Technological diversity of persistent innovators in Japan

Research Policy 33 (2004) 531–549

Technological diversity of persistent innovators in JapanTwo case studies of large Japanese firms

Jun Suzukia,∗, Fumio Kodamaba The Institute for Future Technology, 2-6-11 Fukagawa, Koto-ku, Tokyo 135-8473, Japan

b Graduate School of Engineering Management (MOT Program), Shibaura Institute of Technology,5-37-8 Shiba, Minato-ku, Tokyo 108-0014, Japan

Received 14 January 2003; received in revised form 15 April 2003; accepted 29 October 2003

Abstract

We have investigated two large Japanese firms with their patent data, technological histories and product sales data of over 30years especially in terms of intra-firm technology diversification and interactions between multiple technological trajectories.Patent data showed the process of emergencies of technological trajectories and interactions (cross-fertilization) betweenthem quantitatively. Both persistence and diversity of technology have contributed to product diversification and sales growth.Based on our findings we have demonstrated that taking advantage of economies of scope in technology through persistenceand diversification is necessary for a technology-based firm if it is to survive and to grow for a prolonged period of time.© 2003 Elsevier B.V. All rights reserved.

Keywords: Patent analysis; Trajectory; Diversification; Persistence; Innovation

1. Introduction

The issue of persistence in innovative activitieshas been discussed from the viewpoint of patterns ofinnovative activities.Malerba and Orsenigo (1993,1995) have examined the way innovative activitiesare organized and take place within an industry interms of technological opportunity, appropriability,cumulativeness and properties of the knowledge base.Among these conditions, they focussed on cumula-tiveness conditions which refer to the probability thatexisting innovators may continue to be innovatorsin the future. They found that the Schumpeterian‘deepening’ pattern of innovation, characterized by

∗ Corresponding author. Tel.:+81-3-5245-1013;fax: +81-3-5245-1063.E-mail address: [email protected] (J. Suzuki).

the prevalence of large established firms and relevantbarriers against entry by new innovators, is closelyrelated to the cumulativeness conditions and is wellapplicable to the chemical and electronics industries.

Moreover;Malerba and Orsenigo (1996, 1999)ex-amined the patterns of innovative entry and exit usingEuropean Patent Office data combined with 49 tech-nology classes. They defined any firm that had consis-tently applied patents in a certain field of technologyas a ‘persistent innovator’. Their data indicates thatalthough turbulence is widely observed and a largepart of innovators are ‘occasional innovators’, inno-vative activities are generated by a relatively stablecore of large and persistent innovators. They also sug-gested that innovative activities in Japan and Germanyshow the typical Schumpeterian ‘deepening’ pattern,whilst those in Italy show the typical Schumpeterian‘widening’ pattern.

0048-7333/$ – see front matter © 2003 Elsevier B.V. All rights reserved.doi:10.1016/j.respol.2003.10.005

532 J. Suzuki, F. Kodama / Research Policy 33 (2004) 531–549

Recently,Cefis and Orsenigo (2001)investigatedthe prevalence of persistent innovators across coun-tries, industries and firms by size. They found that,although country-specific and industry-specific prop-erties exist and highly aggregated analysis might bemisleading, both great innovators and non-innovatorshave a strong tendency to remain in their state. Theypointed out again that Japanese and German firms havea stronger tendency to be persistent innovators thanfirms in the USA, the UK, France and Italy. They alsoimplied that in order to maintain innovative activities,persistence rather than the size of R&D expendituremight be important. Although they have not drawn anystrong theoretical conclusions, they implied that sus-tained innovative performance is generated and has tobe supported by a systematic and continuous processof accumulation of resources and competencies.

These preceding studies have successfully intro-duced a novel viewpoint of ‘persistence’ into innova-tion study and are very suggestive. However; we thinkanother important viewpoint, that of ‘diversification’,should also be considered in the discussion. The ex-istence of persistence is considered to be the evi-dence of ‘creative accumulation’ according to Cefisand Orsenigo, (Cefis and Orsenigo, 2001)and Japanis identified as a typical case. However; we know thatthere have also been a lot of instances of ‘creativedestructions’ or Schumpeterian widening patterns ofinnovation in Japan such as Canon, Sony, Honda andso on. A significant suggestion can be seen in the dataof Malerba and Orsenigo (1999). They found that theshare of ‘lateral (inter-industry)’ entrants is the highestin Japan. Their data seems to reflect the consequenceof the well-managed persistent innovators.

