Stakeholders from a dynamic and network perspective
NobodyKnows (Talk | contribs) (→Overview of a Stakeholder Analysis today) |
|||
(152 intermediate revisions by 2 users not shown) | |||
Line 1: | Line 1: | ||
− | + | ''Developed by Jens Henrik Lilleskov Nielsen'' | |
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | |||
− | + | =Overview= | |
− | + | The identification of stakeholders has proven as a crucial factor for the success of projects<ref name="nottodo"/>, due to the fact that it helps to give an overview of the current state of the analysed scenario. For this, numerous models have been developed for mapping stakeholders for knowing which, when and how to prioritise them. This type of analysis works as a macro analysis, not only framed to contain a specific area of investigation, but takes into account all the different kind of actors (person, group or organisation) that currently has an interest and that can affect, or be affected, by the given action/project. | |
− | + | This article takes a critical point of view of the current practices, regarding stakeholder analyses. It will be based on the assumption, that not only is the identified stakeholders important, but the way stakeholders are interacting will also be important for the success of a project. Therefore an implementation of Social Network Theory is considered and investigated. This proves advantageously because, it will now be possible to uncover shadow networks and how stakeholders with e.g. low formal power, can still be crucial to take into account, due to their informal power through their interconnection with other stakeholders. Different methods within Social Network Theory concerning centrality algorithms, are discussed and examples are provided, showing how these can be beneficial in e.g. project management. | |
− | + | = Stakeholder Analysis Overview = | |
+ | Conducting a stakeholder analysis, is done with the purpose of getting an overview of different actors that currently has an interest and that can affect, or be affected, by a given action/project. This is useful during projects, due to the fact it will be possible to get an understanding on how much attention different stakeholders should get. When using the term “stakeholder”, it covers a broad range of actors such as; individuals, groups and organisations<ref> Solaimani Sam, Guldemond Nick and Bouwman Harry, (2013), ''Dynamic stakeholder interaction analysis: Innovative smart living design cases'', ELECTRONIC MARKETS, Vol.23(4), PP.317-328</ref>. A stakeholder analysis is what can be characterised as a macro analysis because, it not only takes into account for the organisation or a specific area of investigation, but also takes into account for the external environment. From this, the analysis can be broaden to take multiple levels into consideration, which includes local, regional, national and even international <ref name="nottodo"> Brugha Ruari and Varvasovszky Zsuzsa, (2000), ''How to do (or not to do). . . A stakeholder analy''sis, Health Policy and planning, Vol. 15, PP:239-246 </ref> . This will affect the researcher and how this person will have to collect the necessary data. A “local stakeholder analysis” usually means that the stakeholders are reachable for individual interviews, which can result in more qualitatively data and otherwise the analysis has to use other kind of existing documentation, such as e.g. reports, if interviews are not a possibility. | ||
− | + | Conducting this kind of analysis is therefore to get a more in-depth understanding about the involved stakeholders and their interest, intentions, agendas and their influence or potential resources the project could benefit/dis-benefit from <ref name="review"> Brugha Ruari and Varvasovszky Zsuzsa, (2000), ''Stakeholder Analysis: a review'', Health Policy and planning, Vol. 15, PP:239-246 </ref>. It can be relevant to distinguish between stakeholders by categorising them as primary (crucial for the survival of the project) and secondary (important, but not essential for the survival) stakeholders <ref name="review" />. | |
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | + | ||
− | Conducting this kind of analysis is therefore to get a more in-depth understanding about the involved stakeholders and their interest, intentions, agendas and their influence or potential resources the project could benefit/dis-benefit from <ref name="review"> Brugha Ruari and Varvasovszky Zsuzsa, (2000), ''Stakeholder Analysis: a review'', Health Policy and planning, Vol. 15, PP:239-246 </ref>. | + | |
Line 28: | Line 17: | ||
== Current practices, when conducting a Stakeholder Analysis == | == Current practices, when conducting a Stakeholder Analysis == | ||
[[File:Matrix.PNG|300px | thumb| |'''Figure 1:''' A stakeholder matrix, mapping stakeholders by their power and interest, towards a project.]] | [[File:Matrix.PNG|300px | thumb| |'''Figure 1:''' A stakeholder matrix, mapping stakeholders by their power and interest, towards a project.]] | ||
− | Multiple kinds of stakeholder analyses can be found in the literature <ref name=" nottodo"/><ref name="matters"> Kennon Nicole, Howden Peter and Hartley Meredith, ''Who really matters? A stakeholder Analysis'', Extension Farming Systems Journal, Vol.2 | + | Multiple kinds of stakeholder analyses can be found in the literature <ref name=" nottodo"/><ref name="matters"> Kennon Nicole, Howden Peter and Hartley Meredith, ''Who really matters? A stakeholder Analysis'', Extension Farming Systems Journal, Vol.2(2)</ref> , depending on what kind of aspects is considered most important. Usually the models have in common that they are grid based, usually in two dimensional matrix tables, where this can represent ''power vs. interest'', see figure 1. This is just one example of several methods, where other stakeholder analyses could e.g. incorporate ''power vs. dynamism''<ref>Gardner, J.R., Rachlin, R., Sweeny, H.W.A. (1986) ''Handbook on strategic planning'', John Wiley & Sons Inc. Hoboken, NJ</ref>. Colleting the necessary information can be divided into two subgroups; primary and secondary data gathering <ref name=" nottodo"/>. |
+ | |||
