An Integrated Approach To The Diffusion Of Technological Innovation

Technovation, 15(1) (1995) 11-24

An integrated approach to the diffusion of technological innovation: the non-manufacturing advanced robotics field Francesca Sgobbi c/o Dipartimento di Economia e Produzione, Politecnico di Milano, Via Ponzio 34/c, 20133 Milano, Italy

Abstract The search for new investment evaluation techniques is justified by the increasing gap between the current scenario and the technological and environmental conditions where the 'traditional' techniques have been developed. This is particularly true for 'embryonic technologies'--those technologies so innovative that the actual number of their operative applications is still negligible. It is important to evaluate the relevance of the technology since its embryonic stage springs from the need to assess its impact over society as a whole. A survey of economic literature underlines that neither traditional nor innovative techniques are able properly to evaluate embryonic technologies: beyond the need for historic data, they barely stress the role played by linkages between the expected benefits and the economic participants involved in determining the success of the technology. An evaluation model is thus developed which focuses on those elements. The way to build a set of indicators useful for assessing the potential of embryonic technology and monitoring its evolution is also depicted. An example related to an embryonic technology (advanced robotics) is discussed.

1. Introduction The debate on evaluation techniques used to assess the development possibilities of innovative technologies gains strength from a correct appraisal of the importance of their impact. This acquires new relevance after the introduction of technologies whose main characteristics are: • the importance of factors not immediately connected to monetary flows;

Technovation Vol. 15 No. 1

• the pervasiveness of their effects on both personal and space-time requirements; • the concomitant growth of disturbance in the global socio-economic context of their introduction. Traditional appraisal approaches do not include these new elements in their evaluation. The less available and reliable the information is, the lower the possibility of evaluating the chances of success. This is the case for new

0166-4972/95/US$07.00 © 1995 Elsevier Science Ltd

11

F. Sgobbi

technologies characterized both by absolutely new contents and by a small range of commercial solutions because of the development stage where most of the proposed applications still happen to be. In this situation, the shortcomings of currently disposable methodological tools seem to be large. This work focuses on the appraisal of possible uses for this kind of technology, referred to as 'embryonic technology'. The target of the work is the development of a framework to evaluate the impact of embryonic technologies over the economic structure they affect. The framework will consider both quantitative and qualitative elements. The paper is divided into four sections. The first section deals with the unusual aspects of embryonic technologies compared with those of generic innovative technologies. The aim is to stress the problems which cannot be solved by traditional approaches, and to define the requirements which the proposed model should answer. The next section reviews the main contributions in the literature in order to show the state of the art concerning the appraisal of radically new technologies. The target is to assess the possibility of extending existing approaches to embryonic technologies. The third section is devoted to the formalization of the proposed framework, the advantages of which are underlined, while the last section displays the possible application of this methodology in the field of advanced robotics for non-manufacturing activities. The chosen example concerns the hospital-service robot PAM.

2.

The problem

The paper deals with the probability of success of those technologies (product and/or process) characterized by: • extremely

new technological content, when compared with the current state of the art; • commercial exploitation limited to a reduced set of products, since most conceived applications are in the project and prototype stages

12

because of the large number of technical solutions still to be tested and standardized; pervasiveness of the economic impact, which spreads throughout the different stages of the production chain and affects all the people involved. We define such technologies as embryonic to underline the fact of their potential for many applications. An example of such a technology is advanced robotics, on which the last section of this paper focuses. Any innovative technology acquires tangible economic relevance only when it is actually in use: "the availability of a new product, or a new process, confers neither private nor public benefit before the innovation is adopted" [1]. This means that the innovation shows its true relevance and its impact on the socio-economic setting only when it has been accepted. The need to quantify the overall impact of the innovation in order to assess its scope is underlined by Gibbons and Metcalfe [2] and by Stoneman [3] as the main justification for adoption/diffusion studies. From this perspective, the problem of evaluating the chances of a generic innovative technology turns to the challenge of how to define and measure the scope of the technology itself. The first step is the selection of the proper tools to ensure the most correct and complete appraisal. The more uncertainty there is concerning possible uses for the new technology, the more difficult the choices appear. Uncertainty about their prospects seems to be the main feature of all technologies in the embryonic stage, and this involves both endogenous and exogenous aspects. Evolutionary paths are not yet defined, and meanwhile the 'absolute' or radical novelty of the embryonic technology prevents anyone from referring to past experience. This situation stems from the lack of reliable historical data, and therefore the impossibility of depicting the current state of the art. Thus quantitative methods cannot extrapolate future trends for either the technology or its market. Uncertainty increases when the technology is

