An explorative study on the role of silo-communication on creativity and efficiency during innovation projects.

(1)

“Outside the Silo”

An explorative study on the role of silo-communication on creativity and efficiency during innovation projects.

T.J. Biesterbos, University of Groningen - The Netherlands

In this research we investigated the role of silo communication on creativity and efficiency in social networks during innovation projects in a single Dutch company. Functional and divisional silo communication was expected to negatively influence creativity and to stimulate efficiency. Progress in design phase would shift emphasis from creativity in early stage to efficiency in later stage. The findings showed that divisional silo communication stimulates creativity in early stages of design and later decreases in relevancy. The influence of silo- communication on efficiency did not show as relevant as expected, but did slightly increase in strength during late design stages. The influence of silo communication on efficiency and creativity showed to have limited relevancy in the empirical setting used.

I

NTRODUCTION

What is the role of silo communication on creativity and efficiency in social networks during innovation projects?

A high number of organizations take a project-based approach to develop their new business ideas. Often, these projects involve members of a variety of skills and business divisions, crossing intra-organizational boundaries to creatively develop their innovations. Structuring and organizing these innovation processes is inevitable when working with these groups of people. In profit-oriented environments, labour hours and time-to-market will need to be reduced as much as possible to lower costs and to strive for efficiency. But when maximizing their business imperatives like productivity and efficiency, the team’s creativity can be undermined in these kind of work environments (Amabile, 1998). As creativity and idea forming are an approach to work that leads to the generation of novel and appropriate ideas, processes, or solutions (Amabile, 1996;

Ford, 1996; Shalley, 1991), these ingredients cannot be missed in any business development effort. A proper alignment of efficiency as well as creativity is needed.

Team communication is indisputably connected with this theme, as we will explain in the next section.

(2)

Team communication in practice: the silo-effect explained

Working on innovation projects and collaborating with others means communication between project team members and other groups that have the information one needs to complete development tasks. When actors involved in an certain innovation project perform this interaction of information, they communicate with each other in a certain pattern: a social communication network.

A social network can be described as the pattern of linkages and the relationships built through exchanges (Nahapiet & Gloshal, 1998), in this case the exchange of information and knowledge about new ideas, new business opportunities, new processes or products. By measuring these exchanges, we can identify the rather informal structures that are not found in formal organizational charts (Cross, Borgatti & Parker, 2002). This is particularly true for communication during innovation projects as when doing new things and ‘creating the unknown’, one – obviously – does not yet have a structured communication pattern in the first place. These patterns of communication among the team members and organizational structures may have an interesting role on the creativity and efficiency of the project’s result.

Let’s use an example of an organization consisting of 4 business divisions. This company uses a project-based approach to innovation and is currently working on a project. The project team involves members from different divisions and different functional areas, and the members are communicating –at this moment- with each other in a certain pattern (see figure 1).

As shown in the example, some actors involved in the project are communicating with members of the same functional area, and have more links with them (resembled by the black arrows)

Division A

Division B

Division C

Division D

= Function 1

= Function 2 Key

Figure 1 - Communication network example

(3)

than with the rest of the team or with members of their own division. In this way, they may be more ‘embedded’ into the group of people with the same divisional or functional area, possibly because of communication about their tasks or knowledge. We define this tendency to communicate more intense inside a group of people with the same characteristic compared with the people outside this group as a communication silo. In this research we will focus on communication silos of divisions and functional areas, as this resembles a matrix-type organization approach which is widely used in businesses today. Also these two types of silos are important because of the emerging effort to be more customer-oriented instead of product-oriented: an effort that can lead to improved business successes (Day, 2006).

By communicating more inside this silo, a project member may lack communication between others outside the silo that may regard a critical design issue, resulting in incompatibility. Also, to come up with the more radical ideas, out-of-the-box thinking can only be done in full potential by crossing the traditional boundaries of a division or functional area. Research shows that we feel closest to others who are in close physical proximity to us (Allen 1977; Kiesler and Cummings 2002). Gulati (2007) states that breaking down silos can increase coordination and cooperation, develop connections with external parties and can induce capability development. Eesley & Longecker (2006) found that poor communication and structural silos (including divisional and functional silos) can prevent the information flow of useful information, causing intrapreneurship (including the development of new business ideas) to suffer greatly. Powerful product, country, and functional silos are jeopardizing companies' marketing efforts as each silo has only part of the picture (Alberg, 2007). Matrix organizations (structuring both on divisions and functions) can be challenged with these silo-focused employees (Sie &

Annunzio, 2005).