+ | Primary data gathering cover direct interactions between stakeholders and researchers, which includes different kinds of interviews including semi-structured, structured etc., but also the use of focus groups. This type of data gathering is especially well suited for exploring the external environment, where experts can be interviewed and sufficient time can be made, for going into depth of the investigated environment. | ||
− | + | Secondary data gathering is more data oriented e.g. reports, internal documentation and is usually discovered during the semi-structured interviews <ref name="nottodo"/>. Current practices concerning stakeholder analyses are primarily based on primary data gathering <ref name="fine"/>, where it is usually investigated “who is important” and “who will be affected” <ref name="fine"/>. From this perspective it can be concluded that the current practice of a stakeholder analysis, is primarily based on qualitatively data. | |
− | + | For more in-depth description of a traditional stakeholder analysis approach, please go to the following links: | |
* [[Stakeholder Management]] | * [[Stakeholder Management]] | ||
* [[Stakeholder Analysis]] | * [[Stakeholder Analysis]] | ||
Line 38: | Line 29: | ||
<div style="clear:both"> </div> | <div style="clear:both"> </div> | ||
− | == | + | == Disadvantages in the current practices of stakeholder analysis== |
− | Even though a lot of the literature states, when conducting a stakeholder analysis it is crucial it is done from a dynamic and iterative perspective, very little actually states how to do this in practice. | + | Even though a lot of the literature states, when conducting a stakeholder analysis it is crucial that it is done from a dynamic and iterative perspective, very little actually states how to do this in practice<ref name="Reed"/>. The reason for this is due to the fact as a project progress existing stakeholders may change attitude towards the project and also and new stakeholders may emerge, which needs to be taken into consideration. |
− | + | Incorporating emerging stakeholders can be a difficult process, when a specific stakeholder tool has been chosen, especially if the specific model does not apply for the new identified stakeholders. From this, it can therefore be concluded that existing stakeholder analyses are rigid and a consequence, by applying matrix models, is that researchers can potentially be forced to fit stakeholders to the matrix. From the discussed variation of different models, confusion can be common, not knowing which model would be most applicable, or when it should be applied <ref name="Reed"> Reed Mark, Graves Anil, Norman Dandy, Posthumus Helena, Hubacek Klaus, Morris Joe, Prell Christina, Quinn Claire, Stringer Lindsay, ( 2009), ''Who’s in and why? A typology of stakeholder analysis methods for natural resource management'', Journal of Environmental management, Vol.90 (5), PP.1933-1949 </ref>. | |
− | + | From this a stakeholder analysis has to be a more dynamic tool and recognise that not only the identified stakeholders are important, but also the way they are interacting is important. Knowing how stakeholders are interconnected can help to better forecast on the future, regarding stakeholders potential behaviour. Therefore current stakeholder analyses needs to be expanded to take into account for more quantitative data gathering, through e.g. questionnaires <ref name="fine"/>. | |
− | + | Getting an overview of how stakeholders are interacting can help investigating the ''shadow network'', which is the way stakeholders are interconnected through self-organisation, and not from the perspective of the designed network, e.g. an organisational diagram <ref name="change"> Battilana Julie and Casciaro Tiziana, (2013) ''The Network Secrets of Great Change Agents'', Harvard Business Review, PP.62-68</ref><ref name="shaw"> Shaw Patricia, (1997), ''Intervening in the shadow systems of organisations'', Journal of Organisational Change Management, Vol.10(3) PP.235-250</ref>. Incorporating these interactions into a traditional stakeholder analysis will also secure less subjective opinions from the researcher, when a larger quantitative collection of data is considered, regarding power, interest, influence etc.<ref name="Reed"/>. This can create a better understanding of the power structure between stakeholders, especially due to the fact that a traditional stakeholder analysis does not always take into account for the informal power aspect. | |
− | + | ||
− | + | = Combining Social Network Theory with stakeholder analysis = | |
− | + | As mentioned earlier, a stakeholder analysis is primarily based on a qualitatively approach to identify relevant stakeholders from both a present and a future perspective. During projects, multiple stakeholders can at some point show some sort of interest, which either can be positive or negative. This is crucial for project managers, having an idea for what to anticipate and then how to accommodate this before it is too late. Applying Social Network Theory (SNT) into a stakeholder analysis implies that not only are the individual stakeholders important, but also the way they are interacting is important. | |
− | + | The reason why the interactions of stakeholders are relevant for investigation is due to the fact that stakeholders with strong ties are more likely to be able to influence each other <ref name="Natural"> Prell Christina, Huback Klaus and Reed Mark, (2009) ''Stakeholder Analysis and Social Network Analysis in Natural Resource Management'', Society and Natural Resources, Vol.22 PP.501-518</ref> . | |
+ | This kind of influence can be either positive or negative because it can indirectly also illustrate trust, respect, communication, support etc., which all can have a crucial impact for the success of getting the maximal benefit from the identified stakeholders. SNT can therefore help to discover informal power structures between stakeholders, where as formal power usually can be extracted by looking directly into the organisational diagram. The procedure for investigating power structures (looking into power, influence and interest) are explored in section "[[#section| Application of nodes]]". | ||
+ | |||
