Collaborate with Your Competitorsand Win

Gary Hammel, Yves L. Doz, and C. K. Prahalad

Collaboration between competitors is in fashion. General Motors and Toyota assemble automobiles, Siemens and Philips develop semiconductors, Canon supplies photocopiers to Kodak, rance!s Thomson and "apan!s "#C manufacture videocassette recorders. $ut the spread of what we call %competitive collaboration&'(oint ventures, outsourcin) a)reements, product licensin)s, cooperative research'has tri))ered unease about the lon)*term conse+uences. , strate)ic alliance an stren)then both companies a)ainst outsiders even as it weakens one partner vis*-*vis the other. .n particular, alliances between ,sian companies and /estern rivals seem to work a)ainst the /estern partner. Cooperation becomes a low*cost route for new competitors to )ain technolo)y and market access. 0et the case for collaboration is stron)er than ever. .t takes so much money to develop new products and to penetrate new markets that few companies can )o it alone in very situation. .C1, the $ritish Computer Company, could not have developed its current )eneration of mainframes without u(itsu. Motorola needs Toshiba!s distribution capacity to break into the "apanese semiconductor market. Time is another critical factor. ,lliances can provide shortcuts for /estern companies racin) to improve their production efficiency and +uality control. /e have spent more than five years studyin) the inner workin)s of 23 strate)ic alliances and monitorin) scores of others. 4ur research 5see the insert %,bout 4ur 6esearch&7 involves cooperative ventures between competitors from the 8nited States and "apan, 9urope and "apan, and the 8nited States and 9urope. /e did not (ud)e the success of failure of each partnership by its lon)evity'a common mistake when evaluatin) strate)ic alliances'but by the shifts in competitive stren)th on each side. /e focused on how companies use competitive collaboration to enhance their internal skills and technolo)ies while they )uard a)ainst transferrin) competitive advanta)es to ambitious partners. There is no immutable law that strate)ic alliances must be a windfall for "apanese or Korean partners. Many /estern companies do )ive away more than they )ain'but that!s because they enter partnerships without knowin) what it takes to win. Companies that benefit most from competitive collaboration adhere to a set of simple but powerful principles. Collaboration is competition in a different form. Successful companies never for)et that their new partners may be out to disarm them. They enter alliances with clear strate)ic ob(ectives, and they also understand how their partners! ob(ectives will affect their success. Harmony is not the most important measure of success. .ndeed, occasional conflict may be the best evidence of mutually beneficial collaboration. ew alliances remain win*win undertakin)s forever. , partner may be content even as it unknowin)ly surrenders core skills. Cooperation has limits. Companies must defend against competitive compromise. , strate)ic alliance is a constantly evolvin) bar)ain whose real terms )o beyond the le)al a)reement or the aims of top mana)ement. /hat information )ets traded is determined

day to day, often by en)ineers and operatin) mana)ers. Successful companies inform employees at all levels about what skills and technolo)ies are off*limits to the partner and monitor what the partner re+uests and receives. Learning from partners is paramount. Successful companies view each alliance as a window on their partners! broad capabilities. They use the alliance to build skills in areas outside the formal a)reement and systematically diffuse new knowled)e throu)hout their or)ani:ations. Why collaborate? 8sin) an alliance with a competitor to ac+uire new technolo)ies or skills is not devious. .t reflects the commitment and capacity of each partner to absorb the skills of the other. /e found that in every case in which a "apanese company emer)ed from an alliance stron)er than its /estern partner, the "apanese company had made a )reater effort to learn. Strate)ic intent is an essential in)redient in the commitment and capacity of each partner to absorb the skills of the other. /e found that in every case in which a "apanese company emer)ed from an alliance stron)er than its /estern partner, the "apanese company had made a )reater effort to learn. Strate)ic intent is an essential in)redient in the commitment to learnin). The willin)ness of ,sian companies to enter alliances represents a chan)e in competitive tactics, not competitive )oals. ;9C, for e<ample, has used a series of collaborative ventures to enhance its technolo)y and product competences. ;9C is the only company in the world with a leadin) position in telecommunications, computers, and semiconductors'despite its investin) less in 6=> 5as a percenta)e of revenues7 than competitors like Te<as .nstruments, ;orthern Telecom, and 1.M. 9ricsson. .ts strin) of partnerships, most notably with ?oneywell, allowed ;9C to levera)e its in*house 6=> over the last tow decades. /estern companies, on the other hand, often enter alliances to avoid investments. They are interested in reducin) the costs and risks of enterin) new businesses or markets than to ac+uirin) new skills. , senior 8.S. mana)er offered this analysis of his company!s venture with a "apanese rival@ %/e complement each other well'our distribution capability and their manufacturin) skill. . see no reason to invest upstream if we can find a secure source of product. This is a comfortable relationship for us&. ,n e<ecutive from this company!s "apanese partner offered a different perspective@ %/hen it is necessary to collaborate, . )o to my employees and say, AThis is bad, . wish we had these sills ourselves. Collaboration is second best. $ut . will feel worse if after four years we do not know how to do what our partner knows how to do.! /e must di)est their skills.& The problem here is not that the 8.S. company wants to share investment risk 5its "apanese partner does too7 but that the 8.S. company has no ambition beyond avoidance. /hen the commitment to learnin) is so one*sided, collaboration invariably leads to competitive compromise. Many so*called alliances between /estern companies and their ,sian rivals are little more than sophisticated outsourcin) arran)ements 5see the bo< %Competition for Competence&7. General Motors buys cars and component from Korea!s >aewoo. Siemens buys computers from e(itsu. ,pple buys laser printer en)ines from Canon. The traffic is almost entirely one way. These 49M deals offer ,sian partners a way t capture investment initiative from /estern competitors and displace customer competitors from

