1 Visions sustained experimentalism for industrial clusters development - from the “Porter’s diamond” to the “organic space” Working note by P. Defourny (Paper Draft – November 2004) Introduction Beginning of 2002, a mission of creating industrial clusters has been initiated in the province of Hainaut1. The proposed target was hopefully to create a sustainable growth for the enterprises of the subcontracting sector, including both the traditional technologies as well the ICT sector. Slightly more than a million people live in this province located in the French-speaking part of Belgium. Some fairly big production centers of international companies own the major asset of the exporting activities, in sectors like steel production, chemical products, glass, cement, heavy transport and industrial equipments, aeronautic and aerospace components, … Around those major production centers, the targeted sectors for applying the clustering strategy were embedded in a diffuse cloud of numerous small companies. In the province itself, the number of enterprises potentially concerned by a cluster action, was hundreds either thousands by assuming that interactions would occur with enterprises of the nearby surrounding areas. The cluster team being made only three people, an initial question was how a so small team could significantly handle a so big number of enterprises. A survey of the literature did provide rational methodologies about how this could be achieved. In example, “value chain” analysis would bring insides to create alliances and enterprises clusters, based on assets complementarities and optimizations. However we activated a great number of conferences, meetings and workshops on the subject, the rational way revealed itself to be rather inefficient to actually initiate company’s interactions, hence clusters initiation. A coach specialized in human resources management underlined that “clusters building” was probably similar to “team building”, the key being in this case creating visions instead of rational. We looped on the following sequence “visions – inferred actions – results benchmarking” and pursued only the vision that gave the best results. We never checked whether any vision would actually correspond to any scientific knowledge or actual reality, but we just retained the ones inducing motivations and progresses. The following paragraphs describe how the visions and the results gradually evolved. Selecting the initial vision At the beginning, we found that an amazingly big number of visions were co-existing “about what a cluster is”. We took an experimentalist methodology to extract our first best vision, by simply testing several ones and benchmarking the business results achieved after a six months period. 1 In the frame of the Objectf 1 Phasing out, with the support of the FEDER and the Walloon Region founding 2 We observed that the best efficient behavior, by far, was obtained by the companies building up a free relational network with complementary enterprises, with the explicit target to obtain the more direct business increases. By a complete absence of activity overlap, those companies could easily join their competences in multidisciplinary market opportunities. Such groups were naturally outwardly oriented, opened to any new opportunity and could easily exchange information about their clients needs. Just by meeting and talking, five to six “unexpected” business results have been obtained in less than six months, in a group of more or less fifteen enterprises. We also observed that those companies indicated a trend to naturally fill up the “Michael Porter Diamond” diagram, i.e. by tending to incorporate in they relationship network, any stakeholders in relation with their targeted market, such as other sector’s companies, institutions, investors, … We decided to pursue this track and to work out a bit further the “Porter” vision, in order to obtain a more acute version. Digging the “diamond” vision : an instability at growth The Porter Diamond is a system similar to a four groups of components interacting into a field2. If a component is excited, it induces the activation of at least one other component of the system. In such system, the field may induce feedbacks between elements that are otherwise independent. Under certain conditions, feedbacks may induce self-sustained cycles or motions that self-amplify. Conditions at which this happens are known as “instability conditions”. Under those conditions, any unexpected motion of one component may induce motion of others, feedbacks making global motions growing and achieving a self sustained level. Important also is the fact that when instable motions initiate, they often become non predictable. You can set up the instability conditions but you can not predict what the motions will be. By analogy, we took the vision that we had to install the instability conditions and not focus on particular actions, those ones being probably non predictable3. Obtaining self-amplification and self-sustainability Not knowing exactly what the instability conditions could be in practice, we assume that the “field” could be induced by a common demand or client type and that the independency would be reached by complementarities (i.e. no overlap between the enterprises activities). We assumed also that the instability conditions would be reached by increasing the number of components4 to a certain level, providing that some of them occur in several instantiations. 2 A field is a concept in physics that describes a kind of force that is distributed everywhere over a certain space. The presence of a field may induce interactions between elements that are otherwise independent of each other. I.e., the electro-magnetic field makes that interferences may build up between a fix and a mobile phone. 3 By not concentrating on particular actions, this option gave us a strategy different of many other clusters that concentrate on particular activities and services organisations. 