+ | Identifying the right stakeholders can provide access to information and knowledge. This is due to the fact that knowledge is not only embedded in formal channels, such as books, reports etc., but crucial knowledge can also be discovered through the social interconnections <ref name="Klaus">Prell Christina, Hubacek Klaus, Quinn Claire, Reed Mark, (2008), ‘’''Who’s in the Network? When Stakeholders Influence Data Analysis''’’, Syst Pract Actions Res, Vol 21, PP.443-458.</ref> . From this statement it can be argued that a subjectivist stand, from the epistemology, is taken here<ref name="subj"> Hatch Mary Jo, Cunliffe Ann L., (2006), ''Organization Theory: Modern, Symbolic, and Postmodern Perspectives'', OUP Oxford, ISBN 9780199260218 </ref>. The knowledge can only be brought forward (created) and understood by investigating the interactions of the stakeholders. Therefore identifying highly interconnected stakeholders, and interacting with them, can potentially bring crucial knowledge forward, that otherwise would not have been available. | ||
== Applying Social Network Theory == | == Applying Social Network Theory == | ||
− | The most efficient way for gathering the relevant kind of data would be through questionnaires, which | + | The most cost- and time efficient way for gathering the relevant kind of data would be through questionnaires, which in a higher degree will ensure more quantitative data, that through SNT can be analysed. There exists several different social network analysis (SNA) software, such as UCINET<ref>[https://sites.google.com/site/ucinetsoftware/home UCINET Homepage], accessed 30.11.2014 </ref> and GEPHI<ref> [http://gephi.github.io/ Gephi Homepage], accessed 30.11.2014 </ref>, where most automatically already has incorporated a various portfolio of different mathematical algorithms, which can be used for analysing the data. |
− | Combining a stakeholder analysis and SNT, primarily two types of networks can be of great importance | + | Combining a stakeholder analysis and SNT, primarily two types of networks can be of great importance being able to detect in portfolio, program and project management, which will be described further down in the article <ref name="change"/>: |
* Cohesive Networks, | * Cohesive Networks, | ||
* Bridging Networks | * Bridging Networks | ||
− | Applying SNT can also help to get the whole picture of a stakeholder, due to the fact | + | Applying SNT can also help to get the whole picture of a stakeholder, due to the fact that when otherwise only the qualitatively stakeholder analysis is conducted; it can be very hard for uncovering hidden agendas. Incorporating SNT it is possible from a statistically point of view to uncover hidden agendas because other stakeholders can be asked for their individual opinion regarding each other interest, attitude, influence etc. towards the project. This is where the quantitatively perspective of the collected data really creates value. |
− | === | + | === Application of nodes === |
− | [[File:Nodes.PNG|300px | thumb| |'''Figure 2:''' An internal organisational network, where the node sizes represents the ''"Attitude towards the project"''.]] | + | [[File:Nodes.PNG|300px | thumb| |'''Figure 2:''' An internal organisational network, where the node sizes represents the ''"Attitude towards the project"''. The bigger the nodes, the more positive is a stakeholder towards the project.]] |
− | By applying SNT, stakeholders can be given several different attributes, either assigned by themselves or by stakeholders. These different values (based on a scale) can be illustrated through the size of the nodes | + | By applying SNT, stakeholders can be given several different attributes, either assigned by themselves or by other stakeholders. These different values (based on a scale) can be illustrated through e.g. the size of the nodes, or by colour, which potentially can help detecting risks or opportunities that otherwise would not have been realised. A list of different attributes can be seen below <ref name="fine"/>: |
* '''Themselves:''' Age, knowledge of the project, seniority, Attitude towards the project, interest, influence, power, involvement of the project, who they are communicating with etc.. | * '''Themselves:''' Age, knowledge of the project, seniority, Attitude towards the project, interest, influence, power, involvement of the project, who they are communicating with etc.. | ||
− | * '''Assigned by | + | * '''Assigned by other stakeholders:''' Attitude towards the project, interest, influence, power etc.. |
− | In figure 2 an example has been created for illustrating how a network could like, where the node size represent the “Attitude towards the project”. From figure 2 it could be argued that the project manager (PM) has not spend enough time promoting the project for the IT- and Production department, due to the fact that their attitude is quite low compared to the stakeholders closer to the PM. | + | In figure 2, an example has been created for illustrating how a network could look like, where the node size represent the “Attitude towards the project” (the larger the node, the more positive a stakeholder is towards the project). From figure 2 it could be argued that the project manager (PM) has not spend enough time promoting the project for the IT- and Production department, due to the fact that their attitude is quite low compared to the stakeholders closer to the PM. If these stakeholders were crucial for the project, an obvious solution would be for the PM to reach out to the departments, open-minded, and investigating the reason for the low attitude. Another argument why it is relevant to reach out is because one of the IT employees actually has a high positive attitude, but he is more or less only interconnected with his own department. Taking into a time/ risk perspective he could potentially over time change his opinion. |
+ | Several attributes would also make sense to illustrate through the interconnections (edges) between stakeholders, but this article will only focus on applying colour and changing sizes of the nodes for illustrating important discoveries throughout the article. | ||
<div style="clear:both"> </div> | <div style="clear:both"> </div> | ||
=== Betweenness Centrality === | === Betweenness Centrality === | ||
− | [[File:BN1.PNG|350px | thumb| |'''Figure 3:''' The stakeholder in the center of the network, can act as a bridge between the two groups of stakeholders]] | + | [[File:BN1.PNG|350px | thumb| |'''Figure 3:''' The stakeholder in the center of the network, can act as a bridge between the two groups of stakeholders, represented by different colours.]] |