value*creatin) activities. .n many cases this )oal meshes with that of the /estern partner@ to re)ain competitiveness +uickly and with minimum effort. Consider the (oint venture between 6over, the $ritish auto*maker, and ?onda. Some B3 years a)o, 6over!s forerunners were world leaders in small car desi)n. ?onda had not even entered the automobile business. $ut in the min*2CDEs, after failin) to penetrate forei)n markets, 6over turned to ?onda for technolo)y and product*development support. 6over has used the alliance to avoid investments to desi)n and build new cars. ?onda has cultivated skills in 9uropean stylin) and marketin) as well as multinational manufacturin). The is little doubt which company will emer)e stron)er over the lon) term. Troubled la))ards like 6over often strike alliances with sur)in) latecomers like ?onda. ?avin) fallen behind in a key skills are 5in this case, manufacturin) small cars7, the la))ard attempts to compensate for past failures. The latecomer uses the alliance to close a specific skills )ap 5in this case, learnin) to build cars for a re)ional market7. $ut a la))ard that for)es a partnership for short*term )ain may find itself in a dependency spiral@ as it contributes fewer and fewer distinctive skills, it must reveal more and more of its internal operations to keep the partner interested. or the weaker company, the issue shifts from %Should we collaborateF& to %/ith whom should we collaborateF& to %?ow do we keep our partner interested as we lose the advanta)es that made us attractive to them in the first placeF& There!s a certain parado< here. /hen both partners are e+ually intent on internali:in) the other!s skills, distrust and conflict may spoil the alliance and threaten its very survival. That!s one reason (oint ventures between Korean and "apanese companies have been few and tempestuous. ;either side wants to %open the kimono.& ,lliances seem to run most smoothly when one partner is intent on learnin) and the other is intent on avoidance'in essence, when one partner is willin) to )row dependent on the other. $ut runnin) smoothly is not the pointG the point is for a company to emer)e from an alliance more competitive than when it entered it. One partner does not always have to )ive up more than it )ains to ensure the survival of an alliance. There are certain conditions under which mutual )ain is possible, at least for a time. The partners strategic goals converge while their competitive goals diverge. That is, each partner allows for the other!s continued prosperity in the shared business. Phillips and >u Pont collaborate to develop and manufacture compact discs, but neither side invades the other!s market. There is a clear upstreamHdownstream division of effort. The size and market power of both partners is modes compared with industry leaders. This forces each side to accept that mutual dependence may have to continue for many years. 1on)*term collaboration may be so critical to both partners that neither will risk anta)oni:in) the other by an overtly competitive bid to appropriate skills or competence. u(itsu!s 2 to 3 disadvanta)e with .$M means it will be a lon) time, if ever, before u(itsu can break away from its forei)n partners and )o it alone. ach partner believes it can learn from the other and at the same time limit access to proprietary skill. "#C and Thomson, both of whom make #C6!s, know that they are tradin) skills. $ut the two companies are lookin) for very different thin)s. Thomson needs product technolo)y and manufacturin) prowessG "#C needs to learn how to succeed in the fra)mented 9uropean market. $oth sides believe there is and e+uitable chance for )ain.

