Experience curves as a planning tool. Part 2. Pyrrhic victories and practical problems

Robinson, Ced ic G. A icle Expe ience cu es as a planning ool. Pa 2. Py hic ic o ies and p ac ical p oblems Sou h A ican Jou nal o Business Managemen P o ided in Coope a ion wi h: Uni e si y o S ellenbosch Business School (USB), Bell ille, Sou h A ica Sugges ed Ci a ion: Robinson, Ced ic G. (1982) : Expe ience cu es as a planning ool. Pa 2. Py hic ic o ies and p ac ical p oblems, Sou h A ican Jou nal o Business Managemen , ISSN 2078-5976, A ican Online Scien i ic In o ma ion Sys ems (AOSIS), Cape Town, Vol. 13, Iss. 4, pp. 159-168, h ps://doi.o g/10.4102/sajbm. 13i4.1193 This Ve sion is a ailable a : h ps://hdl.handle.ne /10419/217811 S anda d-Nu zungsbedingungen: Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen Zwecken und zum P i a geb auch gespeiche und kopie we den. Sie dü en die Dokumen e nich ü ö en liche ode komme zielle Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich machen, e eiben ode ande wei ig nu zen. So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen (insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en, gel en abweichend on diesen Nu zungsbedingungen die in de do genann en Lizenz gewäh en Nu zungs ech e. Te ms o use: Documen s in EconS o may be sa ed and copied o you pe sonal and schola ly pu poses. You a e no o copy documen s o public o comme cial pu poses, o exhibi he documen s publicly, o make hem publicly a ailable on he in e ne , o o dis ibu e o o he wise use he documen s in public. I he documen s ha e been made a ailable unde an Open Con en Licence (especially C ea i e Commons Licences), you may exe cise u he usage igh s as speci ied in he indica ed licence. h ps://c ea i ecommons.o g/licenses/by/4.0/ Expe ience cu es as a planning ool. 1 Pa 2. Py hic ic o ies and p ac ical p oblems Ced ic G an Robinson School o Business Leade ship, Uni e si y o Sou h A ica S a egies based on he expe ience cu e e ec equi e ha p eda o y p ice cu ing leads o enhanced ma ke sha e and an inc easingly compe i i e cos s uc u e is-a- is he com- pe i ion. The company ha ides he expe ience cu e o he bank has cos ad an ages, p icing disc e ion and eaps hand- some p o i s. Co po a e g a eya ds a e li e ed wi h he co pses o companies who ha e adop ed oo nai e an app oach o, and oo simplis ic an accep ance o , he concep s i s pionee ed by he Bos on Consul ing G oup. This, he second o wo a icles on he expe ience cu e, highligh s py hic ic o ies in he ques o ma ke sha e and asks six key ques ions ha need o be answe ed be o e adop ing an expe ience-d i en s a egy. Likewise 12 p oblem a eas in he applica ion o his app oach a e iden i ied. S. A . J. Bus. Mgm . 1982, 13: 159-168 'n Voo e eis e i s a egiee gebasee op die e a ingsku we- e ek is da oo sug ige p yssnye y wel lei o e g o e ma kaandeel en 'n oenemend mee mededingende kos es uk uu eenoo die an die mededinge s. Die maa - skappy wa olle oo deel ui die e a ingsku we ek, he kos e oo dele, disk esie oo p ys ass elling en maak in- d ukwekkende wins e. Die 'beg aa plase' an die sakewl! eld he in onlangse ja e heelwa g a e bygek y an maa skappye wa onde gegaan he weens 'n e na'iewe benade ing eenoo , en 'n e simplis iese aan aa ding an die konsep e wa deu die Bos on Consul ing G oup oo ges el is. Hie die a ikel, die weede an wee a ikels oa die e a ingsku we, es ig die aandag op duu gekoop e skynoo winnings in die s yd om ma kaandeel en s el ses sleu el ae wa bean woo d moe wo d oo da 'n maa skappy hul s a egie op die e a- ingsku we basee . Twaal p obleema eas in die oepassing an hie die benade ing wo d ook ge'iden i isee . S.·A . Tydsk . Bed y sl. 1982, 13: 159-168 Second in a se ies o wo a icles C.G, Robinson School o Business Leade ship, Uni e si y o Sou h A ica, P.O. Box 392, P e o ia 0001, Republic o Sou h A ica Recei ed Ap il 1982; accep ed May 1982 S a egies based on he expe ience cu e e ec a e immensely appealing o manage s. The key o se ing s a egy is s a egic ma ke segmen a ion ollowed by p eda o y p ice cu ing o c ea e ba ie s o en y, o gain ma ke sha e and he eby o dec ease cos s. Lowe cos s gene a e enhanced ma gins and keep compe i o s a bay. The company ha ides he expe ience cu e o he bank domina es he chosen ma ke segmen , has cos ad an ages and p icing disc e ion and eaps handsome p o i s, pa icula ly on indus y ma u i y. The e is e idence ha a simplis ic accep ance o he ex- pe ience cu e as a s a egic ool wi hou conside a ion o a numbe o key ques ions may be suicidal. This does no nega e he use o he expe ience cu e, bu a he cau ions agains he nai e accep ance o expe ience cu e s a egies. Conside some o he e idence. Py hic ic o ies in he ques o ma ke sha e In 1972 Texas Ins umen s cu he p ice o calcula o s by 42"7o. They con e ed he slide ule business in o he silicon chip business and gained o e 65D/o o he ma ke by 1978. In 1974 Black and Decke cu hei p ices by 361170 in all domes ic ma ke s o he mains eam chain saw business. This ga e hem a ma ke sha e jump o 20 poin s and a com- manding lead in he US ma ke .2 In 1975 Black and Decke 's majo compe i o s cu hei p ices 280/o and began a long ha d climb owa ds losing hei shi s agains Black and Decke . In 1976 a well-known $3 billion u no e pape p o- duc s company cu p ices o gain ma ke sha e. Fou yea s la e hey we e s ill engaged in a uinous p ice wa which was e mina ed by he 1979 - 80 ecession and hei demise. 