− | The type of network illustrated in figure 3 can be characterised as a bridging network <ref name="change"/>. Being aware of this kind of network can be very beneficial for a PM during projects. Stakeholders identified in the center of bridging networks are also called gatekeepers/brokers <ref name="fine"/>, which can help the PM in numerous ways. Using gatekeepers can help the PM controlling the type of information and the flow of what should be directed where. This type of stakeholder will usually have strong ties with his interconnected stakeholders, which is very common, due to the fact that as more interactions a stakeholder have, the weaker they will become. Therefore a good starting point for a project would be for the PM to get the gatekeepers support so they can act as ambassadors throughout their network, when communicating the project. | + | The type of network illustrated in figure 3 can be characterised as a bridging network <ref name="change"/>. Being aware of this kind of network can be very beneficial for a PM during projects. Stakeholders identified in the center of bridging networks are also called gatekeepers/brokers <ref name="fine"/>, which can help the PM in numerous ways. Using gatekeepers can help the PM controlling the type of information and the flow of what should be directed where. This type of stakeholder will usually have strong ties with his interconnected stakeholders, which is very common, due to the fact that as more interactions a stakeholder have, the weaker they will become<ref name="Klaus"/>. Therefore, a good starting point for a project would be for the PM to get the gatekeepers support so they can act as ambassadors throughout their network, when communicating the project. |
− | This type of network is a good example of how e.g. | + | This type of network is a good example of how stakeholders with e.g. low formal power actually still can have a strong influence through informal power. Without this stakeholder, the whole network would otherwise decay and consequences of this could be damaging to the success of the project. An example of this could be an employee that has moved to another department, but still has a lot of social interrelations with his previous department. |
The algorithm calculating the betweenness centrality in nodes, are based on counting how many times a stakeholder is the path between two stakeholders that are not directly interconnected (acting as the bridge) <ref name="fine"> Lienert Judit, Schnetzer Florian and Ingold Karin, (2013) ''Stakeholder analysis combined with social network analysis provides fine-grained insights into water infrastructure planning processes'', Extension Farming Systems Journal, Vol.125, PP.134-148</ref> <ref name="Stanley">Wasserman Stanley, Faust Katherine (1994). ''Social Network Analysis: Methods and Applications'', Cambridge University Press. ISBN 9780521387071.</ref>. | The algorithm calculating the betweenness centrality in nodes, are based on counting how many times a stakeholder is the path between two stakeholders that are not directly interconnected (acting as the bridge) <ref name="fine"> Lienert Judit, Schnetzer Florian and Ingold Karin, (2013) ''Stakeholder analysis combined with social network analysis provides fine-grained insights into water infrastructure planning processes'', Extension Farming Systems Journal, Vol.125, PP.134-148</ref> <ref name="Stanley">Wasserman Stanley, Faust Katherine (1994). ''Social Network Analysis: Methods and Applications'', Cambridge University Press. ISBN 9780521387071.</ref>. | ||
Line 83: | Line 77: | ||
=== Degree Centrality === | === Degree Centrality === | ||
− | [[File:CN.PNG|300px | thumb| |'''Figure 4:''' The stakeholder in the center of the network can be seen as a potential important stakeholder, due the many interactions]] | + | [[File:CN.PNG|300px | thumb| |'''Figure 4:''' The stakeholder in the center of the network can be seen as a potential important stakeholder, due to the many interactions, illustrated through colouring.]] |
− | From figure 4 a network has been created, which illustrates what can be defined as a cohesive network <ref name="change"/>. This type of network emerges, when interactions between many stakeholders exist and especially one or few are highly interlinked. These stakeholders can also be stated to have a high degree of centrality. From figure 4 a visual illustration shows that the algorithm can colour the | + | From figure 4, a network has been created, which illustrates what can be defined as a cohesive network <ref name="change"/>. This type of network emerges, when interactions between many stakeholders exist and especially one or few are highly interlinked. These stakeholders can also be stated to have a high degree of centrality. From figure 4 a visual illustration shows that the algorithm can colour the individual stakeholders from the way they are interconnected, where the “pink” stakeholder will be the person with highest degree of centrality. |
− | Overall can this type of network be very important to identify | + | Overall can this type of network be very important to identify because, usually there will be a lot of trust in this kind of network, and the central stakeholders can help to bring a network together towards a project, due to their internal relations. This can again be related back to some degree of informal power. It can therefore be a good suggestion for a PM to approach this stakeholder and create an alliance instead of approaching every single stakeholder himself. As a project progresses these interrelations will probably change and it is therefore important that the network is updated and time is spend analysing possible trends of interrelations. Identifying these highly interconnected stakeholders can also, from a cost- or time perspective, be very efficient because the PM can use the highly interconnected stakeholders for disseminate information throughout the network <ref>Freeman Linton, (1978), ''Centrality in Social Networks Conceptual Clarification'', Social Networks, Vol.1, PP.215-239 </ref>. |