How to B !ld "ec re De#enses or collaboration to succeed, each partner must contribute somethin) distinctive@ basic research, product development skills, manufacturin) capacity, access to distribution. The challen)e is to share enou)h skills to create advanta)e vis*-*vis companies outside the alliance while preventin) a whole sale transfer of core skills to the partner. This is a very thin line to walk. Companies must carefully select what skills and technolo)ies they pass to their partners. They must develop safe)uards a)ainst unintended, informal transfers of information. The )oal is to limit the transparency of their operations. The type of skill a company contributes is an important factor in how easily its partner can internali:e the skills. The potential for transfer is )reater when a partner!s contribution is easily transported 5in en)ineerin) drawin)s, on computer tapes, or in the head of a few technical e<perts7G easily interpreted 5it can be reduced to commonly understood e+uations or symbols7G and easily absorbed 5the skill or competence is independent of any particular cultural conte<t7. /estern companies face an inherent disadvanta)e because their skills are )enerally more vulnerable to transfer. The ma)net that attracts so many companies to alliances with ,sian competitors is their manufacturin) e<cellence'a competence that is less transferable than most. "ust*in*time inventory systems and +uality circles can be imitated, but this is like pullin) a few threads out of an oriental carpet. Manufacturin) e<cellence is a comple< web of employee trainin), inte)ration with suppliers, statistical process controls, employee involvement, value en)ineerin), and desi)n for manufacture. .t is difficult to e<tract such a subtle competence in any way but a piecemeal fashion. There is an important distinction between technolo)y and competence. , discrete, stand*alone technolo)y 5for e<ample, the desi)n of a semiconductor ship7 is more easily transferred than a process competence, which is entwined in the social fabric of a company. ,sian companies often earn more from their /estern partners than vice versa because they contribute difficult*to*unravel stren)ths, while /estern partners contribute easy*to*imitate technolo)y. So companies must take steps to limit transparency. 4ne approach is to limit the scope of the formal a)reement. .t mi)ht cover a sin)le technolo)y rather than an entire ran)e of technolo)iesG par of a product line rather than the entire lineG limited period of time. The ob(ective is to circumscribe a partner!s opportunities to learn. Moreover, a)reements, should establish specific performance re+uirements. Motorola, for e<ample, takes an incremental, incentive*based approach to technolo)y transfer in its venture with Toshiba. The a)reement calls for Motorola to release its microprocessor technolo)y incrementally as Toshiba delivers on its promise to increase Motorola!s penetration in the "apanese semiconductor market. The )reater Motorola!s market share, the )reater Toshiba!s access to Motorola!s market technolo)y. Many of the skills that mi)rate between companies are not covered in the formal terms of collaboration. Top mana)ement puts to)ether strate)ic alliances and set the le)al parameters for e<chan)e. $ut what actually )ets traded id determined by day*to*day interactions of en)ineers, marketers, and product developers@ who says what who to whom, who )ets access t what facilities, who sits on what (oint committees. The most important deals 5%.!ll share this with your if you share that with me&7 may be struck four or five or)ani:ational levels below where the deal was si)ned. ?ere lurks the )reatest risk of unintended transfers of important skills. Consider one technolo)y*sharin) alliance between 9uropean and "apanese competitors. The 9uropean Company valued the partnership as a way to ac+uire a