4 Components = entreprises, institutions, etc. 3 With the criteria of common specific clients, we selected different type of enterprises groups, invited them to meet and tried to coach them in different way such that each group would eventually growth their number of “associates”. The only criteria of success we had, was “if we stop our coaching action, the “good” systems would self-amplify and self-sustain their own actions” by themselves. Several groups did not pass the test and their activity stopped with our retrieval. Fortunately, a few groups actually did get trough the test and gradually increased their “components” numbers and their business volume5. The companies clustered in such a groups, were not having big official meetings either conferences, but their interactions gradually became part of their day-to-day activities. For those successful groups, the results growth appeared as driven by a kind of stochastic mechanism. As the people said, “successes appeared to happen by chance. While the network grew, the “chance” was increasing too over time, exhibiting so a kind of amazing selfamplification”. We diagnosed for those groups that, by increasing their associate’s number adequately, they first increased the positive externalities occurrences while in the mean time they increased the number of externalities to which they were able to answer. They created a reinforcing loop between the probabilities to find a new opportunity and the probability to be able to answer it. This is an important factor as it creates on each enterprise a demand that can be higher than the mean market demand on an isolated company. Because the external origin of any initiated actions, we effectively loose predictability about what a group could actually do, however we become able to give a probability that something positive would happen or not for a given group. Transforming the “diamond” into the “soccer team” vision The “instable system” vision is actually not accurate to be popularized and to create a simple adhesion on how the groups were behaving. We did need a more accessible vision in order to better promote and enhance the cluster construction. Several alternative visions have been tested. By far the one of a “soccer team” received the highest agreement of the involved players, this vision being actually enhancing many aspects of each group behavior and management style. First of all, a team is often initiated by one or two leaders that like to play together. Scouting, selecting and playing some training matches will be necessary to complete and strengthen their cooperation prior to enter challenging confrontations. Often, they need to create a pool that will gather more than one player at every position. According to the challenge to face, they may prefer one player to an other, but still keep both in the team. Obviously, like in any soccer team, the key is to maintain cooperative strength during the match, while competition exists in between. 5 Most of the groups that succeed did not promote friendship and collegiality per se. They embedded one or two leaders who believed that their personal vision would be better achieved by growing the group. Their leadership did progressively drive the others to regular actions. 4 A good team attracts good players, specializes itself, can explain what are the matches they perform best, etc. Particularly, we became able to create an image of what each team was doing, without explaining how he was internally organized. This gave a break into the identity building process. Many clusters tend to explain who is in the cluster, everyone’s company profile, how the cluster is organized, but still the potential customer does not see its advantage. By having an image of what a team provides, we unified people around an outwardly driven goal, even if some of the people were also playing with other team. The people were not unified by a group ownership, but by a match delivering. It became also more understandable that the clusters were not organizing big meetings and conferences. A football team gathers mainly for a match preparation or to play a match. The cluster support role became also better defined and mainly specified to new player’s detections, intermediations, coaching and visions building. All the other clusters specific actions were actually at the initiative of the players. In particular, the teams became forced to innovate in order to maintain their competitiveness. There are still many other aspects to report with the installation of the soccer team vision. In example, we should mention that people like to work with some other that become friends, in the same way that soccer players do prefer some partners to others. Because of this, you might be a good player, yet not accepted in a given team. You either need to find an other team or to initiate yourself a new one. This gave us the only unexpected result which has been the self-generation of other new groups. People looking at those visions, did adopt the same track, but with their own team members. This later aspect was taken as important one as it demonstrated that for a given and constant cluster support, the cluster building might become a self amplifying process. Over a certain member’s number threshold, we actually observed that a group might enter a self growing phase. Similarly, over a certain “soccer team’s” number, the global economy should also enter a self growing phase. Self-organization in complexity: the “organic cell” By following the vision of installing instability conditions in “Porter Diamonds” and its operational translation into the soccer team’s vision, we first worked out the enterprises of the ICT sector. We obtained, in a 12 months period, around 10 groups activated, each group being made of between 5 to 10 companies. The more mature groups did start to incorprated collborations with R&D facilities and private or public institutions, demonstrating