Technology diversification of firm or industry hasonly recently attracted an interest among researchers.After pioneering work byKodama (1986)andPavittet al. (1989), many researchers have shown in sev-eral micro-level case studies that cumulativeness orpath-dependency is a fundamental property of inno-vative activities whereas diversity is another factor.For example,Hamel and Prahalad (1994)emphasizedfrom the respective managements’ point of view thatmany once-successful companies had failed becauseof their lack of regeneration and their erroneous be-lief in persistence of yesterday’s business practices.Among the ways to successful corporate regenera-tion they credited corporate persistent diversity of

thinking. On the other hand,Patel and Pavitt (1997)andGranstrand et al. (1997)pointed out based on ananalysis of US patent records and case studies thatfirms’ technological competencies are dispersed over awider range of sectors than their production activities,and that firms are on the whole becoming more tech-nologically diversified over time. They suggested thathigh growth firms often followed a sequential strategywith technology diversification followed by productand/or market diversification. While they pointed outthat Japanese firms typically had the most developedmanagerial capability for concerted technology andbusiness diversification into new product areas, theyhave not shown how such firms exploited and/orcorrelated their technological knowledge practically.Also, they have not mentioned anything about theprocess of generation and renovation of technologicaltrajectories.

Gambardella and Torrisi (1998)found that a largeelectrical firm’s performance was positively asso-ciated with its technology diversification. However,they found that good performance was also positivelyassociated with greater focus on business operations.Gemba and Kodama (2001)analyzed the relation-ship between the diversification dynamics of R&Dactivity and the profitability of Japanese industries.They found that R&D diversification strategy intodownstream activities contributed to increased prof-itability. More recently,Breschi et al. (2003)testedthe extent to which firms diversify their innovativeactivities across related technological fields utiliz-ing knowledge-relatedness index. Their index wasmade on the assumption that the relatedness betweenfields of technology can be measured analyzing theco-occurrence of IPC codes assigned to individualpatent documents. Based upon findings, they con-cluded that firms diversify technologically along cer-tain directions that depend on the links and distanceamong technological fields.

Granstrand (1998)presented a theoretical frame-work for the technology-based firm and technologydiversification. He demonstrated the central roleplayed by technology diversification in the evolutionof a technology-based firm from the viewpoints ofeconomies of scale, scope, speed and space, respec-tively. He pointed out that the quality of the manage-ment of technology is a critical factor for this type offirm.

J. Suzuki, F. Kodama / Research Policy 33 (2004) 531–549 533

These results imply that persistence and diversifi-cation in innovative activities are closely related toeach other constituting principal elements of innova-tion. However, most of the existing research has notstudied in depth into the substance of technology di-versification adequately, and only a few researchershave shown the correlation between technological tra-jectories quantitatively. Why persistence does matter?Through which mechanism can technological diversi-fication contribute to the business expansion?

In this article, we adopt bibliometrics approach ofpatent analysis in order to clarify the process of gen-eration of novel technological trajectories and interac-tions between them. We selected two large Japanesefirms for case studies that have innovated and grownpersistently with the expectation that the typical‘lateral’ diversification (by Malerba and Orsenigo,1999) and ‘most developed managerial capability’(by Granstrand et al., 1997) can be observed.

2. Data and methods of analysis

2.1. The data

Data on the sales composition of a firm’s businessdomains was used as a proxy of the output of its in-novative process. Data on sales was taken from “TheJapan Company Handbook (Toyo Keizai, Inc.)”. How-ever; as firms have occasionally altered the definitionsof their business domains by themselves, and some-times have not released detailed data, the data on salesinevitably contains some degree of estimation. The35-year serial data sets adjusted by Japan’s GDP de-flator were examined.

In concordance with each business domain, tech-nological domain was identified and codified usingthe International Patent Classification (IPC) version 5.Data on patent applications from 1965 to 1999 camefrom Patent On-line Information System (PATOLIS)database based on the official bulletin of the JapanesePatent Office (JPO). We also used the JPO’s specificclassification code “facet code”. A facet code is ap-plied to some (not every) patents in addition to the IPCcode. It offers a more detailed description or a catego-rization that is different from what the IPC code offers(Ueno and Amano, 1998). For example, most medic-inal compounds are classified into a single “A61K”

sub-class by IPC, but these can be classified in furtherdetail by applicability to symptoms using their facetcode. We used the facet code to identify detailed tech-nological fields of Takeda’s medical compounds. InJPO patent records, IPC codes and facet codes appearas below:

A61K 31/50 (IPC code); ABR (facet code)

In order to identify firm’s technological domains,frequently observed IPC codes and facet codes in thefirm’s patent records were identified and sorted inrank order. Top ranking IPC codes and facet codeswere classified into technology fields representingthe respective firm’s major business domains such as“organic synthesis technology”, “genetic engineer-ing technology”, “medicinal compound technology”,“camera technology” and “copier technology”, etc.(Tables 1 and 2). Patent application in each field indi-cates an accumulation of knowledge and advancementof technological trajectory.