− | This type of algorithm is quite simple structured, due the fact that it works by counting the edges for every single node, and from that it is possible to either colour grade, or alter the | + | This type of algorithm is quite simple structured, due the fact that it works by counting the edges for every single node, and from that it is possible to either colour grade, or alter the size of, the nodes <ref name="fine"/> <ref name="Stanley" />. |
Line 99: | Line 93: | ||
=== Eigenvector Centrality=== | === Eigenvector Centrality=== | ||
− | [[File:Eigenvector.PNG|300px | thumb| |'''Figure 5:''' Stakeholders have been coloured | + | [[File:Eigenvector.PNG|300px | thumb| |'''Figure 5:''' Stakeholders have been coloured illustrating their power, either formal or informal, which need investigation.]] |
− | Taking into account that networks can actually be quite complex, and not as simple as figure 3 and figure 4, more advanced methods can be applied. One example | + | Taking into account that networks can actually be quite complex, and not as simple as figure 3 and figure 4, more advanced methods can be applied. One example is the ''eigenvector centrality'' algorithm <ref name="Stanley" />. This algorithm bases the level of power of not only how interconnected a stakeholder is, but also how interconnected the stakeholder is with other stakeholders with a high interconnection density. This has been illustrated in figure 5. |
− | Looking into figure 5, three stakeholders are of particularly interest; stakeholder one (S1) in ''cluster A'' could | + | Looking into figure 5, three stakeholders are of particularly interest; stakeholder one (S1) in ''cluster A'', could be identified as the manager for the department and stakeholder two (S2) in ''cluster B'' could be the other manager. These two stakeholders are both rated as high powered, due to the fact that they are well interconnected within their own respective networks, but are also the connection between the two clusters. This could therefore illustrate a more formal power scenario. The third stakeholder (S3) is rated as an equal high powered stakeholder as S2, where S3 could be a respected employee, other trust and listen to (informal power). This will be important for a PM to look into because, S2 and S3 could be the essential combination for creating the necessary support so the whole cluster/department would commit to the project. When analysing this kind of networks, it is of course still important to take into account for the designed network, such as organisational diagrams, which directly can illustrate the formal power. |
<div style="clear:both"> </div> | <div style="clear:both"> </div> | ||
− | = Conclusion = | + | = Conclusion= |
− | By implementing a more in-depth quantitative | + | This article has been based on the assumption that stakeholders themselves will not only be important for the success of a project, but also understanding the way they are interacting will be crucial. By implementing a more in-depth quantitative SNA into the current qualitative stakeholder analysis will provide a stronger framework for identifying crucial stakeholders. This will help to get an overview for how to focus attention and for how to categories the most relevant stakeholders. This will help for identifying not only the formal power structure within a project, but also how informal power can influence the success of the project. |
+ | Applying the presented algorithms will give an overview of the more active and communicative stakeholders in the network, which will be beneficial to consider, when developing e.g. communication plans. This is done by analysing the centrality in the network of stakeholders, which will help narrowing down, which will be crucial for approaching and convincing for committing to a project. One should not neglect the importance of already established networks, such organisational structures, which still provides important information regarding e.g. power structures that can be useful for taking into consideration. | ||
− | Possible downsides for combing these methods | + | It is very important to emphasise and recognise that SNA does not replace the traditional stakeholder analysis, but should be applied as a combination of both. This especially concerns the data gathering process. It will be primarily through the qualitatively data collection, e.g. interviews, where the external environment can be explored and gone into depth, where internal perspectives are primarily explored through quantitative data collection. Possible downsides, for combing these methods, are that it can be a more costly and time consuming data collecting process, compared to e.g. large focus groups<ref name="Reed"/>. |
− | + | ||
− | + | ||
= References = | = References = | ||
<References/> | <References/> | ||
+ | |||
+ | [[Category: Stakeholder Analysis]] |
Latest revision as of 13:29, 20 December 2018
Developed by Jens Henrik Lilleskov Nielsen
Contents |
[edit] Overview
The identification of stakeholders has proven as a crucial factor for the success of projects[1], due to the fact that it helps to give an overview of the current state of the analysed scenario. For this, numerous models have been developed for mapping stakeholders for knowing which, when and how to prioritise them. This type of analysis works as a macro analysis, not only framed to contain a specific area of investigation, but takes into account all the different kind of actors (person, group or organisation) that currently has an interest and that can affect, or be affected, by the given action/project.
This article takes a critical point of view of the current practices, regarding stakeholder analyses. It will be based on the assumption, that not only is the identified stakeholders important, but the way stakeholders are interacting will also be important for the success of a project. Therefore an implementation of Social Network Theory is considered and investigated. This proves advantageously because, it will now be possible to uncover shadow networks and how stakeholders with e.g. low formal power, can still be crucial to take into account, due to their informal power through their interconnection with other stakeholders. Different methods within Social Network Theory concerning centrality algorithms, are discussed and examples are provided, showing how these can be beneficial in e.g. project management.