T e c h n o v a t i o n Vol. 15 No. 1

Diffusion of technological innovation

introduced within a turbulent environment, where structural changes are in progress. The uncertainty of the external context thus sums up the ongoing feature of applications. This results in increased difficulty in identifying those factors which are important in order to achieve the expected benefits. The continuous updating of appraisal elements becomes especially important in turbulent situations, in order to verify their relevance to future constraints and opportunities. Moreover, linkages between significant factors change over time, as do their relative degrees of critical importance, according to the state of the art and the evolution of the general environment. Pervasiveness represents one particular aspect of many recently introduced technologies (e.g. microelectronics, computer-integrated manufacture (CIM), telecommunications, etc.). This increases the uncertainty about the prospects of an embryonic technology, since the consequences of the innovation affect more stages of the productive chain and modify the innovator's connections within that chain, changing previously existing balances. In this situation, any attempt to assess the relevance of the new technology must evaluate its impact over all the economic areas involved. In fact, the probability of the innovator's success increases when people in other related areas can also expect some benefits from the new technology. Analysis of other people's reactions is always relevant for the innovator, since it shows the possible areas of opposition to acceptance of the new technology. It therefore becomes possible to assess the suitability of actions designed to eliminate or reduce the constraints outlined. The pervasiveness of innovation itself underlines the need to enlarge the set of factors to include in the technology appraisal. This indication is further stressed by the increasing importance of intangible aspects (i.e., those not immediately related to monetary flows) in assessing the actual gain of expected costs and benefits. Both the highlighted aspects acquire further relevance when considered together with the growing turbulence of the global context in which the innovative technology takes place.

Technovation Vol. 15 No. 1

This brief analysis has outlined the main questions that a model aimed at the correct appraisal of a technology in the embryonic stage should answer. They can be summed up as: • significance of the innovation for all the economic participants likely to be involved; • determination of those factors acting as the main constraints or incentives to development; since they influence the occurrence of expected costs and benefits the identification of their mutual linkages also becomes relevant. An appraisal of the scope of an embryonic technology involves the choice of the tools used by the decision maker to make up his (or her) mind about the technology itself. A framework able to identify the connections between important aspects of the achievement of potential benefits by the different participants involved becomes really important after the evaluation has convinced the decision maker to introduce the new technology. It actually enables him to adopt an 'active' management of the technology development, since it allows him • to identify the potential improvements (measured in terms of attainable benefits) consequent to specific programmes undertaken by the different people involved economically, and to define the measurable targets which the programme should aim at; • to dynamically control the evolution of intervention priorities in time.

3. Background Both diffusion studies and adoption justification methods deal with the process of introducing an innovative technology within a certain socioeconomic setting. The two approaches differ in the way they look at the problem, Diffusion studies focus on impact dynamics, i.e. the trend of adoptions in time, while adoption justification routines are concerned with the effects of the introduction of a particular application.

13

F. Sgobbi

In the case of embryonic technologies, the employment of diffusion models seems unsatisfactory to explain development trends. In fact, those models implicitly assume that the techno-economic feasibility of the innovation has already been proved and that the number of single adoptions which took place in the past is great enough to generate a diffusion curve. This scenario is misleading for an embryonic technology as defined in the previous section, since the development time span covers at least the very first part of the logistic curve generally used to represent the innovation diffusion process. Analysis of recent trends seems worthwhile from the point of view of a global approach to the understanding of such phenomena. Indeed, it allows the identification of further considerable factors able to influence the way in which the innovative technology spreads.1 The strategic approach, which refers to game theory, has the advantage of explicitly considering the weight of linkages among competitors in determining the speed of diffusion. The same factor is thus included among those affecting the probabilities of success for the technological innovation: "the decision to adopt is an eminently individualistic choice for the company, which also has to consider the strategic interdependencies with competitors" [1]. The most recent developments of the model attempt to overcome the main limits of the original formulation by Reinganum [4], introducing learning curves and further improvements in technology following the first introduction. The strategic approach does not weigh the role of uncertainty and, moreover, includes only the innovator's direct competitors among the game participants; i.e., it considers only horizontal linkages. Among epidemic models, the research is concerned with analysing the role of information in determining the level of risk faced by the adopter. On the other hand, evolutionary models consider the diffusion of a new technological pattern as the result of selective processes which operate from an initial variety of proposed solutions. Attention has concentrated upon selection mechanisms, while