(4)

The dynamic aspect of innovation projects

Our example in figure 1 shows the links between team members on a certain moment in time. As the nature of all development projects is very dynamic, this communication network will change over time. Myers and Marquis (1969) provide a generic model that distinguishes three phases in innovation processes: idea generation, problem solving and implementation. Thusman (1977) adapted this model to focus on which type of communication is needed at what particular phase: communication depends on the nature of the tasks to be completed for that phase.

The idea generation phase involves front-end activities like questioning the status quo, the emergence of a problem or dissatisfaction with the current state of affairs (Dasgupta, 1996). Preliminary market, technical and business assessments are performed with spending very limited time and resources (Cooper, 1997) as an input for the screening decision to proceed with the innovation. In this first phase there should be communication between people form different backgrounds and functions to complete the task of generating a valuable, useful and original idea.

In the problem solving phase, the concept is being researched and developed. A technical solution is developed by engineers (Myers & Marquis, 1969). Detailed investigations take place and the business case is being built. Being the last phase before actual development and “spending the money”, a justification for the investment must be made (Cooper, 1997) as well as a detailed plan for the implementation. To complete these tasks, there will be strong communication between functional experts like the marketing and engineering staff. In complex environments, a lot of cross-functional communication is also expected, because of the high level of iterations needed.

The implementation phase is there to implement the project as designed in the previous one. For the implementation or ‘launch’ stage, the marketing capabilities like market investigation, market testing and promotion can be dominant communication needs

(5)

(Adams, Bessant, & Phelps, 2006). Also, specific organizational skills and tasks can be required in case of a process innovation. Groups of people will be assigned a specific and clear task, are able to work efficiently and will likely communicate more intense with their co-assigned peers than with the rest of the co-workers.

The communication pattern changes when the innovation project matures because of different tasks during the different phases in design. The agile nature of the early innovation project stages becomes less uncertain and will be easier to plan and structure later in the project. Risk and uncertainty are reduced (Young, 2007). Project teams might be formed along the way to facilitate the structure and will be dividing their tasks and set goals for the further project development. So the relative importance of creativity is dominant in early stages or the ‘front end’, whereas efficiency becomes more important in the later stages of design. The silo-effects may therefore be harmful for creativity in the early stages, but can be positively influencing efficiency in the later stages.

H

YPOTHESES

Silo communication on divisional and functional level is expected to influence both creativity and efficiency. The dynamic effects of proceeding through the development process will influence these relations because the nature of the different tasks required in each phase, making that effect weaker or stronger depending on the need for creativity or efficiency.

If people communicate stronger with peers within their silo compared to outside their silo, they will be able to structure their tasks and goals more efficient. Because they already work together on a day-to-day basis, they share the same knowledge and skills and may have a small physical or mental distance to each other. The communication links that already exist are used instead of communicating outside the silo where it would take much more time and effort to get the same result.

(6)

Silo-communication Efficiency

Hypothesis 1a: The silo-communication strength of divisional and functional areas will be positively related to efficiency.

In later design stages where efficiency is more important and it is possible to coordinate and assign tasks, the influence of silo-communication on efficiency is expected to become stronger compared to early design stages - where creative and out of the box thinking is the goal. The increase in design stage makes this positive relation between silo effects and efficiency stronger.

Silo-communication Efficiency

Progress in design phase

Hypothesis 1b: The relation between the two silo-communication factors and efficiency will be stronger if moderated by the phase of the design project; the further in the development process, the stronger this relation.

The more people from different backgrounds will communicate, the more creative the outcome is expected to be. To think out-of-the box and create the radical and innovative ideas, we need input from multiple different sources and take a high distance from day- to-day business practices. Strong communication inside a functional or divisional silo where the differences between people are smaller, will therefore negatively influence creativity. Also, when forming project-teams that consist of members from the same division or function, we can expect this effect.

Silo-communication Creativity

+

_ +

(7)

Hypothesis 2a: The silo-communication strength of divisional and functional areas will be negatively related to creativity.

Because of the changing nature of innovation projects, less creativity is needed in later stages of design. Tasks can be more structured, and falling back into the silo will not obstruct the development process. Creativity is still needed for a successful launch but on a more incremental or small basis. This kind of creativity can be found inside a silo because of the clear task structure in the problem solving and implementation phase.

Communication between members of the same silo will be enough to complete the innovation tasks for that phase and will not lower creativity as much as in the front end of the development process.