2.2. Primary IPC and IPC co-occurrence analysis

In concordance with WIPO’s manual, Japanesepatent application records have single or multiple IPCcodes. The IPC code placed at the primary position,namely “primary IPC”, has special meaning. The pri-mary IPC represents the core technology which theinvention pertains to. Any additional IPC code whichmay exist represents the specific application field ora closely related technology field pertinent to the in-vention. When multiple IPC codes exist in a singlepatent and those codes belong to different technologyfields, we call this a case of “IPC co-occurrence”.

Verspagen (1996)argued inter-sectoral knowledgespillovers assuming that the primary IPC provides agood proxy of the knowledge-producing sector, andthat the additional IPC codes provides a good proxy ofthe spillover-receiving sectors (see alsoGrupp, 1996for a similar idea). He used the MERIT concordancetable to assign IPC codes to the International StandardIndustrial Classification (ISIC) sectors (Verspagenet al., 1994). In comparison with other methodologiesmeasuring knowledge spillovers like the Yale ma-trix, Verspagen insisted that IPC co-occurrence-basedmeasures can figure out technological linkages purerthan “transaction-based approach”.

534J.

Suzuki,F.

Kodam

a/R

esearchPolicy

33(2004)

531–549

Table 1Frequently observed IPC codes on Canon’s patent and its classification into technology fields

Technology fields IPC codes

Optical lens B29C B29D B29F B29H B29L B32B C03B C03C C04B C09D C09J G02B G02F

Camera G01B G01C G01D G01F G01G G01J G01K G01P G01R G03B G03C G03D G03G G03HG04F G04G G05B G05D G05F G07C G08C

Printer and copier B05B B05C B05D B43L F24HB41C B41F B41J B42B B42C B42D B65D 665G B65H C09B C09C F26B G11B G11CB41L B41M H04N

Semiconductormanufacturingequipment

B23H B23K B23P B24B C21D C23C C25D C30B G03F H01L H05K

Digital data processing G06F G06K G06T

OthersChemistry C07C C07D C08C C08F C08G C08J C08K C08LMechanical F16C F16D F16F F16H F16M H02KDisplay C09K G09F G09G H01JCommunication H04B H04J H04L H04MBattery H01MOther (all other IPC codes)

J.Suzuki,

F.K

odama

/Research

Policy33

(2004)531–549

535

Table 2Frequently observed IPC and facet codes on Takeda’s patent and its classification into technology fields

Technology fields IPC codes Facet codes for A61K sub-class

Foods and vitamins A23L A61K ADE ADF ADG ADH ADJ ADK ADL

Pharmaceuticals A61KBrain and nerve A61K AAB AAC AAD AAE AAF AAG AAH AAJ AAK AAL AAM AANImmune systems A61K ABN ABP ABC ABD ABE ABF ABG ABHCardiovascular systems A61K ABN ABP ABQ ABR ABS ABT ABU ABV ABW ABXDigestive systems A61K ACJ ACK ACL ACM ACN ACP ACQ ACR ACS ACT ACUCancer A61K ADU ADV ACJInfectious diseases A61K ADW ADX ADY ADZ AEA AEB AECEndocrine systems A61K AED AEE AEF AEG AEH AEJ AEK AEL AEM AEN AEP AEQ

Organic chemistry C07B C07C C07D C07F C07G C07H C07J C07M

Microbes and fermentation C12P C12R

Genetic engineering andprotein engineering

C07K C12N C12Q

OthersInsecticides A01NOther (all other IPC codes)


As mentioned earlier,Breschi et al. (2003)used IPCco-occurrence approach for knowledge-relatedness in-dex. However; they utilized only the frequencies ofIPC co-occurrence regardless of primary-additional re-lationship.

In this paper, we will show here that applicationof “primary IPC” and “IPC co-occurrence” approachin analysis of a firm’s patent application historygives us a useful tool for better understanding of thelinkages between technological trajectories and theirdynamics.

3. Case studies

3.1. Canon

Although Canon is classified as an electronicsequipment manufacturer these days, it started out asa camera manufacturer in the 1930s. Canon has ag-gressively diversified its technological development,business development and marketing activities espe-cially over the last 30 years. Consequently, the firm’snon-consolidated net sales have grown to about 1.7trillion yen (FY 2000) and domestic sales accountsfor just 29% of this amount. In addition, the firm iswidely recognized as one of the most active patentapplicants in the world. Canon is regarded as a goodrepresentative example of a Japanese persistent inno-vator.

3.1.1. Canon’s business domain and technologicaltrajectories

Canon started its R&D activity with reverse engi-neering of foreign camera products and with creatingits own product by adding some new ideas. Its tech-nologies had been restricted to the field of cameramanufacturing until the late 1950s. In 1957, their firstattempt in technology and business diversification hadbegun with the development of an electro-magneticdata recorder/reader, namely “The Synchroreader”.Although business itself was not successful with thedata recorder, the newly hired engineers and acquiredknowledge of digital processing constituted a noveltechnology core of digital data processing.