14

the recognition of the reasons leading to the initial variety still presents an open problem, even if its importance has been clearly assessed [2]. Diffusion models dealing with the interaction between demand and supply acquire a particular relevance, according to the aim of this study, since they stress the importance of interconnections between economic participants, even if the incorporation into formal models is limited by methodological challenges. Stoneman [3] regards the introduction of interaction between demand and supply into diffusion models as a fundamental progress of recent years: not only does it allow more complete information about the determinants of the diffusion process, but it also represents a first step in relating diffusion both to research and development and to industrial policy. On the part of adoption studies, a growing dissatisfaction towards 'traditional' investment evaluation methods has been growing within economic literature since the beginning of the 1980s. 2 The need to develop different evaluation techniques arises both from the appearance of new technologies connected to automation of production (among other factors) and from the evolution of generic context. Adopters' requirements and traditional evaluation techniques are moving apart because of the widening gap between the environment in which traditional methodologies were suggested at first, and current scenarios, characterized by greater complexity and growing turbulence. Both phenomena have shown the suitability of enlarging the set of factors included in the evaluation process, together with their connections. Gesternfeld and Berger [5] and Kaplan [6] have already stressed the need to consider factors that are not immediately quantifiable, as they play an important role in assessing the novelty of technology compared to the state of the art and environmental conditions. Lambrinos and Johnson [7] provide an example of the strong tie between appraisal methods and the evolution of values within society. Increased sensibility towards workplace safety and health acts so as to include in the evaluation factors able to influence its economic weight. The change of priority among

Technovation Vol. 15 No. 1

Diffusion of technologica/ innovation

different expected benefits implies a different emphasis on the factors related to the attainment of the benefits themselves. The solutions recently suggested in the literature aim at increasing the completeness and thoroughness of appraisal methods. We can recognize two main streams: • quantitative approaches; • strategic-analytical approaches. 'Modified discounted cash flow' methods [8] can be mentioned among quantitative approaches, as they attempt to increase the correctness of traditional methods, including even the so-called intangible aspects, into the assessment of the economic value created through the investment. Other suggestions can be found which try to give a better evaluation of the impact of a new technology through the analysis of linkages among important elements. 3 Both options are useless for technologies in the embryonic stage, since they aim to quantify the economic value created, thus needing available historical data and reliable appraisals of future trends of the relevant variables. Strategic-analytical approaches clearly point to an integrated evaluation of both qualitative and quantitative elements related to technology and environment. Beyond the general methods described in the surveys quoted, more particular solutions stress the trend to refer to different decision techniques at the same time. 4 The main feature arising from the survey of recent literature is the consciousness that some aspects deserve more attention nowadays than they received in the past. Nevertheless, only a few models explicitly consider the dynamics of linkages among relevant elements and the change of their relative critical weights in time. Problems arise even on the side of the significance of innovation for all the economic participants potentially involved. The models suggested in the literature mainly focus on benefits attainable by the decision maker only, who generally coincides with the buyer or user. 5 This seems to be a limit, as even the decision maker's probabilities of success increase when other people involved regard

Technovation Vol. 15 No. 1

the introduction of the new technology as favourable. Under this enlarged perspective, it becomes the decision maker's interest to evaluate the benefits achievable by all the economic participants concerned, i.e. to determine the whole significance of the technology. 6 The study of technological innovation sources (i.e. the people who develop the innovation and the reasons for their behaviour) leads von Hippel [9] to underline the appropriateness of actively looking for the advantages connected to a technological innovation among the different economic participants. His asserted thesis is that "the innovators are those able to foresee in a reasonable way benefits greater than those expected by noninnovators". The only classes of functional sources 7 identified by von Hippel are producers, users and suppliers of technology.

4.

The model

The task faced by the proposed model is to supply a helpful tool to evaluate a technology still in the embryonic stage. It is not possible to estimate future values of important variables on a statistical basis, since the extremely innovative content of such a technology prevents one from referring to previous experience and currently available data seem unreliable. The potential adopter thus faces not risk, but uncertainty. The suggested framework aims at reducing uncertainty by enlarging the set of elements included in the decision process. The criterion used to verify the opportunity of introducing an embryonic technology is the capacity of the adoption, regarded as an investment, to generate economic value for the decision maker. The concept of created economic value indeed represents a proper synthesis between the expected benefits and costs of the adoption. It has already been emphasized that the decision process has to consider not only those elements whose weight can be measured through monetary cash flows, but also the so-called intangible aspects, i.e. those difficult to measure. The concept of

15

F. Sgobbi

enlarged economic value (EEV) is introduced to underline this perspective. It can be defined, for every projected application and for every economic participant interested in its development, as: EEV,,, = f(R)

(1)

where EEVij is the enlarged economic value generated by the ith application for the jth involved economic subject (i = 1,2,...,n; j = 1,2 .... ,m), and R is the vector of expected benefits and costs, related to technology and/or environment: R = [ra, r2

.....

rp]

(2)

As the lack and low reliability of available information represent peculiar characteristics of an embryonic technology, the exact quantification of the economic value generated by the investment seems an unattainable target. The model presented here, the logical scheme of which is depicted in Fig. 1, thus proposes a number of steps useful to evaluate at least the sign, and only incidentally the extent, of the EEV connected with an investment in embryonic technologies---and the suitability of the

TECHNOLOGY AN/) EXTERNAL ENVIRONMENT

_~

EXPECTED ~--~ BENEFITS

EEV = f(13)

i

INDICATORS SYSTEM

COMPARISON BETWEEN EXPECTED/ACTUAL ---~ INDICATORS

I ADOPTION/ ACTIVE POSTPONEMENT REJECTION/ PASSIVE POSTPONEMENT Fig. 1.