Silo-communication Creativity

Progress in design phase

Hypothesis 2b: The relation between the two silo-communication factors and creativity will be stronger if moderated by the phase of the design project; the further in the development process, the weaker this relation.

To test these four hypotheses in an empirical setting we used the method as explained in the next section.

M

ETHOD

A single medium sized divisional company with about 400 employees was used for this research. First, a round of explorative interviews was taken from the management of each division and the senior management to gain insight in the dynamics of innovation project efforts at the company. Questions were asked on ‘what projects are completed or in development’ and ‘who is involved in these innovation projects’. Information about

_ _

(8)

the organizational structure (including the divisions and functional areas) was also taken from these explorative interviews.

The second step involved taking a survey to assess who of the company’s R&D team members were involved in which project and at what stage of design (being ‘idea generation’, ’problem solving’ or ‘implementation’). For each project they were asked whether they participated in any way during each of the three phases. Also, their functional area and division they were working for were identified. A brief project description was given as well as background information on design processes to be sure that the subject could identify the design phases they had been working on during the projects and unambiguously recognized the project itself (see appendix A for the questionnaire used for this first survey).

That first survey formed the input for a second survey that assessed the communication pattern between network members involved at a certain project at a certain design phase, as well as the perceived level of creativity and efficiency in a particular phase.

Statements were used like ‘I consider the result of this part of the innovation project as very original’, where the respondent could agree or disagree on a 7 point Likert scale.

The subject was then asked with which actors he or she communicated with during a particular project at a particular phase, but only about “new ideas, new ways of seeing things, giving solutions to problems and asking for advice”. If so, the subject was asked how often on average this communication took place on a 5 point scale (0=never, 1=less than 5 times a month, 2=between 5 and 10 times a month, 3=between 10 and 15 times a month, 4=more than 15 times a month). The subject was given a list with all internal project members and some open fields to fill in other people who they communicated with. Also, a list of groups from outside the organization was given: “customers”,

“suppliers”, “research institutes” and “other” to fill in any group that was not identified yet. This list was used the same way as the internal members (see appendix B for an example of this questionnaire).

After this, a limited number background interviews were taken on randomly selected project members that were involved in one or more of the projects, to interpret the

(9)

quantitative results and gain some broad extra insights in the context of the company.

Questions were asked on how they perceived their efficiency and creativity while working on different projects and design phases, and why they communicated with their own functional and divisional silos.

R

ESULTS AND ANALYSIS

The explorative interviews with the management board revealed insights in the dynamics of the company’s innovation efforts. A total of 11 innovation projects could be identified. At the time of measurement, 3 out of the 11 projects were in phase 2. A total of 8 projects reached phase 3 or completion. The company was structured around 7 different divisions or division, each partly delivering to other divisions and partly to third parties. The functional areas were split up into Engineering, Commercial, Production and support staff and Management as this resembled the structure used by the company. A total of 35 people were directly involved in innovation projects.

From the first survey, 35 out of 35 questionnaires were received (100%). It showed 160

“participations” (the involvement of a project member in a certain project at a certain phase of design) in the different innovation projects. These participations each were used for the second questionnaire.

From a total of 160 different participations, 135 were obtained (84,4%). The individual communication flows were combined into network files for each innovation project and phase of design. These communication intensity measures were then dichotomized in order to use the calculation method properly. Because the communication frequency did not show big differences and was less important than evidence of an existing communication ‘link’ regardless of how often this link was used, all communication (starting at less then 5 times a month but higher than 0) was regarded as ‘1’. The strength of the silo-communication of divisions and functional areas was then measured using the “External-Internal ties” embeddedness calculation method, developed by Krackhardt and Stern (1988). The calculations were made using the Ucinet 6.0 software tool by Analytic Technologies (Borgatti, Everett and Freeman, 2002). To minimize

(10)

problems of multicolineairity, the measures were standardized. These measures were then tested on correlations and regressions. For each of the three phases of design we used two regression models; one to test creativity and one for efficiency.

MODEL 1: IDEA GENERATION PHASE

The first model involves 11 communication networks that was taken from innovation projects in phase 1. Short descriptives are given (see table 1) as well as a correlation test

and the two regression models for both creativity and efficiency.

The correlation table (see table 2) shows a moderately strong correlation between creativity and Silo-Divisional (sign. ,091). Also, some correlation exists between the two silo-variables. Efficiency does not correlate on a relevant way at all with our two independent variables.