Our investigation of Canon’s technological his-tory combined with its business history reveals thatthe firm has developed four major technology cores

continuously, and each of these is closely relatedto the respective business domain. Their “cameratechnology” core has accumulated and evolved dueto persistent development of cameras and optical in-struments. Its trajectory is characterized mainly byautomation. Their “digital processing technology”core which originated from development of TheSynchroreader has evolved into electronic calcula-tors, word processors, personal computers, and soon. The “electro-photographic technology” core hasbeen extended to copiers and printers, and this is thelargest business domain of Canon today. Its trajec-tory is characterized mainly by higher resolution andmaintenance free. The “semiconductor manufacturingtechnology” core has evolved with photo-lithography,and Canon is now one of the largest manufacturers ofmask aligners and steppers in the world. The patentapplication count is taken as a proxy of the accumu-lation of knowledge in each technology core, thusshowing technological trajectory. Of course, therehave been many other technologies developed andapplied by Canon, but most of those are generic tech-nologies and cannot be readily classified into discretebusiness domains.

Fig. 1 illustrates Canon’s history of business struc-ture (non-consolidated net sales) and its diversity rep-resented by entropy.1 Canon’s business diversity hashad tendency to increase continuously for 35 yearswith steep rises in the late 1960s and the early 1980s.The former steep rise corresponds to the growth insales of electronic calculators, and the latter corre-sponds to sales of printers, word processors and fac-similes. The rise in sales of copiers compensated fordiminishing sales of electronic calculators.

Fig. 2 illustrates Canon’s technological accumula-tion and diversity represented by patent applicationcounts and entropy. It is apparent that the progressionof technology diversity went in parallel with that ofbusiness diversity.

Fig. 3 visualizes the assumable of interactionsamong several Canon’s technological trajectories.We have drawn this chart by investigating public

1 The entropy value is an index of the degree of diversificationdefined by�iPi ln(Pi): where Pi is usually the product sales ratioand ln stands for natural log. The merit of using the entropy valueIn comparison with the Herfindahl index was reviewed byGembaand Kodama (2001).


Fig. 1. Canon’s history of business structure and its diversity.

literatures (firm’s business history and commerciallyavailable business books) and by interviews. Boxedarrows indicate technological domains and thin ar-rowgraphs between domains indicate technologicalapplications or spillovers identified from specificevents or projects.

3.1.2. Interactions between technological trajectoriesIn addition to the descriptive approach, the quan-

titative “IPC co-occurrence” and “primary IPC”

Fig. 2. Canon’s technological accumulation and diversity.

analyses were applied to shed light on the relationshipof Canon’s technological trajectories.

3.1.2.1. From “camera” to “copier and printer”. In1965, Canon applied for a total of 126 patents. Amongthose, according to their respective primary IPC codes,were 91 of camera technologies, 14 of copier technolo-gies, 7 of optical equipment technologies, 4 of semico-nductor manufacturing technologies, 3 of digital dataprocessing technologies, and 7 of other technologies.


Fig. 3. Interactions between Canon’s core technological trajectories.

Patent applications for camera technologies hadincreased gradually until 1981 reaching an almostconsistent flat level of 2000 per year. On the otherhand, patent applications for copiers and printers(those are not distinctive) had increased rapidly fromaround 1977, exceeding that for cameras in 1983,reaching nearly 4000 per year in 1986.

Fig. 4illustrates detailed data on IPC co-occurrencebetween the fields of “camera” and “copier & printer”.The persistent IPC co-occurrence implies interactionbetween these fields. Furthermore, IPC co-occurrencewith primary IPC in “camera technology” had beendominant until 1977, but those with primary IPCfor “copier & printer technology” had overtakenthem after 1978. It is concluded that the applica-tion of camera technology had generated and helpedto develop the technological trajectory of copiersthroughout the 1960s and the early 1970s. How-ever, the technological trajectory of copiers hadprobably come to stand somewhat independentlyon its own legs in the late 1970s and this has

contributed significantly to the copier and printerbusiness.

3.1.2.2. From “camera” to “semiconductor manu-facturing equipment”. Canon’s patent applicationcounts for semiconductor manufacturing equipmenttechnologies had increased gradually throughoutthe late 1960s forming a small peak at the begin-ning of the 1970s which reflects the emergence ofa “mask aligner”. In the early 1980s, patent ap-plications had increased sharply almost reaching aplateau in the middle of the 1980s. The sharp in-crease is considered to account for the emergenceof the second-generation semiconductor manufactur-ing equipment, namely “stepper”. Patent applicationsin the 1990s should account for the refinement on“stepper” and “post-stepper” technology correspond-ing to the diminishing design rules for semiconductormanufacturing technology.