2 A li le sel es ain could be an admi able hing. The e a e imes when managemen should pu he ceiling o ealism on i s ambi ions when a h us o a la ge piece o he ac- ion could spell disas e . Al hough ROI is closely ela ed o ma ke sha e,3 com- panies some imes launch campaigns o building ma ke sha e wi hou g ea o esigh . Six key ques ions appea o be c i ical: Does he expe ience cu e apply? Does he company ha e he inancial esou ces?4 Will he company's s a egic posi ion emain iable i hwa ed?4 Will egula o y au ho i ies pe mi he company o n,-hi .. .. i ,;: ohiec i e?4 160 Do ade and a i ba ie s con use he pic u e? A e he s a egic cos implica ions changing? The expe iences o a numbe o companies, in sepa a e indus ies, ha ough disas ously o inc ease hei ma ke sha es illus a es he poin .4 Does he expe ience cu e apply? The expe ience e ec is no a law o na u e. I is mo e akin o a s a is ical law o any gi en company in an indus y. I exis s, i i exis s, o can be made o occu . This sounds a no e o wa ning. Sla ish belie in he exis ence o an ex- pe ience e ec can be ill- ounded.5 Conside he e ec o a p ice cu which is no ollowed by a gain in ma ke sha e o , i a gain occu s, no cos bene i s a e ealized by he added olumes. The e ec , i demand elas ici y is less han one, o i wo king capi al equi emen s a e g ea , is a d op in ma gin accompanied by a subs an ial d op in p o i abili y o no eal long- e m gain. The e ec can be o d ain p o i s o no long- e m s a egic bene i . Be o e emba king on p eda o y p ice cu - ing in ques s o ma ke sha e, i is wise o analyse indus y and co po a e cos pa e ns o see i an expe ience cu e e - ec does in ac apply. Smalle economies han ha o he Uni ed S a es, pa - icula ly hose o de eloping na ions, o en eso o o ms o a i p o ec ion o assis he de elopmen o ledgling indus ies. No de ini i e wo k appea s o ha e been done on companies wi hin such economies, bu i seems p uden no o bli hely assume he exis ence o expe ience e ec s unless hey can be p o ed o exis . This aises a seconda y p oblem o hose in cha ge o public policy o a de eloping na ion. I an expe ience e - ec does exis ou side o he bounda ies a o ded by p o- ec ion, he p o ec ion does no o e su icien suppo o local companies in he e en o he a i ba ie being d op- ped. The local company canno compe e in e na ionally because: cos s will no ha e been agg essi ely managed downwa ds due o he p o ec i e umb ella, o he small size o local e hnic ma ke s may ha e p eclud- ed he abili y o he local company o gain by d op- ping ma ginal cos due o low cumula i e olumes. Does he company ha e he inancial esou ces? Du ing Sep embe 1970 Gene al Elec ic's s udies indica ed ha i equi ed a 150/o sha e o he main ame compu e ma ke i he company we e o become compe i i e in he indus y.2•6 A abou he same ime, RCA concluded ha i needed a IOOJo sha e o he main ame compu e ma ke and commi ed i sel , publicly, o mee ing ha goal by he mid 1970s.2•7 Modes as hey seemed when ela ed o indus y s andings (IBM 690/o o he ma ke , GE 40/o, RCA 3,2,0Jo) hese a ge s ep esen ed mo e han a 2000/o g ow h in sha e in a high g ow h ma ke .2 Bo h companies decided, in he ligh o po en ial an i- us egula ions, ha expansion would be by means o in e nal g ow h a he han ha o acquisi ion. 8 F uhan 4 poin s ou ha in his indus y ope a ions equi e la ge amoun s o wo king capi al as compu e ma ke ing is by way o leased equipmen . In 1969 IBM equi ed $5,9 S.-A . Tydsk . Bed y sl. 1982, 13 (4) billion o suppo shipmen s o $4,95 billion sugges ing abou $1,20 o capi al o each $1,00 o shipmen s. Fo RCA, his mean ha o g ow ma ke sha e om 3,20/o o IO% o e i e yea s equi ed $1 334 million in capi al wi h yea ly amoun s o new capi al ising om $95 million o $475 million. Simila ly GE would ha e equi ed $2 076 million ising om $138 million o $785 million in he inal s ages.4 Wi h RCA and GE's espec i e co po a e p o i s being $151 million and $357 million in 1969, wi h di idend payou s o $68 million and $235 million, and deb equi y a ios o 0,45 and 0,27, he o al capi al gene a ion po en ial o each company was only $120 million and $135 million a e le e - ing e ained equi y up wi h deb . In addi ion, in 1%9 he compu e di ision e enues accoun ed o less han IO% o he o al co po a e e enues in bo h companies. The implica ions we e ha ei he he deb equi y a ios, al eady high by IBM s anda ds, would ha e o be aised sha ply, o equi y should be ob ained by bo h companies h ough secu i ies ma ke s a lowe p ice ea nings a ios han IBM be o e ei he company could come close o achie ing hei g ow h objec i es. The ma ke sha e aspi a ions e- qui ed capi al commi men s beyond he co po a e capi al gene a ing abili y o ei he company. Bo