Creativity

Table 3 - Creativity regression model

Model

Sum of Squares R²

Mean

Square df F Sig.

Creativity Regression Residual Total

4,167 5,833 10,000

,417 2,064 ,729

2 8 10

2,858 ,116

Coefficients

Unstandardized Coefficients

Standardized Coefficients

Model B Std. Error Beta t Sig.

Creativity (Constant) SiloDivisional SiloFunctional

-7,94E-016 ,751 -,423

,257 ,315 ,315

,751 -,423

,000 2,387 - 1,344

1,000 ,044 ,216 Table 1 - Descriptives

Min Max Mean St. dev

Creativity 3,6 6,5 4,692 1,1117

Efficiency 3,83 6,00 4,5600 ,71881 SiloDivisional -,333 1,000 ,21418 ,435259 SiloFunctional -1,000 1,000 ,30191 ,564544 (n=11 for all variables)

Table 2 - Pearson Correlations

Creativity Efficiency SiloDivisional SiloFunctional

Creativity .

Efficiency ,316 (,344) .

SiloDivisional ,534 (,091) ,024 (,944) .

SiloFunctional -,037 (,914) -,021 (950) ,514 (,106) . (n=11 for all variables, significance is shown in parentheses)

(11)

The creativity regression model (table 3) is showing a significance of ,116. Although this is not nearly significant enough for p<,005 we can use this model as indicative. The residual for this model is high (more than half of the sum of squares). The variables SiloDivisional and SiloFunctional are not significant (see table 4). SiloDivisional has a significance of ,044 and is an indication that silo-communication inside the division is positively related to creativity in the first stage of design. The SiloFunctional factor seems to be negatively related to creativity but is not significant enough to use as a predictor.

The scatterplot (figure 2) reveals the linear relation between creativity and SiloDivisional. It is congruent with our indicative result we found earlier in the regression model: SiloDivisional increases creativity.

As for our SiloFunctional variable, the plot in figure 2 shows a scattered negative relation. When we exclude the two outer values (where SiloFunctional is about -2 and 1,5) a much stronger and significant negative effect appears. Because our small sample size of 11 innovation projects, the influence of these ‘extreme’ values greatly impact the model. Thus, a much stronger negative effect of SiloFunctional on creativity would possibly have been found when more projects were available to be measured.

Figure 2 - Scatterplots SiloDivisional and SiloFunctional

1,00000 0,50000 0,00000 -0,50000 -1,00000

SiloDivisional 2,00000

1,50000

1,00000

0,50000

0,00000

-0,50000

-1,00000

Creativity

R Sq Linear = 0,416

1,00000 0,00000 -1,00000 -2,00000

SiloFunctional 1,50000

1,00000

0,50000

0,00000

-0,50000

-1,00000

-1,50000

Creativity

R Sq Linear = 0,184

(12)

Efficiency

Table 4: Efficiency regression model

Model

Sum of

Squares R²

Mean

Square df F Sig.

Efficiency Regression Residual Total

,021 9,979 10,000

,002 ,011 1,247

2 8 10

,009 ,991

Coefficients

Model B

Std.

Error Beta t Sig.

Efficiency (Constant) SiloDivisional SiloFunctional

1,64E-015 ,048 -,046

,337 ,412 ,412

,048 -,046

,000 ,116 - ,112

1,000

,910 ,914

The efficiency regression model (table 4) is nowhere near significance and has a residual that is practically the same as the total sum of squares. Efficiency therefore does not seem to be influenced by our Silo-variables in phase 1 as we expected in our hypotheses.

When looking on the scatterplot from SiloDivisional on efficiency, we can see a near- horizontal regression line. Even when we exclude the two extreme values with both high efficiency and SiloDivisional (in the right top corner), the relation still has a very small slope. Clearly the influence of divisional silo communication on efficiency is minimal in this phase.

The plot with the relation between SiloFunctional and Efficiency however does show that a stronger positive relation might exist. Especially when excluding the very extreme

1,00000 0,50000 0,00000 -0,50000 -1,00000

SiloDivisional

2,00000

1,00000

0,00000

-1,00000

Efficiency

R Sq Linear = 0,002

1,00000 0,00000 -1,00000 -2,00000

SiloFunctional

2,00000 1,50000 1,00000 0,50000 0,00000 -0,50000 -1,00000 -1,50000

Efficiency

R Sq Linear = 0,002

(13)

value where SiloFunctional is about -2,1 we might expect a stronger positive relation in a greater population of samples.