[edit] Stakeholder Analysis Overview
Conducting a stakeholder analysis, is done with the purpose of getting an overview of different actors that currently has an interest and that can affect, or be affected, by a given action/project. This is useful during projects, due to the fact it will be possible to get an understanding on how much attention different stakeholders should get. When using the term “stakeholder”, it covers a broad range of actors such as; individuals, groups and organisations[2]. A stakeholder analysis is what can be characterised as a macro analysis because, it not only takes into account for the organisation or a specific area of investigation, but also takes into account for the external environment. From this, the analysis can be broaden to take multiple levels into consideration, which includes local, regional, national and even international [1] . This will affect the researcher and how this person will have to collect the necessary data. A “local stakeholder analysis” usually means that the stakeholders are reachable for individual interviews, which can result in more qualitatively data and otherwise the analysis has to use other kind of existing documentation, such as e.g. reports, if interviews are not a possibility.
Conducting this kind of analysis is therefore to get a more in-depth understanding about the involved stakeholders and their interest, intentions, agendas and their influence or potential resources the project could benefit/dis-benefit from [3]. It can be relevant to distinguish between stakeholders by categorising them as primary (crucial for the survival of the project) and secondary (important, but not essential for the survival) stakeholders [3].
[edit] Current practices, when conducting a Stakeholder Analysis
Multiple kinds of stakeholder analyses can be found in the literature [1][4] , depending on what kind of aspects is considered most important. Usually the models have in common that they are grid based, usually in two dimensional matrix tables, where this can represent power vs. interest, see figure 1. This is just one example of several methods, where other stakeholder analyses could e.g. incorporate power vs. dynamism[5]. Colleting the necessary information can be divided into two subgroups; primary and secondary data gathering [1].
Primary data gathering cover direct interactions between stakeholders and researchers, which includes different kinds of interviews including semi-structured, structured etc., but also the use of focus groups. This type of data gathering is especially well suited for exploring the external environment, where experts can be interviewed and sufficient time can be made, for going into depth of the investigated environment.
Secondary data gathering is more data oriented e.g. reports, internal documentation and is usually discovered during the semi-structured interviews [1]. Current practices concerning stakeholder analyses are primarily based on primary data gathering [6], where it is usually investigated “who is important” and “who will be affected” [6]. From this perspective it can be concluded that the current practice of a stakeholder analysis, is primarily based on qualitatively data.
For more in-depth description of a traditional stakeholder analysis approach, please go to the following links:
[edit] Disadvantages in the current practices of stakeholder analysis
Even though a lot of the literature states, when conducting a stakeholder analysis it is crucial that it is done from a dynamic and iterative perspective, very little actually states how to do this in practice[7]. The reason for this is due to the fact as a project progress existing stakeholders may change attitude towards the project and also and new stakeholders may emerge, which needs to be taken into consideration.
Incorporating emerging stakeholders can be a difficult process, when a specific stakeholder tool has been chosen, especially if the specific model does not apply for the new identified stakeholders. From this, it can therefore be concluded that existing stakeholder analyses are rigid and a consequence, by applying matrix models, is that researchers can potentially be forced to fit stakeholders to the matrix. From the discussed variation of different models, confusion can be common, not knowing which model would be most applicable, or when it should be applied [7].
From this a stakeholder analysis has to be a more dynamic tool and recognise that not only the identified stakeholders are important, but also the way they are interacting is important. Knowing how stakeholders are interconnected can help to better forecast on the future, regarding stakeholders potential behaviour. Therefore current stakeholder analyses needs to be expanded to take into account for more quantitative data gathering, through e.g. questionnaires [6].
Getting an overview of how stakeholders are interacting can help investigating the shadow network, which is the way stakeholders are interconnected through self-organisation, and not from the perspective of the designed network, e.g. an organisational diagram [8][9]. Incorporating these interactions into a traditional stakeholder analysis will also secure less subjective opinions from the researcher, when a larger quantitative collection of data is considered, regarding power, interest, influence etc.[7]. This can create a better understanding of the power structure between stakeholders, especially due to the fact that a traditional stakeholder analysis does not always take into account for the informal power aspect.
[edit] Combining Social Network Theory with stakeholder analysis
As mentioned earlier, a stakeholder analysis is primarily based on a qualitatively approach to identify relevant stakeholders from both a present and a future perspective. During projects, multiple stakeholders can at some point show some sort of interest, which either can be positive or negative. This is crucial for project managers, having an idea for what to anticipate and then how to accommodate this before it is too late. Applying Social Network Theory (SNT) into a stakeholder analysis implies that not only are the individual stakeholders important, but also the way they are interacting is important.
The reason why the interactions of stakeholders are relevant for investigation is due to the fact that stakeholders with strong ties are more likely to be able to influence each other [10] . This kind of influence can be either positive or negative because it can indirectly also illustrate trust, respect, communication, support etc., which all can have a crucial impact for the success of getting the maximal benefit from the identified stakeholders. SNT can therefore help to discover informal power structures between stakeholders, where as formal power usually can be extracted by looking directly into the organisational diagram. The procedure for investigating power structures (looking into power, influence and interest) are explored in section " Application of nodes".
Identifying the right stakeholders can provide access to information and knowledge. This is due to the fact that knowledge is not only embedded in formal channels, such as books, reports etc., but crucial knowledge can also be discovered through the social interconnections [11] . From this statement it can be argued that a subjectivist stand, from the epistemology, is taken here[12]. The knowledge can only be brought forward (created) and understood by investigating the interactions of the stakeholders. Therefore identifying highly interconnected stakeholders, and interacting with them, can potentially bring crucial knowledge forward, that otherwise would not have been available.