h GE and RCA acknowledged in 1970 and 1971 ha nei he could ma shal he esou ces o achie e a ma ginal ma ke sha e in he compu e indus y a an accep able cos . They qui he a ena. Be o e emba king on a ma ke sha e inc easing s a egy i is p uden o calcula e explici ly he cash low and capi al equi emen s equi ed o gain and hold ma ke sha e. A se o p o o ma p ojec ions o he en isaged ime ho izon will indica e whe he he s a egic a ge is easible o e en desi able. Abell and Hammond 9 ·P· 116 poin ou ha o g ow sha e om 60/o in he ace o an indus y leade wi h 240Jo in an indus y g owing a 80/o pe annum equi es a com- pound g ow h a e o 260/o pe annum o nine yea s o ca ch up o he leade . I implies g ow h a h ee imes indus y le els and a sales and capaci y expansion o 6400/o. The impo an ques ions a e: Wha a e he equi emen s o cash low and capi al unding o bo h ixed and wo king capi al? Wha a e he explici mechanics o gaining ma ke sha e? The expe ience model as a ame model 10 ad- oca es gains in ma ke sha e, bu o e s li le ad ice on how o ope a ionalize g ow h e o . The ollowing ques ions mus be answe ed: A whose expense will sha e be gained? Wha a e he likely s a egic compe i i e eac ions? Will he e be a clea cu ad an age associa ed wi h ma ke dominance se e al yea s down he line? Will he company's s a egic posi ion emain iable i hwa ed? The issue he e is i he company is hwa ed en ou e in a ques o ma ke sha e, he s a egic posi ion o he com- pany may become un enable. In Food Re ailing he US Fede al T ade Commission ound a high co ela ion be- ween he p o i con ibu ion o chain s o es in a gi en geog aphic a ea and he ma ke sha e in he same a ea. This means ha he p o i abili y o a chain o g oce y e ail s o es s. A . J. Bus. Mgm . 1982, 13(4) depends no on he na ional ma ke sha e, bu a he on he weigh ed a e age o i s ela i e ma ke sha es in a ious ci y ma ke a eas in which i pa icipa es, as can be seen in Figu e 1. 8,0 - ai 6,9 ii 7,0 en ~ - 0 6,0 5,9 6,0 - - 0 ii. 4,9 Q) 5,0 iii 0 - 4,3 a. - 0 u 4,0 B C .Q 45 3,0 :g c ~ 2,1 0 - u 2,0 Q) .. CII Ill cii > < 1,0 0,2 - 5% 10% 15% 20% 25% 4,9% o o o o o 35% 9,9% 14,9% 19,9% 24,9% 34,9% Ma ke Sha e (Pe cen ) Figu e 1 Ma ke sha e e sus s o e con ibu ion o co po a e p o i by g oups o ci ies4 Du ing 1948 - 58 many US chains mo ed owa d consoli- da ion. The mos ac i e acqui e , Na ional Tea, e ed by going o na ional geog aphic ma ke sha e a he han o sha e in speci ic me opoli an ma ke s. As he las o he independen s we e being me ged in o he compe i i e chains, he Fede al T ade Commission was aking decisi e ac ion o s op consolida ion by blocking u u e indus y me ge s. 11 As a consequence Na ional Tea could no y o dominance o ci y geog aphic segmen s and has a conside ably lowe ROI han Winn-Dixie who a e dominan in he Sou heas USA. This e oneous oehold s a egy illus a es bo h he need o pay ca e ul a en ion o s a egic ma ke segmen a- ion, and he need o ensu e ha an expe ience cu e eally does apply in he pa icula si ua ion. I also illus a es he dange in ge ing s opped hal -way in a ques o ma ke sha e. Will egula o y au ho i ies pe mi he company o achie e i s objec i e? Regula o y bodies may abo he ques o ma ke sha e o a ious easons, especially o an i-monopoly easons o o p e en s a egic dependence on a single majo supplie . F uhan poin s ou ha in making a el plans cus ome s ini ially con ac he ai ca ie which hey belie e has he mos daily ligh s o hei des ina ion.4 Thus, analogous o he alloca ion o shel space in a supe ma ke , high ligh equencies imply g ea e ma ke sha e and highe load ac- 161 o s pe ip and hence lowe cos s pe passenge mile and highe ma gins. The mino i y ca ie s should ope a e a a signi ican loss. The ai ca ie wi h su icien inancial esou ces o pu chase ex a ai c a and ly highe equen- cies on a pa icula in e ci y ou e should achie e he bene i s o expe ience e ec s. In p ac ice his does no always occu in he Uni ed S a es as he Ci il Ae onau ics Boa d (CAB) ends o in luence ma ke sha es and ela i e p o i abili y o US domes ic unk ai ca ie s. This is achie ed bo h by wi hholding o g an- ing licences o any in e ci y ma ke , and by egula ing he numbe o ca ie s compe ing on he ou e. Hence he CAB exe cised emendous powe o e he ela i e p o i abili y o he a ious ai lines and in ac he big ou , Ame ican, Eas e n, T ans Wo ld and Uni ed a e less p o i able han he se en smalle US ca ie s. This, howe e , does no seem o ha e blun ed he majo s' appe i e o waging expensi e capaci y wa s in a no-win en i onmen . I seems wise, p io o going all ou o ma ke sha e, o check whe he any legal, social, e hical o poli ical obs acles may lie ahead. Such obs acles may p e en he gaining o ma ke sha e, o may o ce companies o incu such cos s ha any expe ience bene i s a e nega ed. Do ade and a i ba ie s con use he pic u e? In economies less compe i i e han he Uni ed S a es, in- e en ion agains ques s o subs an ial ma ke sha e e- quen ly occu s. Me ge s o akeo e s can be blocked o an i- us easons o o p e en s a egic na ional dependence on a majo supplie , pa icula ly