16

The proposed model scheme.

investment itself. The proposed approach develops through two distinct stages: • operational definition of the EEV; • identification of a proper set of indicators to judge the suitability of the adoption. The identification of expected costs and benefits related to the introduction of the embryonic innovation is the first step in defining the EEV. From a formal point of view, this means identifying the elements of the vector R as previously defined.8 These elements arise from examining both the considered technology and the environment in which it is supposed to be introduced. Referring to the formalization of a discounted cash flow criterion such as net present value [10], it is possible to give an operational definition for the EEV as the sum of the suitably discounted cash flows generated by the investment, considered as if it were possible to convert all relevant elements into monetary terms: EEVij

,,

ffi,j(gl ..... Rm, t) (1 ~- r,)

(3)

where f f i d ( R 1 , ..., R,,, t) represents the cash flow generated by the jth application in favour of the ith economic subject during the tth year, and r, is the discount rate. The use of the NPV functional form does not aim, in this case, to quantify the economic value created by the embryonic technology investment, since R contains qualitative elements as well. It is justified by formal needs only: indeed, the additive structure allows a high degree of completeness if the single terms can be regarded as independent. The next step is to isolate the relevant factors, which arise from an examination of the sociotechnical context entered by the new technology, beyond the analysis of characteristics of the eventual benefits. 9 This search has a double purpose: first, to identify the most critical factors and, second, to define the connections between different benefits. The identification of factors affecting the expected benefits answers the need to enlarge the set of elements included in the

Technovation Vol. 15 No. 1

Diffusion of technological innovation

evaluation. The links between benefits arise from an examination of the impact which every single factor causes on the expected distribution of distinct advantages. If we indicate the vector of relevant factors as X = [x,, x2, ..., Xh]

(4)

every generic cost or benefit Rk can be expressed as a suitably constructed function referring to a subset of X: R~ = gk(xl, x2 . . . . . xr)

(5)

where x~, x2. . . . . xr belong to X. By substituting (5) within (3), the EEV is now defined as: EEV, j

,,

fft,l(gl . . . . . gin, t) (1 + r,)

(6)

This EEV coincides with the so-called modified DCF criteria when all the important factors are quantifiable, their mutual interactions are well known and even their present and future values are estimable. But the peculiarities of embryonic technologies prevent this from coming true. A set of indicators is then referred to in order to operatively use the values of eq. (6), i.e. to evaluate the suitability of adoption. The indicators should be able to estimate, qualitatively or quantitatively, the present situation and to monitor its evolution over time. An array of parameters can be identified through the 'explosion' of the relevant factors. ~° The different indicators are made up by the parameters themselves, or by proper aggregations. H It should be possible to assess the defined indicators, i.e. to measure them or at least to forecast their trend, ~2 and to monitor their evolution as well. The formalization of the EEV, written for every involved participant, is thus used to build up a system of indicators able to depict the current situation and to monitor its evolution. Both general and specific indicators can be recognized within the set: the former kind is useful for all the applications within a certain area, the latter concerns the indicators peculiar to a specific implementation.

Technovation Vol. 15 No. 1

'Pseudo-formalization' of the EEV, written for every economic participant, allows the decision maker to turn his/her mind towards the selection of capital budgeting. Even if a global appraisal of the potential economic value seems unlikely, it is possible to verify whether the parameters related to critical factors reach the previously defined hurdle levels. It should be noted that the EEVs defined for people other than the decision maker become additional indicators. Their relevance is weighted by the subjective importance given to the various economic entities. When the choice results in adoption of the innovative technology, or even in the definition of incentive measures aimed at further development ('active postponement'), the understanding of critical factors and their links allows measurable targets to be set. The goals of the defined improvements are thus defined and, since they are measurable, they can be controlled later. A particular characteristic of the proposed model is its suitability to monitor a dynamic situation such as that of a technology still in the embryonic stage, developing within a turbulent context. Indeed, the relative relevance of the different expected benefits shifts in time, because of changes both in the technological state of the art and in the external environment. The same happens with the relative relevance of critical factors. Both these phenomena can be monitored through the set of indicators, which can be updated by adding new parameters and deleting the no longer meaningful ones.