[edit] Applying Social Network Theory
The most cost- and time efficient way for gathering the relevant kind of data would be through questionnaires, which in a higher degree will ensure more quantitative data, that through SNT can be analysed. There exists several different social network analysis (SNA) software, such as UCINET[13] and GEPHI[14], where most automatically already has incorporated a various portfolio of different mathematical algorithms, which can be used for analysing the data. Combining a stakeholder analysis and SNT, primarily two types of networks can be of great importance being able to detect in portfolio, program and project management, which will be described further down in the article [8]:
- Cohesive Networks,
- Bridging Networks
Applying SNT can also help to get the whole picture of a stakeholder, due to the fact that when otherwise only the qualitatively stakeholder analysis is conducted; it can be very hard for uncovering hidden agendas. Incorporating SNT it is possible from a statistically point of view to uncover hidden agendas because other stakeholders can be asked for their individual opinion regarding each other interest, attitude, influence etc. towards the project. This is where the quantitatively perspective of the collected data really creates value.
[edit] Application of nodes
By applying SNT, stakeholders can be given several different attributes, either assigned by themselves or by other stakeholders. These different values (based on a scale) can be illustrated through e.g. the size of the nodes, or by colour, which potentially can help detecting risks or opportunities that otherwise would not have been realised. A list of different attributes can be seen below [6]:
- Themselves: Age, knowledge of the project, seniority, Attitude towards the project, interest, influence, power, involvement of the project, who they are communicating with etc..
- Assigned by other stakeholders: Attitude towards the project, interest, influence, power etc..
In figure 2, an example has been created for illustrating how a network could look like, where the node size represent the “Attitude towards the project” (the larger the node, the more positive a stakeholder is towards the project). From figure 2 it could be argued that the project manager (PM) has not spend enough time promoting the project for the IT- and Production department, due to the fact that their attitude is quite low compared to the stakeholders closer to the PM. If these stakeholders were crucial for the project, an obvious solution would be for the PM to reach out to the departments, open-minded, and investigating the reason for the low attitude. Another argument why it is relevant to reach out is because one of the IT employees actually has a high positive attitude, but he is more or less only interconnected with his own department. Taking into a time/ risk perspective he could potentially over time change his opinion.
Several attributes would also make sense to illustrate through the interconnections (edges) between stakeholders, but this article will only focus on applying colour and changing sizes of the nodes for illustrating important discoveries throughout the article.
[edit] Betweenness Centrality
The type of network illustrated in figure 3 can be characterised as a bridging network [8]. Being aware of this kind of network can be very beneficial for a PM during projects. Stakeholders identified in the center of bridging networks are also called gatekeepers/brokers [6], which can help the PM in numerous ways. Using gatekeepers can help the PM controlling the type of information and the flow of what should be directed where. This type of stakeholder will usually have strong ties with his interconnected stakeholders, which is very common, due to the fact that as more interactions a stakeholder have, the weaker they will become[11]. Therefore, a good starting point for a project would be for the PM to get the gatekeepers support so they can act as ambassadors throughout their network, when communicating the project.
This type of network is a good example of how stakeholders with e.g. low formal power actually still can have a strong influence through informal power. Without this stakeholder, the whole network would otherwise decay and consequences of this could be damaging to the success of the project. An example of this could be an employee that has moved to another department, but still has a lot of social interrelations with his previous department.
The algorithm calculating the betweenness centrality in nodes, are based on counting how many times a stakeholder is the path between two stakeholders that are not directly interconnected (acting as the bridge) [6] [15].
[edit] Degree Centrality
From figure 4, a network has been created, which illustrates what can be defined as a cohesive network [8]. This type of network emerges, when interactions between many stakeholders exist and especially one or few are highly interlinked. These stakeholders can also be stated to have a high degree of centrality. From figure 4 a visual illustration shows that the algorithm can colour the individual stakeholders from the way they are interconnected, where the “pink” stakeholder will be the person with highest degree of centrality.
Overall can this type of network be very important to identify because, usually there will be a lot of trust in this kind of network, and the central stakeholders can help to bring a network together towards a project, due to their internal relations. This can again be related back to some degree of informal power. It can therefore be a good suggestion for a PM to approach this stakeholder and create an alliance instead of approaching every single stakeholder himself. As a project progresses these interrelations will probably change and it is therefore important that the network is updated and time is spend analysing possible trends of interrelations. Identifying these highly interconnected stakeholders can also, from a cost- or time perspective, be very efficient because the PM can use the highly interconnected stakeholders for disseminate information throughout the network [16].
This type of algorithm is quite simple structured, due the fact that it works by counting the edges for every single node, and from that it is possible to either colour grade, or alter the size of, the nodes [6] [15].
[edit] Eigenvector Centrality
Taking into account that networks can actually be quite complex, and not as simple as figure 3 and figure 4, more advanced methods can be applied. One example is the eigenvector centrality algorithm [15]. This algorithm bases the level of power of not only how interconnected a stakeholder is, but also how interconnected the stakeholder is with other stakeholders with a high interconnection density. This has been illustrated in figure 5.