i he supplie is a sub- sidia y o a mul i-na ional. A sage piece o ad ice, o en gi en by s a o he S a egic Planning Ins i u e o clien s wi h ma ke sha es app oxi- ma ing 100%, is o cul i a e a small docile and appa en ly independen compe i o o suppo p ices. F equen ly he la ge dominan compe i o has o gi e away expe ience cu e gains so as no o suppo a p ice umb ella which in i es compe i i e en y o ac ion by he hos go e nmen . A e he s a egic cos implica ions changed? Al hough hey ha e achie ed ame mainly owing o hei wo k on he expe ience cu e e ec and hei g ow h sha e ma ix, discussed la e , he Bos on Consul ing G oup has also done excellen wo k in he a ea o s a egic cos ing associa ed wi h he expe ience e ec . This has no achie ed such widesp ead publici y. The expe ience e ec o any gi en company ope a es only on he company's con ibu- ion o he p oduc and is applicable o he alue-added com- ponen o o al cos . (Cos o goods sold minus cos o aw ma e ials, including ene gy, bough in.) I , o easons o simplici y, expe ience e ec s a e being analysed using sel- ling p ices, he expe ience cu e ope a es on he alue ad- ded componen in e ms o selling p ice minus cos o aw ma e ials. I includes ha alue added con ibu ion due o p o i , he alue added by he ma ke place. I , as ad oca ed by he Bos on Consul ing G oup, ma ginal p icing is used and no p ice umb ella is sus ained, he p ice and cos cu es ollow simila pa e ns. This means ha o a gi en p oduc a numbe o ex- . pe ience e ec s, o en sha ed wi h o he p oduc lines, 162 some imes un ela ed, can ope a e. This can ha e a p o ound e ec on he op imum s a egy o he mos inno a i e and success ul numbe o a complex p oduc chain. This has occu ed in he wa ch indus y. The classic Swiss wa ch was he handc a ed p oduc o a illage indus y. A e e ol ing h ough a numbe o p o o ypes, he clockwo k iming and mechanical sp ing mechanism we e in ecen yea s eplaced by solid s a e componen s and elec- ical s o age de ices. Ini ially qui e cos ly, and o ming a la ge pa o he cos o a wa ch, he semiconduc o com- ponen s ha e sped down an expe ience cu e ou s ipping he o he cos componen s o such an ex en ha semi- conduc o componen cos s a e no longe o majo impo - ance in he p oduc ion and ma ke ing o wa ches. This has led o Texas Ins umen s, a ho a ou i e in he s akes o wa ch indus y dominance because o hei compe ence in semiconduc o s, exi ing he indus y in he ealiza ion ha success will depend mo e on design and dis ibu ion han on expe ience e ec s applied o one o he mino , once ma- jo , cos componen s. Elec onic calcula o s, which in he space o a ew yea s ha e been ans o med om expensi e luxu ies o e e y-day i ems, p o ide a good example o he s a egic impac o changing cos pa e ns. The e en s in he calcula o ma ke illus a e he need o manage s o eally unde s and he dynamics o expe ience cu e cos educ ion and i s e ec on s a egy and hei abili y o emain compe i i e. Table I p esen s a simpli ied his o y o he de elopmen o solid s a e elec onic calcula o s up o 1975. 12 Table 1 His o y o solid s a e elec onic calcula o s up o 1975 Majo cos Dominan elemen s compe i o s Phase I Semiconduc o s Ame icans, ( disc e e de ices) e.g. Wang Phase 2 Assembly Japanese Sha p, Casio Phase 3 In eg a ed Ame icans, ci cui s e.g. Texas Ins umen s Phase 4 Assembly and Un esol ed dis ibu ion Tesco, Woolwo hs, Boo s In he ea ly days he majo cos componen s we e he disc e e semiconduc o de ices such as ansis o s and diodes. Leading compe i o s we e based in he USA close o he sou ce o solid s a e componen s and echnology. The calcula o s we e expensi e and he ma ke was limi ed o a ew applica ions o which hey we e cos e ec i e. Rapid ma ke g ow h occu ed. The ma ke o solid s a e diodes expanded a mo e han 500Jo pe annum du ing he ea ly and mid 1960s. Rapid expe ience-based cos and p ice declines occu ed. The p ice o an a e age ge manium diode ell by a ac o o 7 in eal e ms be ween 1960 and 1965. The main cos componen soon became ha o he labou equi ed o assemble he calcula o , and he compe i i e ad- an age passed o he Japanese who assumed i ual con- ol o he ma ke by he end o he decade. S.-A . Tydsk . Bed y sl. 1982, IJ( 4) Cos s con inued o decline and he ma ke became so la ge and a ac i e ha he Ame ican semiconduc o manu ac. u e s hemsel es en e ed he ma ke . Thei s a egy was o use la ge-scale in eg a ed ci cui s (LSI). The LSI chips made many assembly ope a ions edundan and Texas Ins umen s se ou o domina e he indus y. Calcula o s we e now a consume i em. The consume calcula o ma ke had expanded om only 2 million uni s in 1972 o 20 million uni s in 1974. P ices d opped d ama ically bo h as a esul o expe ience-based gains and a shakeou due o compe i ion o ma ke sha e. The shakeou is no ye comple e. The cos s uc u e is again changing wi h he in eg a ed ci cui s and displays becoming a smalle pa o he o al cos . Labou cos s a e again becoming impo an . This im- plies a e u n o compe i i e edge o low labou cos coun- ies, o o highly au oma ed cos -e ec i e mass p oduc ion in ad anced coun ies. The