MODEL 2: PROBLEM SOLVING PHASE

The second model also involved 11 communication networks that were taken from innovation projects in phase 2. Again, short descriptives (see table 1), correlation and two regression models for creativity and efficiency are given.

The correlations (table 6) show no significant relations between our variables. The highest significance level found is the relation between our two independent variables, indicating an inverse relation.

Creativity

Table 7 - Creativity regression model

Model

Sum of Squares R²

Mean

Square F df Sig.

3,093 6,907 10,000

,044 1,547 ,863

1,791 2 8 10

,228

Coefficients

Model B

Std.

Error Beta t Sig.

Creativity (Constant) SiloDivisional SiloFunctional

8,16E-017 -,393 -,313

,280 ,302 ,302

-,393 -,313

,000 -1,301 -1,036

1,000 ,230 ,331 Table 5: Descriptives

Creativity 3,0 6,0 4,470 ,9407

Efficiency 2,97 5,50 ,07155 ,69771 SiloDivisional -,500 1,000 ,23291 ,455818 SiloFunctional -1,000 ,750 4,1942 ,459200 (n=11 for all variables)

Table 6: Pearson Correlations

Creativity .

SiloDivisional -,046 (,893) -,404 (,218) .

SiloFunctional ,203 (,550) ,232 (,492) -,466 (,149) . (n=11 for all variables, significance is shown between parentheses)

(14)

1,5000 0 1,0000

0 0,5000

0 0,0000

0 -0,500

00 -1,000

00 -1,500

00 -2,000

00

SiloFunctional

2,00000

0,00000

-2,00000

Creativity

R Sq Linear = 0,042

The creativity model for this phase (table 7) does not have a high significance (,228) and also has a large residual. In the model, our silo-variables show a small negative influence on creativity.

The scatterplots in figure 4 confirm our findings in the regression model. The SiloDivisional variable is widely scattered and has an ambiguous relation with creativity.

SiloFunctional’s plot again indicates a scattered relation, but also seems to have a parabolic-like shape (when excluding the extreme value where SiloFunctional is about -1,7). To stimulate creativity one might have to minimize or maximize functional silo communication here. Unfortunately, 11 cases is not enough to state this strongly.

Efficiency

Model Sum of Squares R²

Mean

Square F df Sig.

Efficiency Regression Residual Total

,433 9,567 10,000

,166 ,217 1,196

1,791 ,181

2 8 10

,228 ,838

Coefficients

Model B

Std.

Error Beta t Sig.

1,00000 0,50000 0,00000 -0,50000 -1,00000 -1,50000

0,00000

-2,00000

Creativity

R Sq Linear = 0,003

(15)

-1,17E-016 ,214 -,049

,330 ,356 ,356

,214 -,049

-,139 1,000

,564 ,893

The efficiency regression model (table 7) is, again, nowhere near significance. The residual resembles almost all of the total sum of squares. In the second phase of design, efficiency does not seem to be influenced by our silo-effect in a relevant way according to our linear regression model. The significance of both our silo-communication variables is very low (,564 and ,893).

The scatterplot from SiloDivisional in figure 5 surprisingly indicates a parabolic-like pattern. A positive effect can be seen on efficiency with either a high or a low score on SiloDivisional. This is an interesting finding. Apparently, communicating very strong or very weak with people from the same division will stimulate efficiency in this phase.

This might depend on the nature of the ‘problem’ to be solved in this phase. A highly complex problem will likely require intensive cross-functional communication (low divisional silo communication) in order to work efficient whereas low complex problems can be solved inside the silo.

The pattern of SiloFunctional (figure 5) is widely scattered and thus confirms that the role of functional silo-communication does not have a relevant influence on efficiency in this second phase of design.

1,00000 0,50000 0,00000 -0,50000 -1,00000 -1,50000

0,00000

-2,00000

Efficiency

R Sq Linear = 0,118

1,5000 0 1,0000

0 0,5000

0 0,0000

0 -0,500

00 -1,000

00 -1,500

00 -2,000

00

SiloFunctional

1,50000 1,00000 0,50000 0,00000 -0,50000 -1,00000 -1,50000 -2,00000

Efficiency

R Sq Linear = 0,003

(16)

MODEL 3: IMPLEMENTATION PHASE

The third and last model also involved 8 communication networks that were taken from the innovation projects that reached phase 3 or were completed. The descriptives (see table 9), correlation and two regression models for creativity and efficiency are given.