5. Non-manufacturing robotic applications By 'advanced robotics' we mean those robots able to collect from the external environment the data needed to operate in ill-structured surroundings, i.e. in deterministically unknown settings. Their field of application is represented by nonrecurring activities, mainly occurring when the important characteristics of the working environment cannot be forecast in advance with certainty.

17

F. Sgobbi

The adoption of advanced robots is thus justified by the need to export automation even outside structured manufacturing activities. The introduction of advanced robotics is particularly meaningful when: • traditional mechanization and automation prevent the enterprise from achieving the desired level of effectiveness and efficiency, thus they are not currently in use; • at the same time, the results obtained through the usual processes, both manual and mechanized, are regarded as unsatisfactory. The analysis of constraints and opportunities for the development of advanced robotics has been the object of research within the Italian National Research Council's 'Progetto Finalizzato Robotica'. Advanced robotics relies upon a set of innovative technologies (machine vision, autonomous navigation, adaptive object manipulation, 'friendly' interaction with users, etc.) which can be called embryonic, since the technical choices for converting them to practical solutions are not yet defined. This scenario is due to the need of both further developing the innovative technologies beyond the current state of the art, in order to increase their effectiveness and efficiency, and facing the problem with a multi-disciplinary approach. Moreover, since the main areas of application are not concerned with manufacturing, they result in being far away from those fields where industrial robotics first developed. The technological uncertainty amounts to both geographical and industrial disconnectedness of the market faced by advanced robotics. The relevance of innovative technologies used within advanced robotics cannot be properly evaluated through the use of 'traditional' quantitative methodologies, because of the relevance of intangible aspects. The applications that are identified are projected not only to increase the productivity of operations through automation--improved safety levels for users, better quality of the product or service supplied by the robot, environmental safeguards are often quoted as expected benefits.

18

The listed elements are certainly able to create value for the economic participants involved, as they answer to needs actually present within the social context, but they are not immediately quantifiable in monetary terms. Beyond this, the totally innovative content of advanced robotics prevents one from directly comparing either the experimental or the operational applications with the previous state of the art: it is not possible to forecast any evolution referring to previous experience. The depicted scenario makes it clear that neither the factors affecting the success of the technology nor their interconnections are clearly defined yet. Indeed, advanced robotics represents a proper field to test the feasibility of the proposed approach to evaluate the relevance of embryonic technologies. It may now be useful to illustrate the method through an example. The proposed example concerns PAM (patient assistant for mobility), a robotic mobile platform projected to adaptively handle bedridden patients within a hospital. PAM is developed by Armstrong Projects Ltd within the UK Department of Trade and Industry's Advanced Robotics Initiative. A comprehensive description of PAM is given by Finlay [11,12]. The witnessing of a similar project by the Japanese company Sanyo Electric Ltd is reported in [13]. PAM, which is now undergoing technology demonstrations, aims to substitute the nurses in the handling of bedridden patients, since it is able to lift the patient from the bed and to transport him/her to the desired area while keeping the patient in the most suitable position. The nurse pilots the robot, but lets it decide about optimizing the path and avoiding collisions. The classes of economic participants affected by the introduction of a similar robot are: • suppliers; • buyers; • direct users, which we can further divide into • patients, • nurses; • public administration.

T e c h n o v a t i o n Vol. 15 No. 1

Diffusion of technologica/ innovation

Expected costs and benefits related to the introduction of the robot can be identified for each of these subjects. In the following example, we focus upon the buyer's choice. This role can be played by the management of either a public or a private hospital facing the decision to adopt PAM. The task performed by the potential buyer becomes quite important in this case, as she/he is directly involved in the economic transmission associated with the adoption, i.e. the purchase of the robot. Two phenomena currently characterize the health service in industrialized countries, where the adoption of a hospital robot seems reasonable. The growing demand for nurses is connected to the general increase in population age, which means more medical treatment for longer periods. This trend joins a parallel reduction in the availability of nurses. The need to handle bedridden patients, which usually requires the presence of more assistants, gets more and more frequent with the growth of the mean age and weight of the hospital population. The introduction of a robot able to perform such a task, repetitive and yet dangerous for both hospital personnel and patients [12], looks like a proper solution to the need both to reduce the hazard of accidents and to amend the lack of nurses, who can thus be free to attend to other tasks that cannot yet be automated. Since only one nurse is now needed to control the robot and to perform the whole task, the simplification of operational flow, together with the reduced complexity of required skills, appear as further benefits related to the adoption of PAM. It is possible to increase the work rate, thus the productivity, of the involved resources. The elements depicted highlight the possibility of delivering a better quality service to the patients, as well as increasing the interest of the medical personnel in their job. The simpler operational flow even implies a fall in the direct costs of patient handling as a consequence of the reduced number of nurses needed for each transfer. Nevertheless, the achievement of the expected benefits involves some incremental costs: these are due to the starting investment and to the