dis ibu ion cos s ha e become a signi ican cos componen now ha calcula o s a e so cheap. The leading compe i o s o he u u e may e y well be hose wi h he bes dis ibu ion channels and mass me - chandising ou le s. Wi h p ices declining mo e apidly han cos s many p o- duce s mus be squeezed, and Texas Ins umen s is p epa ed o ole a e hea y losses o e ain he ma ke . A equen e-analysis is essen ial in an expe ience cu e d i en indus y o see i he cos componen s upon which ea ly s a egies we e based a e s ill o s a egic signi icance. The e is a pa icula dange ha success ul candida es may be lulled in o a s a e o alse secu i y. P ac ical p oblems wi h applying he expe ience e ec Al hough he use o expe ience cu es is in ui i ely appeal- ing as a s a egic planning ool, a numbe o p oblems in applica ion can a ise. Some may be o e come easily, o he s a e no . Twel e basic p oblem a eas appea o exis : The ma ke de ini ion p oblem Analysing cos beha iou pa e ns The e ec o economic cycles Fo eign exchange a ia ions The ime span o analysis The choice o uni o analysis Sepa a ing ou cos componen s The e ec o sha ed expe ience Analysing cos componen s o compe i o s The issue o e ec i eness e sus e iciency P oblems in o ganiza ion P oblems in mo i a ion. Many o he issues a e in e linked o one ano he , o a e mani es a ions o unde lying p oblems in he s a egic p ocess. The ma ke de lni ion p oblem The p oblem o accu a ely de ining he ma ke is such an ubiqui ous one ha a ull book can be de o ed o he issues in ol ed. I mani es s i sel in he expe ience con ex e en hough he use o expe ience cu e s a egies does no e- qui e a de ini i e measu e o ma ke sha e o a gi en p oduc . s. A . J. Bus. Mgm . 1982, 13(4) The expe ience e ec elies on cumula i e p oduc ion olume o a p oduc o d i e cos s down he expe ience cu e. This depends on accu a ely de ining wha is a p o- duc and i s associa ed ma ke . Fo example, piping comes in di e en ma e ials o di e en sizes and p ope ies o di e en uses. n ying o analyse cos beha iou pa e ns, decisions ha e o be made as o wha cons i u es a gi en p o- duc in o de o alloca e o e heads (a p oblem discussed la e ), and as o wha cons i u es a gi en ma ke o analysis o .compe i o s and onesel in con ex . Technology ans e . be ween ela ed ma ke s gi es ise o sha ed e ec s wi h ega d o expe ience. Al hough he ma ke de ini ion p oblem is no made ex- . plici in se ing s a egy based on he expe ience e ec , i mus be esol ed soone o la e o enable he s a egis o assess he impac o sha ed expe ience, and o decide on he co ec alloca ion o co po a e o e head. Analysing cos beha iou pa e ns The·analysis o cos beha iou pa e ns is complica ed by ou issues: he e ec o in la ion, he ac ha cos s a e compa ed wi h expe ience, and no ime, he di icul y in ob aining accu a e in o ma ion, he join o e head cos alloca ion p oblem. The essence o he expe ience cu e is ha cos s exp ess- ed in eal e ms co ec ed o in la ion decline in a-p edi~- able way as expe ience is accumula ed. The e ec s o in- la ion a e el in se e al ways. Fi .s ly, inpu cos s, because o a i hme ic, a ec he alue added componen , pa icula ly i p ice, no p oduc cos , is being used as a su oga e a iable. Cos s do no mo e smoo hly and adjus men s; when hey occu , ake he o m o s ep changes, hus dis o - ing he cos pic u e. Secondly, because he expe ience cu e depends on eal cos s, he e is he p oblem o he choice o de la o . Should i be na ional, and i so which one; o should i be a sec- o al de la o ? The Bos on Consul ing G oup a e qui e de ini i e on he non-use o sec o al de la o s which hey claim may hemsel es ei he ha e an expe ience cu e e - ec ela ed o he indus y conce ned o , i he p oduc is a la ge inpu in o an indus y index, may ac ually implici ly inco po a e he expe ience e ec s unde analysis. On he choice o which na ional index, he e appea s no eal guidance o he han he use o he consume p ice index by de aul . Since indices a e man-made and subjec o e o s in concep ual design, i may be wise o ou inely see i any e ec s a e implici in he choice o any gi en de la o . The expe ience e ec is a ela ionship be ween cos s and cumula ed expe ience, no ime. This means ha manage s o ma u e p oduc s wi h la ge cumula i e p oduc ion may no obse e dec eases in cos o e sho ime spans e en a e in la ion has been compensa ed o . I annual olume g ows a a cons an a e, expe ience accumula es a a slowe and slowe a e wi h espec o ime. Fo example, o a new p oduc g owing a 200'/o pe annum, expe ience doubles in less han wo yea s om s a o p oduc ion (100 uni s in yea l, 120 in yea 2, o al 220). In jus 5 yea s 3 mon hs, cumula i e p oduc ion will ha e gone up by a ac o o eigh , 163 bu i will ake an addi ional 2 yea s 7 mon hs o double again. On a 700'/o expe ience cu e, a cos educ ion o 300'/o is achie ed be ween he end o yea one and yea wo, ye i akes om ea ly in he i h yea o close o he end o he se en h yea o achie e an equi alen 300 o educ ion. The ac ual collec ion o company cos in o ma ion is o en e y di icul o a numbe o easons. Fi s ly, cos s o a