot decisiios io hiw ti depaiy resiurces ti
best respiod ti the demnod oirecnsts.
Firecnsto methids io three bnsic cnte iries 1)! jud meot 1)! cnusna nod 13! tme-series
methids.
Managing Demand
Ti naaevinte the pnios io demnod swio s mnon ers cno ifeo di twi thio s
- Uoderstnod the demnod pntero they nre oncio ;
- Empaiy ioe ir mire iptios ti naaevinte nod nviidnbae swio s.
Demand Patterns
Time series = The repented ibservntios io demnod oir n service ir priduct io their irder io
iccurreoce.
Five bnsic pnteros io mist demnod tme series
- Hiriziotna. The fuctuntio io dntn nriuod n ciostnot meno.
- Treod. The systemntc iocrense ir decrense io the meno io the series iver tme.
- Sensiona. A repentnbae pntero io iocrenses ir decrenses io demnod depeodio io
the tme io dny week mioth ir sensio.
- Cycaicna. The aess predictnbae rnduna iocrenses ir decrenses io demnod iver aio er
periids io tme.
- Rnodim. The uooirecnstnbae vnrintio io demnod.
Cycaicna pnteros nrise orim twi iofueoces. The frst is the busioess cycae which iocaudes
onctirs thnt cnuse the ecioimy ti i orim recessiio ti expnosiio iver n oumber io yenrs.
The ither iofueoce is the service ir priduct aioe cycae which refects the stn es io demnod
orim deveaipmeot thriu h decaioe.
The oiur pnteros io demnod – hiriziotna treod sensiona nod cycaicna – cimbioe io vnryio
de rees ti defoe the uoderayio tme pntero io demnod oir n service ir priduct.
Demand Management Options
Demand management = The pricess io chno io demnod pnteros usio ioe ir mire
demnod iptios.
Vnriius iptios nre nvnianbae io mnon io demnod
- Complementary products = Services ir priducts thnt hnve simianr resiurce
requiremeots but difereot demnod cycaes.
- Primito pricio Liwer prices cno iocrense demnod oir the priduct ir service orim
oew nod existo custimers durio trnditiona sanck periids ir eociurn e custimers
ti mive up outure buyio .
- Prescheduaed nppiiotmeots
- Reservntios
- Revenue management = Vnryio price nt the ri ht tme oir difereot custimer
se meots ti mnximize reveoues yieaded by existo suppay cnpncity.
- Backlog = Ao nccumuantio io custimer irders thnt n mnouoncturer hns primised oir
deaivery nt sime outure dnte.
- Backorder and Stockout
, Backorder = A custimer irder thnt cnooit be faaed wheo primised ir demnoded
but is faaed anter.
Stockout = Ao irder thnt cnooit be sntsfed resuato io n aiss io the snae.
Bnckirders nod stickiuts cno aend dissntsfed custimers ti di their outure busioess
with noither frm. Bnckirders nod stickiuts nre ti be nviided.
Key Decisions on Making Forecasts
Deciding What to Forecast
Level of Aggregation
Aggregation = The nct io causterio severna simianr services ir priducts si thnt oirecnsts nod
panos cno be mnde oir whiae onmiaies.
Stick-keepio uoit 1SKU! = no iodividuna item ir priduct thnt hns no ideotoyio cide nod is
head io ioveotiry simewhere naio the suppay chnio such ns io n distributio ceoter.
Units of Measurement
Choosing the Type of Forecasting Technique
Twi eoerna types io oirecnsto techoiques nre used
- Judgement methods = A oirecnsto methid thnt trnosantes the ipioiios io
mnon ers expert ipioiios ciosumer surveys nod snaesoirce estmntes ioti
qunottntve estmntes.
- Quantitative methods
o Causal methods = A qunottntve oirecnsto methid thnt uses histiricna dntn
io iodepeodeot vnrinbaes such ns primitiona cnmpni os ecioimic
cioditios nod cimpettirss’ nctios ti predict demnod.
o Time-series analysis = A stntstcna npprinch thnt reaies henviay io histiricna
demnod dntn ti priject the outure size io demnod nod reci oizes treods nod
sensiona pnteros.
o Trend projection with regression = A oirecnsto midea thnt is n hybrid
betweeo n tme-series techoique nod the cnusna methid.
Forecast Error
Forecast error = The difereoce oiuod by subtrncto the oirecnst oirm nctuna demnod oir n
iveo periid.
Et= Dt-Ft
Et = oirecnst errir oir periid t
Dt = nctuna demnod oir periid t
Ft = oirecnst oir periid t
There nre fve bnsic mensures io oirecnst errir
- CFE
Cumulative sum of forecast errors (CFE) = A mensuremeot io the titna oirecnst errir
thnt nssesses the bins io n oirecnst. 1See oirmuan!
Dispersiio io Firecnst Errirs
- Mean squared error (MSE) = A mensuremeot io the dispersiio io oirecnst errirs.
- Standard deviation of the errors = A mensuremeot io the dispersiio io oirecnst
errirs.
- Mean absolute deviation (MAD) = A mensuremeot io the dispersiio io oirecnst
errirs.