specific technolo)y. The "apanese company considered it a window on its partner!s marketin) and product*development staff. The company mined each contact for as much information as possible. or e<ample, every time the 9uropean Company re+uested a new feature on a product bein) sourced from its partner, the "apanese company asked for detailed customer and competitor analyses to (ustify the re+uest. 4ver time, it developed a sophisticated picture of the 9uropean partner throu)h the formal a)reement had a useful life of three to five years. The competitive insi)hts ac+uired informally by the "apanese company will probably endure lon)er. 1imitin) unintended transfers at the operatin) level re+uires careful attention to the role of )atekeepers, the people who control what information flows to a partner. , )atekeeper can be effective only if there are a limited number of )ateways throu)h which a partner can access people and facilities. u(itsu!s many partners all )o throu)h a sin)le office, the %collaboration section,& to re+uest information and assistance from different division. This way the company can monitor and control access to critical skills and technolo)ies. /e studied one partnership between 9uropean and 8.S. competitors that involved several divisions of each company. /hile the 8.S. Company could only access its partner throu)h a sin)le )ateway, its partner had unfettered access to all participatin) divisions. The 9uropean Company took advanta)e to provide certain information, the 9uropean partner make the same re+uest of another division. ;o sin)le mana)er in the 8.S. company could tell how much information had been transferred or was in a position to piece to)ether patterns in the re+uests. Colle)iality is a prere+uisite for collaborative success. $ut too much colle)iality should set off warnin) bells to senior mana)ers. C94s or divisional presidents should e<pect occasional complaints from their counterparts about the reluctance of lower level employees to share information. That!s a si)n that the )atekeepers are doin) their (obs. ,nd senior mana)ement should re)ularly debrief operatin) personnel to find out what information the partner is re+uestin) and what re+uests are bein) )ranted. 1imitin) unintended transfers ultimately depends on employee loyalty and self* discipline. This was a real issue for many of the /estern companies we studied. .n their e<citement and pride over technical achievements, en)ineerin) staffs sometimes shared information that top mana)ement considered sensitive. "apanese en)ineers were less likely to share proprietary information. There are a host of cultural and professional reasons for the relative openness of /estern technicians. "apanese en)ineers and scientists are more loyal to their company than to their profession. They are less steeped in the open )ive*and*take of university research since they receive much of their trainin) from employers. They consider their themselves team members more than individual scientific contributors. ,s one "apanese mana)er noted, %/e don!t feel any need to reveal what we know. .t is not an issue of pride for us. /e!re )lad to sit and listen. .f we!re patient we usually learn what we want to know.& Controllin) unintended transfers may re+uire restrictin) access to facilities as well as to people. Companies should declare sensitive laboratories and factories off*limits to their partners. $etter yet, they mi)ht house the collaborative venture in an entirely new facility. .$M is buildin) a special site in "apan where u(itsu can review its forthcomin) mainframe software before decidin) whether to license it. .$M will be able to control e<actly what u(itsu see and what information leaves the facility.

inally, which country serves as %home& to the alliance affects transparency. .f the collaborative team is located near one partner!s ma(or facilities, the other partner will have more opportunities to learn*but less control over what information )ets traded. /hen the partner houses, feeds, and looks after en)ineers and operatin) mana)ers, there is a dan)er they will %)o native&. 9<patriate personnel need fre+uent visits from head+uarters as well as re)ular furlou)hs home. $nhance the Ca%ac!ty to Learn /hether collaboration leads to competitive surrender or revitali:ation depends foremost on what employees believe the purpose of the alliance to be. .t is self*evident@ to learn, one must want to learn. /estern companies won!t reali:e the full benefits of competitive collaboration until they overcome arro)ance borne of decades of leadership. .n short, /estern companies must be more receptive. /e asked a senior e<ecutive in a "apanese electronics company about the perception that "apanese companies learn more from their forei)n partners than vice versa. %4ur /estern partners approach us with the attitude of teachers,& he told us. %/e are +uite happy with this, because we have the attitude of students. 1earnin) be)ins at the top. Senior mana)ement must be committed to enhancin) their companies! skills as well as to avoidin) financial risk. $ut most learnin) takes place at the lower levels of an alliance. 4peratin) employees not only represent the front lines in an effective defense but also play a vital role in ac+uirin) knowled)e. They must be well briefed on the partner!s stren)ths and weaknesses and understand how ac+uirin) particular skills will bolster their company!s competitive position. This is already standard practice amon) ,sian companies. /e accompanied a "apanese development en)ineer on a tour throu)h a partner!s factory. This en)ineer dutifully took notes on plant layout, the number of production sta)es, the rate at which the line was runnin), and the number of employees. ?e recorded all this despite the fact that he had no manufacturin) responsibility in this own company, and that the alliance didn!t encompass (oint manufacturin). Such dedication )reatly enhances learnin). Collaboration doesn!t always provide an opportunity to fully internali:e a partner!s skills. 0et (ust ac+uirin) new and more precise benchmarks of a partner!s performance can be of )reat value. , new benchmark can provoke a thorou)h review of internal performance levels and may spur a round of competitive innovation. ,skin) +uestions like, %/hy do their semiconductor lo)ic desi)ns have fewer errors than oursF& and %/hy are they investin) in this technolo)y and we!re notF& may provide the incentive for a vi)orous catch*up pro)ram. Competitive benchmarkin) is a tradition in most of the "apanese companies we studied. .t re+uires many of the same skills associated with competitor analysis@ systematically calibratin) performance a)ainst e<ternal tar)etsG learnin) to use rou)h estimates to determine where a competitor 5or partner7 is better, faster, or cheaperG translatin) those estimates into new internal tar)etsG and recalibratin) to establish the rate of improvement in a competitor!s performance. The )reat advanta)e of competitive collaboration is that pro<imity makes benchmarkin) easier. .ndeed, some analyst ar)ue that one of Toyota!s motivations in collaboratin) with GM in the much*publici:ed ;8MM. venture is to )au)e the +uality of GM!s manufacturin) technolo)y. GM!s top manufacturin) people )et a close look at Toyota, but the reverse is true as well. Toyota may be learnin) whether its )iant 8.S. competitor is capable of closin) the productivity )ap with "apan.