numbe o p oduc lines may ha e been collec ed on an ab- so p ion a h~ han a di ec cos ing basis. One needs o ind some means o e-alloca ing cos s ela ed o uni s p oduc- ed, a he han wi h ega d o di isional e o , such as ma ke ing o esea ch and de elopmen . Secondly, e en ho1,1gh a o m o di ec cos ing may ha e been used, alloca- ion o join o e heads is o en no made on a basis com- mensu a e wi h he ac ual p oduc ion ci cums ances, esul ing in oo la ge o oo small an alloca ion. As poin ed ou elsewhe e, he join o e head cos alloca ion p oblems is a di icul one which has no e en been add essed by many companies, pa icula ly se ice indus y companies. La ge gene al banks, o example, ake in e ail deposi s, lend ou money, unde w i e leases, and so on, and ha e no way o alloca ing cos s o s a egic ma ke segmen s. Ab up changes in accoun ing p ocedu es, aw ma e ial p ices, plan modi ica ion, sales ax changes, p oduc modi ica ion and o he c~uses can also dis o he cos da a. Rela ing he cos da a o he uni o p oduc ion, kilowa hou s o gene a ing capaci y o kilos o chicken p oduced, is ano he s umbling block as cos ing sys ems a e usually se up o cap u e disc e e uni s., sueh as gene a o s manu ac- u ed o chickens sold, a he han cos pe expe ience uni . The ca e ul un a elling o his o ic cos pa e ns can be a . di icul job. The e ec o economic cy<:les Mos businesses a e cyclical. Du ing a down u n olumes con ac , while hey expand in he upwa d phase. O e heads emain ela i ely ixed. As a esul , alloca ion o ixed o e head pe uni o p oduc ion a ies acco ding o he cycle and dis o s he cos da a. Changing demand his o ies and a ying loadings and o e heads need o be smoo hed using a long enough ime span o ob ain da a con- sis en wi h he long- un cos beha iou pa e ns o e ime. Fo eign exchange a ia ions This p oblem is becoming mo e impo an wi h he g ow h o mul ina ional co po a ions, he in e ela ionship be ween geog aphically dispe sed p oduce s, who may p o ide in- pu s o one ano he , and he eali y ha i is inc easingly s a egically nai e o conside indus ies such as au omobile manu ac u e on a geog aphic basis. The e ec s o bo h o eign exchange a ia ions and a ying na ional in la ion . a es can o ally ans o m di e en ela i e expe ience cu es. Thus Q a g aph such as Figu e 2, when cos s a e measu ed in a weake cu ency like s e ling, he slope o he cu e could e e se and a eal ela i e cos inc ease could ake place. 14 ,P· 113 The e is a mi iga ing e ec in ha economies subjec o high a es o in la ion a e also hose wi h poo exchange a e pe o mance and he wo e ec s can cancel one ano he ou o some ex; en . Howe e , no all cu encies a e allow- 164 in ~ - § 100 C 80 Ql C. 60 Ql () &: 40 Ql a <II a; > < 20 126 - 250cc 10 Cumula i e Volume (Million Uni s) Figu e 2 P ice expe ience cu e o he Japanese mo o cycle indus y (1959- )974) 14 ed o loa eely. They a e sho ed up, slide, glide, mo e cyclically, and some imes ab up ly de alue, hus changing he expe ience e ec when measu ed in ano he cu ency. Fo ma ke s whe e ela i e cos posi ions on an in e na- ional basis a e impo an bo h exchange a e luc ua ions, sui ably smoo hed, and in la ion need o be aken in o ac- coun when assessing ela i e cos ad an ages. The ime span o analysis As no ed, luc ua ing economic cycles a ec cos s, as do exchange a e a ia ions, plan modi ica ion, p oduc modi ica ion, wage a e changes and o he unlis ed a iables. As he expe ience e ec is a long- un phenomenon, a su i- cien ly long ime span mus be chosen o he analysis and sho - un da a mus be a e aged o sui ably smoo hed o educe he e ec s o luc ua ions. The choice o uni o analysis This is coupled wi h he ma ke de ini ion p oblem which has been discussed. The uni o analysis should ela e o he s a egic ma ke segmen which is being s udied. Fo exam- ple, i he s a egic conside a ions a e ocussed on a gi en de ined ma ke segmen , e o s can be made in choosing a uni which ela es o he o al ma ke a he han he speci ic segmen . I chosen oo b oadly, he uni neglec s he specialized cos ad an ages ha may acc ue o only a segmen o a o al ma ke due o i s unique equi emen s. The mi o image p oblem exis s i he uni is chosen oo na owly, pa icula ly i sha ed expe ience e ec s acc ue om common p oduc ion, esea ch and de elopmen , ma ke ing, componen s and so on. Too na ow a choice neglec s he impac o pe iphe al ac i i ies which may be sha ed. The uni should also be chosen o be meaning ul in e ms o he bene i s o he use associa ed wi h an indus y. Examples o well-de ined uni s a e kilog ams o chicken pe dolla , no chickens p oduced; kilowa s o elec ici y gene a ed pe dolla , no gene a o s manu ac u ed; and passenge kilome e s pe dolla , no ai lines sea bookings. I may be necessa y o de ine he uni in se e al ways and es he applicabili y and sensi i i y o he expe ience e ec o di e en uni s. Simila ly i may be necessa y in he case S.