IPC analysis of the patents that have IPC codes rang-ing between camera technologies and semiconductor


Fig. 4. IPC co-occurrence in Canon’s patents between the fields of “camera” and “copier and printer”.


manufacturing equipment technologies reveals the dy-namic characteristics of the relationship between coreand emerging technologies.

Fig. 5illustrates detailed data on IPC co-occurrencebetween “camera” and “semiconductor manufacturingequipment”. In the late 1960s and the early 1970s,there had been a certain number of patent applica-tions which had the primary IPC codes for cameratechnologies and the additional codes for semiconduc-tor manufacturing technologies, meaning that Canonhad developed the “mask aligner” by applying cam-era technologies. In other words, the “mask aligner”was merely a subsidiary product of the high- resolu-tion camera at the beginning.

After the era of the “mask aligner”, there was an-other technological flow from “camera” to “semicon-ductor manufacturing equipment” in the early 1980s.From 1980 until 1984 the number of patent appli-cations with the primary IPC codes for camera andthe additional IPC codes for semiconductor manufac-turing equipment technologies had increased sharply.However; the relationship of the primary core tech-nology and the additional technology shown by thecodes in the IPC co-occurrences had been invertedin one leap with the steep rise in patent applicationsconcerning the “stepper” in 1985. It means that thesemiconductor manufacturing equipment technologyhad probably come to stand independently, forming anovel technological core in the middle of the 1980s,and then had begun to follow its own trajectory.

Precise manufacturing technologies using photo-lithography developed with the “stepper” have beenapplied to the manufacturing process of the head com-ponent of the high resolution ink jet printer. Also, thetechnologies for semiconductor manufacturing pro-cesses are essential for manufacturing the exposuredrum with amorphous silicon film for copiers and laserprinters.

Although the contribution to Canon, from sales ofsemiconductor manufacturing equipment, was not sogreat in the era of the “mask aligner”, it became muchmore so with the “stepper”, and now it is one ofCanon’s core businesses.

3.2. Takeda chemical industries

Takeda is the largest pharmaceutical firm in Japan.It was founded as a small medicine wholesaler over

two centuries ago. It was incorporated in 1925 andsubsequently listed on the Tokyo and Osaka stock ex-changes in 1949.

Takeda’s non-consolidated net sales is about 760billion yen (FY 2000). Domestic sales account for71% of this. Although the proportion of domesticsales looks high, the net sales of TAP Pharmaceuti-cal Products Inc. based in the USA which is a 50:50joint venture between Takeda and Abbott Laborato-ries where products are mostly from Takeda is 3.2billion dollars (about 400 billion yen). These salescan also be regarded as additional overseas sales forTakeda. In addition, the firm has long been a top rank-ing patent applicant among Japanese pharmaceuticalfirms. Thus, in the Japanese pharmaceutical industry,Takeda is regarded as an outstanding global firm anda persistent innovator.

3.2.1. Takeda’s business domain and technologicaltrajectories

In 1895, Takeda established its own factory toproduce medicines such as quinine hydrochloride inOsaka and became a pharmaceutical manufacturer.Takeda’s research activities began with the establish-ment of a research division in 1915.

The executive officers of Takeda have modeledtheir firm on Merck & Co. Inc., based in the USAaiming to be a research-driven pharmaceutical prod-ucts company. Although both of the firms make muchof innovative R&D, a difference in product devel-opment strategy had once existed (Morita, 1991).Merck had well-defined targets in each therapeuticarea and its research and development activities havebeen needs-oriented. On the other hand, Takeda hadfocused on the discovery and exploitation of un-known compounds produced by microbes in theirearlier days. This seeds-oriented strategy brought thefirm an opportunity to expand its business domaininto the fields of agriculture, food and industrialchemistry.

Fig. 6 illustrates Takeda’s history of business struc-ture and its business diversity represented by entropyTakeda’s conventional competitive products have tra-ditionally been in the fields of antibiotics and syn-thetic vitamins although the firm’s business diversityhad increased during the late 1960s and the early1970s corresponding to embarkation on a journey intofields other than pharmaceuticals. However, since the


Fig. 5. IPC co-occurrence in Canon’s patents between the fields of “camera” and “semiconductor manufacturing equipment (SME)”.


Fig. 6. Takeda’s history of business structure and its diversity.

middle of the 1970s, needs-oriented targeted R&D ac-tivities had come to the mainstream at Takeda too,and restructuring of the business had begun resultingin decreasing business diversity. Recently, Takeda de-cided to transfer its side businesses such as food sup-plements, balk vitamins, urethane chemical products,and animal health products to joint ventures or com-petitors. This clearly indicates that Takeda is focusingor specializing on a narrow set of core businesses (a“back to basics” strategy).