Technovation Vol. 15 No, 1

costs of operation and maintenance. The starting investment concerns the purchase of the robot and the revision of the operational flow. These two aspects interlock, because the optimal number of robots varies according to organizational choices. The reduction of incremental costs depends upon the buyer's interactions with other people affected by the new technology. The initial investment will vary according to public administration incentive policy, while the costs connected with the introduction of PAM within the work flow are affected by nurses' and patients' reactions. Moreover, the upkeep depends upon the supplier's policy, not only because robot reliability conditions the frequency of the maintenance periods, but also because robot availability depends upon the service level chosen by the supplier himself. From this point of view, the presence of the robot increases the buyer's dependence on the external environment. Costs and benefits expected from PAM adoption have been summarized to underline the links between different advantages and between the behaviours of different involved economic participants. The Appendix offers a possible formalization of the EEVs associated with the introduction of PAM into a hospital, together with a set of indicators aimed at monitoring the present state of the art and its evolution. As a proxy, we can regard as independent the contribution of every term to the creation of economic value. In this case, a possible formalization of the EEV created by PAM for the hospital adopting the technology can be written as: EEV = -Pr*n°r - CAA +

Cur,)*n°r]

+ ~ (n'(1 n)*CP, it

J/- r) t

A detailed explanation of the symbols used is supplied in the Appendix. The present target is to highlight how the various addenda embody buyer's costs and benefits as earlier identified. The first two terms represent the starting investment, since (Pr*n°r) represents the cost of the

19

F.

Sgobbi

robots while the second term represents the costs of introducing the robots within the work flow. The third term represents the impact of the adoption on the hospital's operational cost structure, since it is the incremental difference between cash inflows and outflows due to the presence of the robots. The last t e r m represents the capability of the investment to reduce the c o m m i t m e n t s connected not only with operational m a n a g e m e n t , but also with undesirable events such as handling accidents. The terms featured in this equation refer to both tangible and intangible elements. For instance, C A A embodies the expenses due to the modifications in the r o b o t ' s working environment. It includes both m o n e t a r y costs, such as those resulting from the physical modifications of the surroundings, and definitely qualitative elements, such as nurses' and patients' reactions to the introduction of the robot. In the Appendix (Table A1) a set of possible indicators is suggested. These can be used to define the importance and the mutual critical level of the different factors affecting the achievement of the expected advantages.

Acknowledgements I wish to thank Gian Carlo Cainarca and Giovanni Azzone for their valuable suggestions and comments. O f course, I take the responsibility for every mistake or inaccuracy. This work has been developed within the framework of the 'Progetto Finalizzato R o b o t i c a ' , sponsored by the Italian National Research Council.

Notes Surveys of recent contributions to diffusion studies are provided by Stoneman [3] and Gruber [1]. 2 A survey of the main economic justification techniques traditionally used even for innovative technologies can be found in [14]. Beyond the analysis of merits and faults of traditional approaches, surveys of recent contributions in this area can be found in [15-17].

20

3 Jang and Lin [18] estimate the reduction of direct labour hours resulting from the introduction of a generic innovative technology. Schlesinger and Imany [19] search for possible non-linear correlations between the parameters affecting the performance of industrial robots. 4 Examples are given in [20--23]. 5 Cowan [24] proposes a model which departs from this trend, as the economic subject it refers to is a generic central authority facing two different technologies of unknown merit. The model aims to provide the central authority with a decisional tool able to reduce the danger of favouring the worse technology. 6 The innovator eager to introduce the new technology should consider its effects on both horizontal and vertical relationships. The former aspect becomes relevant mainly when network externalities are present, i.e. when the success of the technology is favoured by diffusion among competitors [1]; an example is provided by the importance of defining new common standards. 7 This term refers to participants relating to the innovation through a particular functional role. 8 For example, expected benefits can be process waste reduction, decrease in lead times, observance of environmental regulations, increase in safety and health of the workplace, etc. Additional costs implied by the introduction of new technology are exemplified by changes in the existing work flow, training of employees, etc. 9 For example, consider a technology that allows a better safeguard of workers' health. The factors affecting the value of the potential advantage can be identified as variation in medical and insurance costs, increased productivity of the work force, higher quality, better external image of the adopter, and so on. 10 For example, parameters such as 'mean number of accidents per year', 'mean cost of each accident', 'frequency of different professional diseases', .... affect the factor 'medical expenses variation'. All the parameters have to be evaluated for both the traditional and the innovative technology. 1~ For instance, a meaningful 'aggregated' indicator could be the trend of the ratio between the mean number of accidents happening with and without the new technology. 12 It is not possible to measure quantitatively a parameter such as 'regulation trend', considered as indicative of an important factor such as 'allowed pollution levels'. It is, however, possible to give a qualitative estimation through the use of degrees such

Technovation Vol. 15 No. 1

Diffusion of technological innovation

as 'stationary', 'towards stiffening of current regulations', etc.