The Pearson correlations (table 10) show a significant and positive correlation between SiloDivisional and Silo- Functional. This finding can be an indication of an overlap between the two variables.

Creativity

Table 11: Creativity regression model

Model

Sum of

Squares R² df

Mean

Square F Sig.

2,533 4,467 7,000

,362 2 7 5

1,266 ,893

1,41 7

,325

Coefficients

Model B Std. Error Beta t Sig.

Creativit y

(Constant) SiloDivisional SiloFunctional

5,94E-016 ,132 ,490

,334 ,603 ,603

,132 ,490

,000 ,220 ,812

1,000 ,835 ,453

The residual of the creativity regression model (table 11) is large: about two-third of the sum of squares. Significance is ,325 showing that the model in itself does not have much relevance. Our silo-variables show no significant influence on creativity in the last phase.

Table 9: Descriptives

Creativity 4,0 6,0 4,854 ,7060

Efficiency 3,07 5,00 3,9567 ,79072 SiloDivisional -,333 1,000 ,30088 ,368337 SiloFunctional -1,000 1,000 ,54825 ,649524 (n=8 for all variables)

Table 10: Pearson Correlations

Creativity .

SiloDivisional ,527 (,180) ,595 (,120) .

SiloFunctional ,596 (,119) ,394 (,335) ,805* (,016) . (n=8 for all variables, significance is shown between parentheses)

* . Correlation is significant at the 0.05 level (2-tailed).

(17)

0,50000 0,25000 0,00000 -0,2500

0 -0,5000

0 -0,7500

0 -1,0000

0

SiloFunctional

1,50000

1,00000

0,50000

0,00000

-0,50000

-1,00000

Creativity

R Sq Linear = 0,117

The scatterplot of SiloDivisional (fig. 6) will show that a stronger positive relation between SiloDivisional and Creativity when excluding the extreme value where SiloDivisional is about 1,4. However, because of the very few cases (n=8) this is questionable.

The SiloFunctional scatterplot (also fig. 6) does not show a relevant relation between SiloFunctional and creativity. That is congruent with our findings from the regression model: Functional silo-communication does not play a significant role in the cases we used in this research.

Efficiency

Model

Sum of

Squares R² df

Mean

Square F Sig.

Efficiency Regression

Residual Total

2,625 4,375 7,000

,375 2 5 7

1,312 ,875

1,500 ,309

Coefficients

Model B

Std.

Error Beta t Sig.

4,58E-017 ,791 -,244

,331 ,597 ,597

,791 -,244

,000 1,327 -,409

1,000 ,242 ,700

1,50000 1,00000 0,50000 0,00000 -0,50000

1,00000

0,50000

0,00000

-0,50000

-1,00000

Creativity

R Sq Linear = 0,01

(18)

0,50000 0,25000 0,00000 -0,2500

0 -0,5000

0 -0,7500

0 -1,0000

0

SiloFunctional

1,00000

0,50000

0,00000

-0,50000

-1,00000

-1,50000

Efficiency

R Sq Linear = 0,032

The efficiency model does gain on significance compared to other phases: in this phase sign. is ,309 compared to ,991 and ,838 for phase 1 and 2 respectively. The SiloDivisional factor indicates a strong (but not significant; significance is ,242) positive effect on efficiency.

The findings in our regression model are similar when looking at the scatter diagram for SiloDivisional (fig. 7). More data is needed to confirm whether this relation is indeed significantly strong and positive.

The SiloFunctional variable does not influence efficiency in a linear way, but here may be a different effect. When SiloFunctional is either low or high, efficiency is stimulated.

Again, we need more data to investigate this in more detail as eight cases is too low to draw conclusions on this effect.

Background interviews

To gain insights in the context of the silo-communication a second interview round was taken. It consisted of 9 short interviews taken by four different project members. They were asked why they communicated inside or outside their divisional and functional silo in a certain phase of design of a certain project. Also, insights in their perceived efficiency and creativity were gained. See appendix C for a table of the complete results of these background interviews.

1,50000 1,00000 0,50000 0,00000 -0,50000

0,50000

0,00000

-0,50000

-1,00000

Efficiency

R Sq Linear = 0,26

(19)

An indication for a functional silo in phase 1 was mentioned by a respondent that formed a “key-user group” that consisted of specialists from each division (thus a functional silo). Also, a respondent mentioned that he perceived the efficiency as low; a lot of time was spent because there were no structures and deadlines. Another respondent showed that in a project during phase 1 there was a need for a very complex solution and he used “breeding time”. These remarks indicate that the respondents made an effort for creativity prior to efficiency. Silo communication on functional or divisional level was mentioned by somebody to be limited because the idea was still vague and not everybody was picking up the idea yet.