-A . Tydsk . Bed y sl. 1982, l3( 4) o sha ed expe ience e ec s, discussed below, o use di - e en uni s o analysis o di e en p oduc componen s in o de o co ec ly model he e ec s. This is especially ue when common ac i i ies o componen s a e used o a numbe o p oduc s. I may be necessa y o analyse cos s di e en ly o di e en unc ions such as manu ac u ing and ma ke ing, o e en communica ion cos s in he case o a se ice business such as a a el o ganiza ion. This de ail- ed analysis e eals he po en ial o expe ience cos gains o he a ious componen s o unc ions ha cons i u e he p oduc . The cos da a is no always linea due o mixed expe ience e ec s. I may be necessa y o disagg ega e cos s and hen, in he ligh o he analysis, o e-agg ega e in o la ge mo e meaning ul uni s o analysis. Di e en uni s o analysis may ha e o be used o di - e en compe i o s in di e en ma ke s. Dependen on he ma ke , o ac o s such as he compe i o s' le el o e ical in eg a ion, di e en uni s may need o be used o he com- pany and i s compe i o s, pa icula ly in he case o some subs i u e p oduc s. Al e na i ely i may some imes be necessa y o choose a uni o ensu e compa abili y be ween wo di e en p oduc s ha a e subs i u able o one ano he , such as s eel piping e sus plas ic piping. Sepa a ing ou cos componen s The cos o an indi idual p oduc educes acco ding o he expe ience cu e e ec . Cos s educe because o expe ience e ec s ac ing on he company's alue-added componen . Cos s also educe because o expe ience e ec s ope a ing on he inpu componen s. Each componen cos , including he company's own, may be d i en by di e en expe ience cu es ope a ing on di e en s a ing posi ions and cumula i e p oduc ion a es. Hence he o al cos , being he sum o he indi idual cos s, ope a es as ollows: To al cos = sum o all indi idual cos s n CT, E k=I (1) = whe e: CT, = he o al cos a ime c,. = he cos o componen k a ime n = he numbe o componen cos s being summed. Subs i u ing o c,.k om equa ion 11 ,P· 102 in e ms o p oduc ion olumes. CT, = 1 c,-;, (:,. ) -A (2) 1-i,k whe e: c,-i.k V ,k = he cos o componen k a ime - i = he cumula i e p oduc ion olume o componen k a ime V,-i,k = he cumula i e p oduc ion olume o componen k a ime -i = he expe ience cu e exponen o componen k. Each componen may ha e di e en cos s C,-i,kandd! - e en cumula i e olumes ;,k and ,-i.k associa ed wi h 1 ~· The olumes can cumula e e y di e en ly pa icula ly i componen k is used in a numbe o applica ions, no jus S. A . J. Bus. Mgm . 1982, 13(4) he p oduc unde analysis. Each componen may also be mo ing down di e en lea ning cu es wi h di e en slopes ela ed o di e en alues o A*. To al cos can be exp essed as n n ~V ) In( CT,) = *~ 1 In( C, _ u> - *~• A* In V ' * . -1,k (3) Because o accoun ing p ocedu es, and a i hme ic, he cos componen s a e addi i e. The successi e cos alues o ln(CT) will, howe e , no be a s aigh line. In gene al, i a non-linea expe ience cu e is encoun e ed, i may mean he exis ence o mo e han one expe ience cu e wi h he same slope, which is unlikely. Figu es 3 and 4 a e g aphic examples. In Figu e 3 a p o- duc uses h ee componen s in equal amoun s o make one uni o he p oduc . Fo simplici y, each is assumed o be on di e en expe ience cu es bu o ha e mo ed om cumula i e expe ience olumes V,-;,i, V,_;,i and V,_;,J wi h V,-;,i = V,-;,2 = V,-;,3 o a inalcumula i eexpe ience , 1 = ,. 2 = ,. The o al ajec o y is plo ed and he esul ing o al cos cu e is he simple sum o he h ee cos s. I is no linea as shown by he s aigh line which has been plo - ed angen ially o he o al cos cu e. (The cos is he linea addi i e alue o all h ee expe ience cu es. I can be seen ha ini ially he cu e o componen 1 will domina e and inally he cu e o componen 3 will domina e he cos pa e n.) Simila ly Figu e 4 ep esen s a p oduc consis ing o h ee componen s each on di e en expe ience cu es bu each cu e has he same slope. I i is assumed ha one uni o p oduc uses one uni o componen 1, wo uni s o com- ponen 2 and ou uni s o componen 3, i he e is no sha ed expe ience om o he p oduc s, he o al cos cu e has he same slope as he componen cu es. The cos cu e does, howe e , ha e a much mo e complex con igu a ion when he di e en componen s a e mo ing a di e en and un ela ed eloci ies down sepa a e cu es. (No e ha in his case all h ee componen cos s ha e been mo ing wi h con- s an eloci ies down he h ee cu es. Non-cons an eloci ies may gene a e a non-linea o al cos cu e.) Fi ing he componen cos da a will allow p edic ions o be accu a ely made as o componen cos s and, p o ided supplie s' u u e p oduc ion olumes can be es ima ed, should p o ide a be e ool o o ecas ing inpu compo- nen cos s. Mul iple expe ience cu es can ha e s a egic ele ance, as al eady poin ed ou . The expe ience e ec appa en ly slows wi h ega d o p oduc ion, and especially ime, as one o he cos componen s apidly diminishes in ela i e im- po ance as was he example wi h he solid s a e semi- conduc o indus y. The ela i e impo ance and change in impo ance be ween cos componen s can indica e he need o a e ised s a egy. The e ec o sha ed expe ience This is simila o he concep , discussed abo e, o compound expe ience cu es. Howe e , he s a egic ele ance is ha wo, o mo e, p oduc s may ha e common expe ience ei he bough in, o