References 1 H. Gruber. La teoria delradozione delle innovazioni tecnologiche: una rassegna dei contributi recenti. Economia Politica, IX(l) (April 1992). 2 M. Gibbons and J.F. Metcalfe. Technological variety and the process of competition. Paper presented at: Conference on Innovation Diffusion, Venice, 17-21 March 1986. 3 P. Stoneman. Technological diffusion: the viewpoint of economic theory. Ricerche Economiche, XL(4) (Oct.-Dec. 1986). 4 J.F. Reinganum. On the diffusion of new technology: a game theoretic approach. Review of Economic Studies, 48 (July 1981). 5 A. Gerstenfeld and P. Berger. A model for economic and social evaluation of industrial robots. Proceedings of the 12th ISIR, Paris, 9-11 June 1982. 6 R.S. Kaplan. Must CIM be justified by faith alone? Harvard Business Review (March-April 1986). 7 J. Lambrinos and W.G. Johnson. Robots to reduce the high cost of illness and injury. Harvard Business Review (May-June 1984). 8 G. Azzone and U. Bertele. Techniques for measuring the economic effectiveness of automation and manufacturing systems. Control and Dynamic Systems, 48 (1991). 9 E. yon Hippel. The Sources of Innovation. Oxford University Press, Oxford and New York, 1988. 10 J.J. Clark, T.J. Hindelang and R.E. Pritchard.

11 12 13 14

15

Capital Budgeting--Planning and Control of Capital Expenditures. 3rd Edition, Prentice-Hall, Englewood Cliffs, NJ, 1989. P.A. Finlay. Medical robotics--why, what and when. Industrial Robot 16 (March 1989). P.A. Finlay. PAM: a robotic solution to patient handling. Industrial Robot, 19(3) (1992). J. Hoiligum. Japanese robotics displayed at Science Museum. Industrial Robot, 18(4) (1991). R.E. Terry, R.A. Branting and D.L. Whitman. A critical review of project analysis techniques. In: H.R. Parsaei, W.G. Sullivan and T.R. Hanley (eds.). Economic and Financial Justification of Advanced Manufacturing Technologies. Elsevier, 1992. J.R. Meredith and N.C. Suresh. Justification techniques for advanced manufacturing technologies.

Technovation Vol. 15 No. 1

International Journal of Production Research, 24(5) (1986). 16 J.P. Lavelle and H.R. Liggett. Economic methods for evaluating investments in advanced technologies. In: H.R. Parsaei, W.G. Sullivan and T.R. Hanley (eds.). Economic and Financial Justification of Advanced Manufacturing Technologies. Elsevier, 1992. 17 S. Kolli, M.R. Wilhelm, H.R. Parsaei and D.H. Liles. A classification scheme for traditional and nontraditional approaches to the economic justification of advanced automated manufacturing systems. In: H.R. Parsaei, W.G. Sullivan and T.R. Hanley (eds.). Economic and Financial Justification of Advanced Manufacturing Technologies. Elsevier, 1992. 18 J. Jang and R.C. Lin. Estimating the reduction in labor hours due to a new technology under uncertain demand. International Journal of Production Research, 29(2) (March 1993). 19 R.J. Schlesinger and M.M. Imany. Statistical investigation of robot performance specifications. Human Systems Management, 6 (1991). 20 K. Knott and R.D. Getto. A model for evaluating alternative robot systems under uncertainty. International Journal of Production Research, 20(2) (1982). 21 K. Knott, B. Bidanda and D. Pennebaker. Economic analysis of robotic arc welding operations. International Journal of Production Research, 26(1) (1988). 22. J. Sarkis. The evolution to strategic justification of advanced manufacturing systems. In: H.R. Parsaei, W.G. Sullivan and T.R. Hanley (eds.). Economic and FinancialJustification of Advanced Manufacturing Technologies. Elsevier, 1992. 23. G. Barbioli. A new method to evaluate the specific and global advantage of a technology. Technovation, 10(2) (1990) 73-93. 24 R. Cowan. Tortoises and hares: choice among technologies of unknown merit. The Economical Journal (July 1991).