Competitive collaboration also provides a way of )ettin) close enou)h to rivals to predict how runs its course. ?ow does the partner respond to price chan)esF ?ow does it measure and reward e<ecutivesF ?ow does it prepare to launch a new productF $y revealin) a competitor!s mana)ement orthodo<ies, collaboration can increase the chances of success in future head*to*head battles. Knowled)e ac+uired from a competitor*partner in only valuable after it is diffused throu)h the established internal clearin)houses to collect and disseminate information. The collaboration mana)er at one "apanese company re)ularly made the rounds of all employees involved in alliances. ?e identified what information had been collected by whom and then passed it on to appropriate department. ,nother company held re)ular meetin)s where employees shared new knowled)e and determined who was best positioned to ac+uire additional information. Proceed w!th Care&B t Proceed ,fter /orld /ar .., "apanese and Korean companies entered alliances with /estern rivals from weak positions. $ut they worked steadfastly toward independence. .n the early 2CIE!s, ;9C!s computer business was one*+uarter the si:e of ?oneywell!s, its primary forei)n partner. .t took only two decades for ;9C to )row lar)er than ?oneywell, which eventually sold its computer operations to an alliance between ;9C and Group $ull of rance. The ;9C e<perience demonstrates that dependence on a forei)n partner doesn!t automatically condemn a company to also*ran status. Collaboration may sometimes be unavoidableG surrender is not. Mana)ers are too often obsessed with the ownership structure of an alliance. /hether a company controls 32J or KCJ of a (oint venture may be much less important than the rate at which each partner learns from the other. Companies that are confident of their ability to learn may even prefer some ambi)uity in the alliance!s le)al structure. ,mbi)uity creates more potential to ac+uire skills and technolo)ies. The challen)e for /estern companies is not to write ti)hter le)al a)reements but to become better learners. 6unnin) away from collaboration is no answer. 9ven the lar)est /estern companies can no lon)er outspend their )lobal rivals. /ith leadership in many industries shiftin) toward the 9ast, companies in the 8nited States and 9urope must become )ood borrowers'much like ,sian companies did in the 2CIE!s and 2CDE!s. Competitive renewal depends on buildin) new process capabilities and winnin) new product and technolo)y battles. Collaboration can be a low*cost strate)y for doin) both. 'bo t ( r )esearch
/e spent more than five years studyin) the internal workin)s of 23 strate)ic alliances around the world. /e sou)ht answers to a series f interrelated +uestions. /hat role have strate)ic alliances and outsourcin) a)reements played in the )lobal success of "apanese and Korean companiesF ?ow doe alliances chan)e the competitive balance between partnersF >oes winnin) at collaboration mean different thin)s to different companiesF /hat factors determine who )ains most from collaborationF To understand who won and who lost and why, we observed the interactions of the partners first*hand and a multiple levels in each or)ani:ation. 4ur sample included four 9uropean*8.S. alliances, two intra* 9uropean alliances, two 9uropean*"apanese alliances, and seven 8.S.*"apanese alliances. /e )ained access to both sides of the partnerships in about half the cases and studied each alliance for an avera)e of three years.

Confidentiality was a paramount concern. /here we did have access to both sides, we often wound up knowin) more about who was doin) what to whom than either of the partners. To preserve confidentiality, our article dis)uises many of the alliances that were part of the study.