Looking into figure 5, three stakeholders are of particularly interest; stakeholder one (S1) in cluster A, could be identified as the manager for the department and stakeholder two (S2) in cluster B could be the other manager. These two stakeholders are both rated as high powered, due to the fact that they are well interconnected within their own respective networks, but are also the connection between the two clusters. This could therefore illustrate a more formal power scenario. The third stakeholder (S3) is rated as an equal high powered stakeholder as S2, where S3 could be a respected employee, other trust and listen to (informal power). This will be important for a PM to look into because, S2 and S3 could be the essential combination for creating the necessary support so the whole cluster/department would commit to the project. When analysing this kind of networks, it is of course still important to take into account for the designed network, such as organisational diagrams, which directly can illustrate the formal power.
[edit] Conclusion
This article has been based on the assumption that stakeholders themselves will not only be important for the success of a project, but also understanding the way they are interacting will be crucial. By implementing a more in-depth quantitative SNA into the current qualitative stakeholder analysis will provide a stronger framework for identifying crucial stakeholders. This will help to get an overview for how to focus attention and for how to categories the most relevant stakeholders. This will help for identifying not only the formal power structure within a project, but also how informal power can influence the success of the project.
Applying the presented algorithms will give an overview of the more active and communicative stakeholders in the network, which will be beneficial to consider, when developing e.g. communication plans. This is done by analysing the centrality in the network of stakeholders, which will help narrowing down, which will be crucial for approaching and convincing for committing to a project. One should not neglect the importance of already established networks, such organisational structures, which still provides important information regarding e.g. power structures that can be useful for taking into consideration.
It is very important to emphasise and recognise that SNA does not replace the traditional stakeholder analysis, but should be applied as a combination of both. This especially concerns the data gathering process. It will be primarily through the qualitatively data collection, e.g. interviews, where the external environment can be explored and gone into depth, where internal perspectives are primarily explored through quantitative data collection. Possible downsides, for combing these methods, are that it can be a more costly and time consuming data collecting process, compared to e.g. large focus groups[7].
[edit] References
- ↑ 1.0 1.1 1.2 1.3 1.4 Brugha Ruari and Varvasovszky Zsuzsa, (2000), How to do (or not to do). . . A stakeholder analysis, Health Policy and planning, Vol. 15, PP:239-246
- ↑ Solaimani Sam, Guldemond Nick and Bouwman Harry, (2013), Dynamic stakeholder interaction analysis: Innovative smart living design cases, ELECTRONIC MARKETS, Vol.23(4), PP.317-328
- ↑ 3.0 3.1 Brugha Ruari and Varvasovszky Zsuzsa, (2000), Stakeholder Analysis: a review, Health Policy and planning, Vol. 15, PP:239-246
- ↑ Kennon Nicole, Howden Peter and Hartley Meredith, Who really matters? A stakeholder Analysis, Extension Farming Systems Journal, Vol.2(2)
- ↑ Gardner, J.R., Rachlin, R., Sweeny, H.W.A. (1986) Handbook on strategic planning, John Wiley & Sons Inc. Hoboken, NJ
- ↑ 6.0 6.1 6.2 6.3 6.4 6.5 6.6 Lienert Judit, Schnetzer Florian and Ingold Karin, (2013) Stakeholder analysis combined with social network analysis provides fine-grained insights into water infrastructure planning processes, Extension Farming Systems Journal, Vol.125, PP.134-148
- ↑ 7.0 7.1 7.2 7.3 Reed Mark, Graves Anil, Norman Dandy, Posthumus Helena, Hubacek Klaus, Morris Joe, Prell Christina, Quinn Claire, Stringer Lindsay, ( 2009), Who’s in and why? A typology of stakeholder analysis methods for natural resource management, Journal of Environmental management, Vol.90 (5), PP.1933-1949
- ↑ 8.0 8.1 8.2 8.3 Battilana Julie and Casciaro Tiziana, (2013) The Network Secrets of Great Change Agents, Harvard Business Review, PP.62-68
- ↑ Shaw Patricia, (1997), Intervening in the shadow systems of organisations, Journal of Organisational Change Management, Vol.10(3) PP.235-250
- ↑ Prell Christina, Huback Klaus and Reed Mark, (2009) Stakeholder Analysis and Social Network Analysis in Natural Resource Management, Society and Natural Resources, Vol.22 PP.501-518
- ↑ 11.0 11.1 Prell Christina, Hubacek Klaus, Quinn Claire, Reed Mark, (2008), ‘’Who’s in the Network? When Stakeholders Influence Data Analysis’’, Syst Pract Actions Res, Vol 21, PP.443-458.
- ↑ Hatch Mary Jo, Cunliffe Ann L., (2006), Organization Theory: Modern, Symbolic, and Postmodern Perspectives, OUP Oxford, ISBN 9780199260218
- ↑ UCINET Homepage, accessed 30.11.2014
- ↑ Gephi Homepage, accessed 30.11.2014
- ↑ 15.0 15.1 15.2 Wasserman Stanley, Faust Katherine (1994). Social Network Analysis: Methods and Applications, Cambridge University Press. ISBN 9780521387071.
- ↑ Freeman Linton, (1978), Centrality in Social Networks Conceptual Clarification, Social Networks, Vol.1, PP.215-239