an actual “bottom-up” evolution. We looked at extending the process to the other technological sectors, more traditional as well as advanced technologies. As per previous experiences, we initiated more and more groups of companies having complementary actions in front of a particular client type. By doing so, the numbers of groups increased, as well as the people involved in several groups. Each team behaving mostly by itself, we enter a real world of incomplete information and we enter an actual complexity as several players did play with several team, several being even unknown to us. 5 Up to five to six more groups, the situation was still “under control”. Over those numbers, things became “unmanageable”, but above ten … eleven, we manage to create a view that made the global situation understandable. We realizede that the teams were in fact embedded into market cells, each of them being defined by the international market they were supplying. This global vision was surprisingly embedding the local cells into a comprehensive space map where we could apparently draw all the potential interactions that could be activated between the companies, no matter whether they would be local or international ones. One of those diagrams is presented in the nearby figure. It represents a self-organized industrial schema of the region. We named this diagram the “regional organic cell”. How to draw the “organic vision” To understand the principles for drawing such a diagram, you have first to consider that the locally restrained activities will be gather in central infield and that a move towards the outside is similar to progressively leaving the area. The international industrial world is located at external part of the graph. The first principle is to locate in the middle infield, the companies and the organizations which are actually acting locally and/or that participate to more than one value chain (i.e.: general servicing companies, standard mechanical workshops, schools …). In the first belt around the infield, we place the companies that mainly supply an international specific business as tier 1 or 2 (i.e.: the glass industry, the automotive industry, the medical and surgery activities …). This belt gathers most of the pro-activity and the innovation outfits of the region. The international clients are placed into the external white belt. The ones that are close to the region are placed more inside and the ones that are faraway can be placed more to the outside. Finally, at the external part, we draw a second belt that replicates the pro-active belt of the region. We are placing there the other major concentrations of “tier 1 – tier 2” suppliers that may exist in the world (i.e.: for the medical equipment: Tüttlingen in Germany, Pakistan …). An organic diagram has to be compared to a map. In such a description, you can imagine all the combined actions sequences than can happen in a given market. In such a map, a company can record the knowledge she has from its environment and investigate the collaboration sequence she would like to activate. This vision told us that in a complex environment like ours, clusters are not simply a few independent diamonds of Michael Porter, but actually that clusters are made of actual interactions between clusters, the organic diagram allowing to “understand” how those interactions would take place. 6 The organic diagram for coaching of the ICT and the subcontracting sectors clusters. First of all, the organic cell diagram allowed us to locate, in an understandable manner, the cloud of companies of both the ICT and the subcontracting sector. A few of those companies are endogen and specialized with regards to a particular market. They will be locating on the inner “pro-active belt”. Most of the endogenous assets that are non-specialized with regards to a particular customer type, are taking place in the inner field. The international companies, specialized or not, can also be located easily in the graph. With such a picture, the sorting is efficient enough to understand easily where potentially interacting companies could be found. To come back on the “soccer team analogy”, we started to compare each cell of the diagram to a championship. We became able to understand how weak or strong we were in the region, whether and how we had to incorporate local or foreigner players in the teams. We could drive better the mix of technologies that might supply a given market. By the use of this diagram, the methodology that allowed us to obtain positive results with the ICT sector, has been applied to other sectors like the supply of equipments for the glass industry, the supply of equipments for mining and minerals, the appliances, decoration and construction furniture production. The first economical results with those “more traditional sectors” have been obtained very quickly, within two to five months. The health care equipment production has been kicked off. This sector exhibiting a higher diversity than the other sectors, the first meeting did not induced results as quickly as expected. At this stage, we did redraw an organic cell describing only the medical sector, each market segment becoming a particular cell. The organic diagram exhibited so, properties close to the self-similarity: the same kind of figure can be used to figure out the whole as well as detailed enhancements. Check is actually performed to see whether this new version allow to induce positives results as well in this more complex sector. Conclusions In the field of industrial clusters building, the literature has produced a large amount of documentation. The subject is seen as a very complex one, however it did occur spontaneously in some place. One of the key questions is certainly how to establish such a mechanism so that it becomes selfgrowing and self-sustained. Our experience suggests that a few simple visions and slight coaching actions can induce selfgrowing mechanisms.