About project communication during phase 2, a number of respondents said they “have to test things”, “coordinate technical information” and that “we have the technical information” as the reasons why they had communicated with peers from their functional area more. Communication inside the division took place because “not all information reaches the right people in one effort”. Some people had delegated their tasks. Some talk about the project because the project leader is member of the same BU.

The self-perceived efficiency shows some interesting insights. “A lot of times we discussed things we had discussed before” and “could be better when tasks were smaller and that strict deadlines would be used; testing should only have begun when the task is complete”. “We didn’t know exactly what the product could and couldn’t do.

It was a process of trial and error”. “The structure becomes more evident now when the project is progressing”. Some kind of balancing act can be seen in the need for structure and deadlines for managing the process efficiently yet there is the same need for technical solutions that still require a fair amount of creativity and therefore communication.

In phase 3, testing was mentioned to be a reason to communicate within the functional silo. According to one respondent, he felt that the efficiency was lowered because the tasks were divided to one person only making the risk of delay very strong. Also, the risk of damaging the day-to-day continuity was mentioned as a reason to cut corners and prioritize normal business processes above innovation projects. It seems that some

(20)

kind of misfit in project task-communication or project management might have a negative influence on some project’s efficiency at the company. This might be a reason for the low perceived efficiency we found in later stages of our quantitative data.

Implications

In the theory section we explained how the different innovation phases of design require different tasks (and therefore adequate information) to be completed. Early stages involve creative thinking and low silo-forming in order to be creative, whereas communication in later stages should be structured and could be inside the silo’s for efficiency reasons.

Contrary, our results from the regression model for phase one found divisional silo forming to be positively related to creativity. Communicating more intensely with team members from the same division compared to others stimulates creativity in the idea generation phase. Functional silo-communication indeed had a negative effect on creativity. A possible explanation for this finding could be that cross-functional effects are causing this: people with different functional backgrounds are simply more creative together. When this cross-functional group communicates inside their division, knowledge sharing around innovation matters will stimulate creativity even more. Also, the dependency between day-to-day business activities and innovation projects might play a role here. When people are not highly ‘connected’ in daily business tasks, they may have more difference in background than highly interconnected people.

In the later stages of design, the relevance of the influence of silo-communication on creativity becomes less. Divisional silo-communication does maintain some positive influence on creativity by either maximizing or minimizing silo-communication. This effect can possibly be explained by the complexity of the (mostly functional) problems in phase two. Functional silo-communication is measured as far less relevant. Shared goals but different backgrounds between members of the same division might cause this positive effect.

(21)

When using the data from this single company, efficiency appeared not to be influenced by silo-communication in a relevant way. However, the role of functional silo- communication on efficiency does seem to have a positive role as indicated by the scatterplot. In later stages of design, this effect is diminished. Divisional silo- communication gains positive influence on efficiency when the silo-communication is either high or low. Also, this variable seems to gain it’s positive strength in the last phase. Progression in the design phase might indeed slightly increase the positive effects of silo-communication on efficiency.

These findings have the following implications for our hypothesis:

Hypothesis 1a: The silo-communication strength of divisional and functional areas will be positively related to efficiency.

The functional silo-communication was positively related to efficiency in the first phase of design only, and did not show relevant in later stages. Divisional silo-communication appeared to be positively related in the second and last stage and was not significant in the first phase. In the background interviews, problems with task-coordination were mentioned by some respondents to negatively influence efficiency in later phases of design, and this might explain why we found the silo-communication’s variables to have some extreme values and distorted some significance. With more data, we can expect to support this hypothesis.

Hypothesis 1b: The relation between the two silo-communication factors and efficiency will be stronger if moderated by the phase of the design project; the further in the development process, the stronger this relation.

The regression models show that the relevancy of the influence of silo-communication on efficiency indeed becomes stronger comparing the first, second and third phase.

Interesting is that divisional silo-communication is not significant in the first phase and did show some positive effect on efficiency later on, possibly because of either high or low complexity of the problem to be solved in phase two. Contrary, functional silo-

(22)

communication did only show some positive effect in phase one. Overall, we can say that there are clearly indications to accept this hypothesis. Again, with more data we might find more significance of these effects.

Hypothesis 2a: The silo-communication strength of divisional and functional areas will be negatively related to creativity.