APPENDIX The proposed formalization of the E E V with regard to the adoption of P A M by a hospital structure is repeated below in order to give a detailed explanation of the symbols used:

21

F. Sgobbi

EEV = -Pr*n°r

-

CAA + ~ [(Ci,

it

-

suggested: these can be used to define the importance and mutual critical level of the different factors affecting the achievement of the expected advantages. As well as for the buyer, a 'formalization' of the enlarged economic value has been written for every identified participant. The formulae result from the analysis of expected costs and benefits. They do not claim mathematical rigour, but are intended as a starting point to the identification of those parameters able to affect the achievement of aimed-for advantages. This is the first step in building up a set of indicators useful for a complete appraisal of the relevance of the application.

Curr)*n°r]

~ ~- r)}-

+ ~ (n' - n)*Cpl

It

(1- _~ ~)t

where EEV Pr n°r CAA

= = = =

Cit Curt n

= = --

Cpi

=

N r

-=

enlarged economic value robot price number of purchased robots cost of setting up the robot's work environment change in direct work costs robot functioning/upkeep costs in year t mean number of handling accidents after robot adoption mean cost per accident sustained by the buyer robot life annual discount rate.

A1. Supplier

EEV = - I +

Primed symbols correspond to the symbols listed above when the robot is not in use. In Table A1 a set of possible indicators is T A B L E A1.

,,

(Fr, Cr,) + ,~ (l+r) t x'

(Rm,- Cm,)

where EEV -- enlarged economic value

A set of possible indicators to assess the relevance of P A M Relevant factors

Indicators

Pr

=

robot price

CAA n°r

= =

costs to fit the environment where the robot works number of purchased robots

Cit

=

change m direct work costs

absenteeism percentage annual lost work hours due to handling accidents annual hours of overtime work change in number of employees due to adoption of robot

Cur,

=

robot functioning/upkeep costs in year t

robot reliability alternative equipment reliabihty

n

22

=

(1 + r)'

trend of public health subsidies (politics) standardizanon level of components price of alternative equipment mean number of training hours number of tasks under nurse's direct control quality of tasks under nurse's direct control

robot availability alternative equipment availability mean number of handling accidents after adoption of robot robot reliability robot availability annual lost work hours due to handling accidents mean patient age mean patient wetght bedridden patient percentage

Technovation Vol. 15 No. 1

Diffusion of techno/ogical innovation

I Fr, Cr, Rmt Cmt N r

= investment needed to enter the new market = new product income in year t = new product direct costs in year t = incomes related to modifications/maintenance in year t = costs related to modifications/maintenance in year t = investment life = annual discount rate.

A2. Nurses N

EEV

EEV gg Cd p, C, Sd

= enlarged economic value (for every hospital treatment) = m e a n n u m b e r of hospital in-patient days after adoption of robot = daily cost of hospital treatment = probability of handling accident after adoption of robot = cost of handling accident sustained by the patient = m o n e t a r y measure of greater satisfaction of patient

Primed symbols correspond to the robot not being in use.

A SALm/

Pi*z~W (1 + rm) t It

A4. Public administration (PA) ~

+ (1 - p ~ ) *

&SALt (1 + r) t

- - +

E E V = - Sr*n°tr +

1t

(1 + r) t

R t

Et , V (l+r)t (p''

+ /_,

p,)+S,

where

where

EEV

EEV Sr n°tr

= enlarged economic value -dismissal probability due to adoption Pl of robot & S A L m , = monthly wage before adoption of robot & S A L t = annual wage after adoption of robot C, = mean work accident cost sustained by the worker in year t p, -- work accident probability after adoption of robot S, = m o n e t a r y measure of greater satisfaction of worker r,,, = monthly discount rate r = annual discount rate. Primed symbols correspond to the robot not being in use.

Pt

T, n

M

R N r

R

~

T,

( 1 + r)'

(n' - n)*C,

= enlarged economic value = subsidy per robot issued by PA = total n u m b e r of robots purchased = pension contributions spared in year t = fiscal incomes lost in year t = mean n u m b e r of handling accidents after adoption of robot mean cost of handling accident sustained by PA mean life left to the nurse substituted by the robot = mean n u m b e r of years to retirement for the nurse substituted by the robot = m e a n robot life = annual discount rate

Primed symbols correspond to the robot not being in use.

A3. Patient E E V = (g-g' - g-g) * Ca + (P', -- P,) * C, + Sd where

Technovation Vol. 15 No. 1

Franeet~m Sgobbi was born in Milan in 1968 and studied at the Politecnico di Milano, where she gained her degree in Industrial Engineering in 1993. She is currently a PhD student at the Economics and Production Department of the same

23

F. Sgobbi

university. Her research interests concern the impact of automation on work organization. Within the framework of the 'Progetto Finalizzato Robotica',

24

sponsored by CNR (the Italian National Research Council), she has written a report dealing with the opportunities and constraints currently faced by advanced robotics.

Technovation Vol. 15 No. 1