Com%et!t!on #or Com%etence

.n the article %>o 0ou 6eally ?ave a Global Strate)yF& 5?$6 "uly*,u)ust 2CL37, Gary ?amel and C.K. Prahalad e<amined one dimension of the )lobal competitive battle@ the race for brand dominance. This is the battle for control of distribution channels and )lobal %share of mind.& ,nother )lobal battle has been much less visible and has received much less mana)ement attention. This is the battle for control over key technolo)y*based competence that fuel new business development. ?onda has built a number of businesses, includin) marine en)ines, lawn mowers, )enerators, motorcycles, and cars, around its en)ine and power train competence. Casio draws on its e<pertise in semiconductors and di)ital display in producin) calculators, small*screen televisions, musical instruments, and watches. Canon relies on its ima)in) and microprocessor competence in its camera, copier, and laser printer businesses. .n the short run, the +uality and performance of a company!s products determine its competitiveness. 4ver the lon)er term, however, what counts is the ability to build and enhance core competences' distinctive skills that spawn new )enerations of products. This is where many mana)ers and commentators fear /estern companies are losin). 4ur research helps e<plain why some companies may be more likely than others to surrender core skills. 'll!ances or ( tso rc!n*? 9nticin) /estern companies into outsourcin) a)reements provides several benefits to ambitious 49M partners. Servin) as a manufacturin) base for a /estern partner is a +uick rout to increased manufacturin) share without the risk or e<pense of buildin) brand share. The /estern partners! distribution capability allows ,sian suppliers to focus all their resources on buildin) absolute product advanta)e. Then 49Ms can enter markets on their own and convert manufacturin) share into brand share. Servin) as a sourcin) platform yields more than (ust volume and process improvements. .t also )enerates low*cost, low*risk market learnin). The down*stream 5usually /estern7 partner typically provides information on how to tailor products to local markets. So every product desi)n transferred to an 49M partner is also a research report on customer preferences and market needs. The 49M partner can use these insi)hts to read the market accurately when it enters on its own. ' )atchet $##ect 4ur research su))ests that once a si)nificant sourcin) relationship has been established, the buyer becomes less willin) and able to reemer)e as a manufacturin) competitor. "apanese and Korean companies are, with few e<ceptions, e<emplary suppliers. .f anythin) the %soft option& of outsourcin) becomes even softer as 49M suppliers routinely e<ceed delivery and +uality e<pectations. 4utsourcin) often be)ins a ratchetlike process. 6elin+uishin) manufacturin) control and parin) back plant investment leads to sacrifices in product desi)n, process technolo)y, and, eventually, 6=> bud)ets. Conse+uently, the 49M partner captures product*development as well as manufacturin) initiative. ,mbitious 49M partners are not content with the old formula of %0ou desi)n it and we!ll make it.& The new reality is, %0ou desi)n it, we!ll learn from your desi)ns, make them more manufacturable, and launch our products alon)side yours.& )evers!n* the +ered!ct This outcome is not inevitable. /estern companies can retain control over their core competences by keepin) a few simple principles in mind. ! competitive product is not the same thing as a competitive organization . /hile an ,sian 49M partner may provide the former, it seldom provides the latter. .n essence, outsourcin) is a way of rentin) someone else!s competitiveness rather than developin) a lon)*term solution to competitive decline. "ethink the make#or#buy decision. Companies often treat component manufacturin) operations as cost centers and transfer their output to assembly units at an arbitrarily set price. This transfer price is an

accountin) fiction, and it is unlikely to yield as hi)h a return as marketin) or distribution investments, which re+uire less research money and capital. $ut companies seldom consider the competitive conse+uences of surrenderin) control over a key value*creatin) activity. $hat out for deepening dependence. Surrender results from a series of outsourcin) decisions that individually make economic sense but collectively amount to a phased e<it from the business. >ifferent mana)ers make outsourcin) decisions at different times unaware of the cumulative impact. "eplenish core competencies. /estern companies must outsource some activitiesG the economics are (ust too compellin). The real issue is whether a company is addin) to its stock of technolo)ies and competences as rapidly as it is surrenderin) them. The +uestion of whether to outsource should always provoke a second +uestion@ /here can we outpace our partner and other rivals in buildin) new sources of competitive advanta)eF