Fig. 7. Takeda’s technological accumulation and diversity.

However, in the field of pharmaceutical products,a different strategy can be observed.Fig. 7 illustratesTakeda’s patent application count andFig. 8illustratesthe share of therapeutic areas identified by their facetcodes with the IPC A61K (preparations for medical,dental, or toilet purposes).

Although we cannot directly analyze the data onsales amounts segmented to each therapeutic area,it appears certain that Takeda has actively diversi-fied its pharmaceutical product related technology


Fig. 8. Takeda’s technological diversity in pharmaceutical field identified from IPC A61K and additional facet codes (seeTable 2).

Consequently, the direct outcomes of these patentsconstitute some of the recent main products of Takeda.These main products include hormone analogue (forprostate cancer, authorized in 1992), proton pumpinhibitor (for peptic ulcers, authorized in 1992), re-ceptor antagonist (for hypertension, authorized in1997), and insulin sensitizer (for diabetes, authorizedin 1999). Those emerging products cannot be catego-rized into Takeda’s conventional mainstream productfields but greatly contribute to the firm’s penetrationinto the world-wide market.

Takeda’s technology which is closely related tothe pharmaceutical product seems to have diversifiedrather suddenly but the process can be better under-stood when considered together with generic leveltechnological trajectories.

3.2.2. Trajectories of generic technologiesIn the 1940s, Takeda began exploratory research of

antibiotics and research of synthetic folic acid in ad-dition to vitamin C and B1. It also started productionresearch for penicillin, which it started manufacturingby quasi-synthetic (with fermentation) technology in1948, and then during the 1960s and the 1970s, Takedahad developed some novel antibiotics especially thethird generation cephalosporins, which were the firstin the world. Also, Takeda had succeeded in launch-ing new businesses with synthetic sodium glutamate

and a mixture of the purine derivatives extracted fromyeast.

These accomplishments ensured its fame as a “vi-tamins and antibiotics expert company”. The impor-tant generic technologies underlying these successeswere “synthetic organic chemistry” and “microbefermentation”.

In addition to these, research into receptors and lig-ands and bioactive substances had begun in the early1980s and advanced significantly with the emerginggenetic engineering (GE) and protein engineering (PE)technologies. Takeda launched its Tsukuba researchlaboratories specialized in basic research on orphanreceptors (receptors with unknown functions) in 1988.In these days, bioinformatic and pharmacogenomictechnologies are utilized together with GE and PE.

Fig. 9 illustrates the impact or application of spe-cific core technologies from certain trajectories on theothers concerning Takeda’s R&D activities.

Fig. 10illustrates detailed data on IPC co-occurrencebetween “microbes and fermentation” and “organicchemistry”. In the early 1970s, most of the appliedpatents that had IPC co-occurrences covering “mi-crobes and fermentation” and “organic chemistry”had primary IPC codes for the former technologyThose patents should primarily regard isolation, pu-rification and structural elucidation of organic com-pounds produced by microbes, whereas, in the 1980s,


Fig. 9. Interactions between Takeda’s technological trajectories.

IPC co-occurrences that had primary IPC codes for“organic chemistry” had risen sharply. It means thatthe relative importance of these technological trajec-tories had changed with the advancement of syntheticorganic chemistry technology and the beginning of amaterial patent. Although the relative importance of“microbes and fermentation” technology had dimin-ished compared to “organic chemistry”, the patentapplication count for the former technology had notdecreased significantly as it had been closely relatedto the other emerging technologies namely GE andPE.

In 1974, Cohen and Boyer applied for the famousgene-splicing patent, and then patent applications con-cerning GE had begun to increase from the beginningof the 1980s in Japan.

Fig. 11illustrates detailed data on IPC co-occurrencebetween “GE and PE” and “organic chemistry”. Itis apparent that IPC co-occurrences with primaryIPC for “GE and PE” had increased in the 1980s in

contrast to those with primary IPC for “microbesand fermentation” inFig. 10. In addition, the pri-mary IPC had shifted from “organic chemistry” to“GE and PE” in the 1990s. It is likely that Takeda’spatent applications for “GE and PE” and “organicchemistry” had been consequences of its concern inpreparation of bioactive substances by GE and PE inthe 1980s, and in genes of receptors and ligands in the1990s.

Takeda’s technological trajectories at the generictechnology level had diversified around the coreof synthetic organic chemistry technology fusingthe exotic technologies such as GE, PE and, morerecently, pharmacogenomics with their conven-tional technologies including microbe culture andfermentation.

It is probable that technology diversification at thegeneric level had impacted on the product level tech-nologies thus driving diversification of the pharma-ceutical products.