as a esul o common componen s, common p oduc ion lines, common unc ions such as ma ke ing and 165 Cumula i e Volume (Uni s) Figu e 3 To al cos and indi idual cos cu es o a h ee componen p oduc Componen 3 Cumula i e Volume (Uni s) Figu e 4 To al cos and indi idual cos o a h ee componen p oduc so on. The equa ions gi en a e alid. Howe e , we a e now in e es ed in he a ious cos pe o mances o se e al p o- duc s which u ilize common componen s. The ela ionship can be exp essed as whe e: CT,J c,-i.kJ V,,k n = E ( ~)-A* C,-;, J V . k=I 1-1,k = he o al cos a ime o p oduc j, = he componen cos o componen k used in p oduc j a ime - i, = he cumula i e olume o componen k p o- duc a ime . I includes he olumes o all o componen k p oduced o all he p oduc s using he componen , 166 = as abo e a ime - i, and = he expe ience componen o componen k's own expe ience cu e. The s a egic signi icance is ha , al hough i may be wise o ollow a gi en s a egy o one p oduc , he sha ed ex- pe ience e ec s and hei impac on o he p oduc s in he line may esul in o al co po a e subop imiza ion. I may, o example, pay o cu p ice on a gi en p oduc o d i e a second p oduc down an expe ience cu e o enhance p o- i s, p o ec a ulne able posi ion o ha es a ma ke . Con- e sely, he emo al o a p oduc om a ange may ha e an unexpec ed impac in he s a egic posi ioning o o he ela ed p oduc s. One o he simples ypes o sha ed expe ience is ha o common componen s, bu common unc ions, ma ke s, echnologies all impac he expe ience. The e ec i e use o sha ed expe ience is illus a ed by he emendous in e na- ional success o he Japanese in he mo o cycle indus y. Ini ially Japanese expo s we e aimed a he low cubic capaci y end o he ma ke . Ha ing achie ed success ul pene a ion and inally dominance o he low capaci y segmen s, he Japanese successi ely mo ed o la ge cubic capaci y segmen s, making e y e ec i e use o modula componen designs and sha ed expe ience on many com- ponen s. The e ec s on B i ish and Ge man manu ac u e s was auma ic. 14 Simila ly, no malized silicon de ice p ice ends 15 can be ela ed o compu ing cos s as shown in Figu e 516 using a no malized ime basis, a he han cumula i e p oduc ion. The conclusion can be eached ha he compu e indus y's spec acula g ow h has been due mainly o i s abili y o p o- duce equipmen which could compu e a e e inc easing speeds and eliabili y le els, and a e e dec easing cos s and sizes. All o hese a ibu es ha e s emmed om ad ances in silicon echnology. 16 S a egically, oppo uni ies o he use o sha ed ex- pe ience can be explo ed, analysed and exploi ed o gain he bene i s o sha ed expe ience. Sha ed expe ience can be a co e s eng h, i s lack a co e weakness. The s a egic ele ance o sha ed expe ience mus be analysed in decisions o add o elimina e p oduc s om he company po olio. Analysing cos componen s o compe i o s Analysing compe i o s' cos s is e en mo e di icul han analysing one's own cos s, owing o he lack o cos in o - ma ion on, and de ailed insigh in o, he compe i o s' business. Howe e , he a emp o do so in i sel gene a es mo e han jus cos in o ma ion. I gene a es a lo o in- sigh in o he compe i o s' way o unning hei businesses. In gene al, cos in o ma ion is no eadily a ailable and es ima es may ha e o be made on he basis o known sel- ling p ices, which may, o may no be, adjus ed o he ex- pec ed ma gins. Fi ms can make inco ec assessmen s o compe i o s' cos s based on an o e simpli ied idea ha a single expe ience cu e exis s down which e e yone is mo ing a a a e ela ed o hei ma ke sha e. This o e simpli ica ion can be e - oneous, and companies may ha e di e en cu es, s a - ing posi ions, eloci ies; o hey may e en play leap og S.-Al . Tydsk . Bed y sl. 1982, 13{4) - - - Compu ing Cos s ---- Silicon De ice P ice T ends 'ii, 10 ~ 0 e. Q) (.) '& OJ C: GI en GI OJ Ill Q) ~ 1,0 L....~.L-~.L-~L....---Ji..........1..____.'-- ...... ~--'100K 0 2 4 6 8 10 12 14 Yea s (No malised) Figu e S Silicon de ice p ice ends ela ed o compu ing cos s • lK -bi Dynamic RAM X A e age IC <> Di used silicon ansis o C High pe o mance compu e cos s down he cu e o a ious easons:9'17 Di e en companies may ha e sub ly di e en ma kelS, ano he mani es a ion o he ma ke de ini ion p o- blem, o di e en ma e ials o cons uc ion such as plas ic e sus s eel pipes. Di e en in e na ional economic cycles and o eign ex- change a e a ia ions will cloud he pic u e. Cos componen s, because o di e en ma e ials o cons uc ion, design di e ences o p oduc ion and o he unc ional ac i i ies, may a y in beha iou owing o he ac ha he compe ing p oduc s may no be iden ical. Common supplie s lead o a sha ing o expe ience wi h compe i o s due o he supplie s' expe ience cu e be- ing a d i ing e ec . Access o echnology om componen manu ac u e s o e en p oduc ion equipmen manu ac u e s and ou le channels educes he compe i i e edge o he dominan p oduce . Pa en and o he p o ec ion, and con ol o e aw ma e ial sou cing, a ec cos s. . Published and o he sou ces o knowledge, including ques ionable p ac ices such as espionage, na ow he expe ience gap.