Fig. 10. IPC co-occurrence in Takeda’s patents between the fields of “microbes and fermentation” and “organic chemistry”.


Fig. 11. IPC co-occurrence in Takeda’s patents between the fields of “genetic engineering and protein engineering” and “organic chemistry”.


4. Discussion

4.1. Technology diversity and business diversity

We have demonstrated that patent cross class anal-ysis can reveal the interactive histories of technolog-ical trajectories and the directions of their progress.This methodology should be useful for in depth un-derstanding of diversity and path-dependency in termsof persistent innovation at the firm level.

Now, we can identify two types of technology diver-sification. Canon diversified its technology by mostlyexploring around their core technology concerning“camera”. It seems to be presenting the model of en-dogenous technology diversification generating per-sistent innovative entry Diversified technology basemade Canon able to diversify their business domaininto copiers, printers, semiconductor manufacturingmachines and so on.

On the other hand, Takeda’s case seems to repre-sent the other model of technology diversification byimportation of exotic technologies and by fusing themwith existing technologies. At the generic technol-ogy level, Takeda imported genetic engineering, pro-tein engineering and genome informatics technologies.Then those technologies were fused with core tech-nologies such as organic synthesis and fermentation.It brought Takeda technology diversification at prod-uct level and persistent innovative entry to a variety ofmedicines for many symptoms.

In both cases, the firms have evolved intoworld-class competitive performers. These resultssuggest a close relationship between internal techno-logical diversity and the competitiveness of the firm.Particularly, attention should be paid to the fact thatpersistent technology diversification (endogenous orby importation) fosters persistent business diversifi-cation.

Leonard-Barton (1992)examined the nature of thecore capabilities of a firm focusing in particular ontheir interaction with new products and process de-velopment projects. She pointed out that although thecore capabilities enable innovation, they sometimeshinder it too, and they cannot be managed as a con-sistently positive entity.Granstrand (1998)concludedbased on empirical studies that technology diversifi-cation leads to the growth of the technology-basedfirm mostly through business diversification. He also

suggested a cross-fertilization effect between differenttechnologies.

We would like to continue the discussion onthe cases of diversification and renovation throughcross-fertilization of technologies.

4.2. The generation and renovation of technologicaltrajectories

Detailed investigation of patent application data anda firm’s technological history provides evidence thatdiversification of the core technology generates newtechnological trajectories adjacent to the existing core.It was also demonstrated that generated technologicaltrajectories sometimes link directly to new product de-velopment and market entry, but sometimes affect newproducts indirectly through regenerating other techno-logical trajectories.

Henderson and Clark (1990)pointed out that a per-sistent process of accumulation of competencies mayoften generate lock-in effects and “competence traps”.Our data suggests that internal technology diversifi-cation may play a preventive role against such trapsthrough generating and renovating technological tra-jectories.

Miyazaki (1995)pointed out that Japanese firms hadbeen building competencies around the key technologyareas that enhance a firm’s core capabilities. She alsosuggested that competence-building can be plannedstrategically to enable organizational transformation.

Our findings are consistent with the conclusions ofMiyazaki (1995)in the sense that Japanese persistentinnovating firms may have had sustaining technolo-gies but sometimes disruptive products. They havesometimes renovated or metamorphosed themselvesdramatically in business like Canon (from camerato printer), Toyota (from loom to car), Sharp (fromstationary to electronics), and so on.Granstrand(1998)pointed out that combined technology/productbase shifts over time. It sometimes shifts partiallywith sticking to the original business area but some-times shifts completely away from the originalarea. He called the former phenomenon “rooteddiversification” and the latter “floating diversifi-cation”.

Analysis of Canon’s patents illustrates that its copy-ing machine was intrinsically an electro-photographicmachine in the 1960s and the core technologies of its


copying machine was not exactly the same as thoseof its camera technologies but bore a close relation-ship. Also, its semiconductor manufacturing machine(mask aligner) was intrinsically ultra high-density ex-posure equipment and had a close relationship withthe camera too. Moreover, Takeda’s technology de-velopment from fermentation to quasi-synthesis andpure synthesis of antibiotics enabled the firm to pro-vide an effective and broader spectrum of productsconsistently. These successful cases of core technol-ogy diversification into adjacent fields were developedunder the definite goal of business diversification.

On the other hand, electronics technology (typicallyan electronic calculator) was exotic and remote fromthe core technology for Canon as a camera manufac-turer in the 1960s. The businesses developed directlywith this exotic technology like the electro-magneticreader or electronic calculator could not become oneof the persistent core businesses. Also, Takeda’s en-terprise into remote technology domains, such asfoods or industrial chemicals production, turned outto be unsuccessful. However, the exotic technologieslike electronics and software for Canon and geneticengineering and protein engineering technologies forTakeda can sometimes persist and be developed alongtrajectories. These exotic or imported technologicaltrajectories have been valuable for the firms becauseof their ripple effect on traditional technological tra-jectories and as the source of prompt countermeasuresagainst the unexpected challenge by outsiders withdisruptive technologies.

In conclusion, case studies of persistent innovatorhave elucidated the importance of persistent knowl-edge accumulation in multiple technology fields inorder to take advantage of cross-fertilization or syn-ergy effect. From this view, technological diversity hasimportance in providing independent trajectories al-though they would not generate new product/processdirectly.

Our findings are consistent with the conclusions ofGranstrand et al. (1997)that the “focus and back tobasics” strategy in R&D activities needs careful con-sideration as technologies are not the same as productsand must be dealt with differently. Besides industrialconditions like those in Japan (i.e. the proportion ofpersistent innovators is quite high), technology diver-sification inside the firms may play an important rolefor innovation.

Acknowledgements

The authors would like to thank Kiminori Gemba,Robert Kneller and other colleagues at the ResearchCenter for Advanced Science and Technology at theUniversity of Tokyo. The authors also would like tothank anonymous referees for comments and helpfulsuggestions. This research was partially funded by TheInstitute For Future Technology (IFTECH) and the di-rectors of IFTECH had given permission to one of theauthors for involvement in this research while work-ing. Their support is gratefully acknowledged.

References

Breschi, S., Lissoni, F., Malerba, F., 2003. Knowledge relatednessin firm technological diversification. Research Policy 32, 69–87.

Cefis, E., Orsenigo, L., 2001. The persistence of innovativeactivities: a cross-countries and cross-sectors comparativeanalysis. Research Policy 30, 1139–1158.

Gambardella, A., Torrisi, S., 1998. Does technological convergenceimply convergence in markets? Evidence from the electronicsindustry. Research Policy 27, 445–463.

Granstrand, O., 1998. Towards a theory of the technology-basedfirm. Research Policy 27, 465–489.

Granstrand, O., Patel, P., Pavitt, K., 1997. Multi-technologycorporations: why they have “distributed” rather than “distinc-tive core” competencies. California Management Review 39 (4),8–25.

Gemba, K., Kodama, F., 2001. Diversification dynamics of theJapanese industry. Research Policy 30, 1165–1184.

Grupp, H., 1996. Spillover effects and the science base ofinnovations reconsidered: an empirical approach. Journal ofEvolutionary Economics 6, 175–197.

Hamel, G., Prahalad, C.K., 1994. Competing for the Future.Harvard Business Press.

Henderson, R., Clark, K., 1990. Architectural innovation: thereconfiguration of existing product technologies and the failureof established firms. Administrative Sciences Quarterly 35, 9–30.

Kodama, F., 1986. Technological diversification of Japaneseindustry. Science 233, 291–296.

Leonard-Barton, D., 1992. Core capabilities and core rigidities:a paradox in managing new product development. StrategicManagement Journal 13, 111–125.

Malerba, F., Orsenigo, L., 1993. Technological regimes and firmbehavior. Industrial and Corporate Change 2, 45–74.

Malerba, F., Orsenigo, L., 1995. Schumpeterian patterns ofinnovation. Cambridge Journal of Economics 19, 47–65.

Malerba, F., Orsenigo, L., 1996. Schumpeterian patterns ofinnovation are technology-specific. Research Policy 25, 451–478.

Malerba, F., Orsenigo, L., 1999. Technological entry, exit andsurvival. Research Policy 28, 643–660.


Miyazaki, K., 1995. Building Competencies in the firm. MacMillanPress Ltd.

Morita, K., 1991. Takeda takes aim at world markets with targetedalliances, Pharmaceutical Executive 11, 1.

Patel, P., Pavitt, K., 1997. The technological competencies of theworld’s largest firms: complex and path-dependent, but notmuch variety. Research Policy 26, 141–156.

Pavitt, K., Robson, M., Townsend, J., 1989. Technologicalaccumulation, diversification and organization in UK companies1945–1983. Management Science 35, 1.

Ueno, S., Amano, H., 1998. Improvement of the IPC. Discussionpaper for advanced seminar on the international patentclassification, Newport, United Kingdom, 7–11 December 1998.IPC/SEM/98/10.

Verspagen, B., 1996. Measuring intersectral technology spillovers:estimates from the European and US Patent Office Databases.Economic Systems Research 9 (1), 47–65.

Verspagen, B., van Moergastel, T., Slabbers, M., 1994. MERITconcordance table: IPC-ISIC (rev. 2). MERIT ResearchMemorandom 94-004.

Technological diversity of persistent innovators in Japan

Documents

Transcript of Technological diversity of persistent innovators in Japan