This hypothesis is not supported. Instead, communicating within a divisional silo is positively related to creativity in the first phase. This finding was the strongest and most significant finding of our research results. Apparently, when communicating more intense with peers from the same division, creativity is stimulated. This might be influenced by cross-functionality: functional silo communication shows to have a slightly negative relation with creativity. People from different functional background but the same division and likely shared goals stimulate creativity. With that said, we cannot support this hypothesis.

Hypothesis 2b: The relation between the two silo-communication factors and creativity will be stronger if moderated by the phase of the design project; the further in the development process, the weaker the relation.

When we compare the influence of silo communication on creativity in the first phase with the communication in later phases, we can see the relevancy of this relation getting weaker. Our regression models showed less significance and higher residual levels and the plots show less strong influences. Therefore, we can accept this hypothesis.

DISCUSSION

In this explorative research we investigated the role of silo communication on creativity and efficiency in social networks during 11^a innovation projects in a single Dutch medium sized company. In our theory section we explained that functional and

a 3 out of 11 projects did not reach the implementation phase at time of measuring. The data of 8 projects were used in the third phase.

(23)

divisional silo communication was expected to negatively influence creativity and to stimulate efficiency. The dynamic environment of innovation projects and the different tasks to be completed in early or late stages of design would cause a difference in the role of silo communication on creativity and efficiency.

Our findings show that divisional silo communication actually stimulates creativity in early stages instead of the expected negative influence whereas functional silo communication indeed showed a negative influence. Cross-functional effects might be the cause for this: different functional backgrounds but shared divisional business goals stimulate the creativity. The positive effect of silo-communication was less strong and significant in later stages, indicating that indeed the influence of silo-communication on creativity is larger in early stages that later stages. The influence of silo-communication on efficiency did not show much relevancy in early stages of development. Only functional silo-communication had some positive effect. At late stage, the relevancy of divisional silo-communication on efficiency slightly increased.

Limitations

This explorative research used data of only a single company because of time and budget constraints. Because of the small number of cases (n=11 or 8) we could not control for factors as innovation radicalness, differences among product-, market- or process oriented projects, team age and team size (which in intra-team communication literature is an important factor). Also, because of the calculation method we used, the role of the actual frequency of communication could not be investigated in detail. We recommend that more research is needed on a much larger scale to be able to control for these factors and to be able to generalize results.

R

EFERENCES

Adams, R., Bessant, J., & Phelps, R. (2006). Innovation management measurement: A review.

International journal of management reviews, 8(1), 21-47.

Alberg, R. (2007). The Silo Mentality. People management, 13(19), 9-9.

Allen, T. (1977). Managing the flow of technology. Cambridge, MA: MIT Press.

Amabile, T.M. (1998). How to Kill Creativity, Harvard Business Review, 76(5), 76-87.

(24)

Amabile, T.M. (1996). Creativity in Context: Update to the Social Psychology of Creativity.

Boulder, CO: Westview Press.

Borgatti, S.P., Everett, M.G., & Freeman, L.C. (2002). Ucinet for Windows: Software for Social Network Analysis. Harvard, MA: Analytic Technologies.

Cross, R., Borgatti, S.P., & Parker, A. (2002). Making Invisible Work Visible: Using Social Network Analysis to Support Strategic Collaboration. California Management review, 44(2), 25-46.

Cooper, R.G. (1997). Fixing the fuzzy front end of the new product process. CMA magazine, 71(8), 21-24.

Dasgupta, (1996). Technology and Creativity, Oxford University Press, New York.

Day, G.S. (2006). Aligning the Organization with the Market. MIT Sloan management review, 48(1), 41-49.

Eesley, D.T., Longenecker, C.O. (2006). Gateways to Intrapreneurship. Industrial management, 48(1), 18-23.

Ford, C.M. (1996). A Theory of Individual Creative Action in Multiple Social Domains. The Academy of Management review, 21(4), 1112-1142.

Gulati, R. (2007). Silo Busting. Harvard business review, 85(5), 98-108.

Kiesler, S., Cummings, J. (2002). What do we know about proximity and distance in work groups: A legacy of research’ in Distributed work. P. J. Hinds and S. Kielser (eds), MIT Press, 57–82.

Krackhardt, D., Stern, R.N. (1988). Informal Networks and Organizational Crises: An Experimental Simulation. Social psychology quarterly, 51(2), 123-140.

Myers, S., Marquis, D.G. (1969). Successful Industrial Innovations: a study of factor underlying innovation in selected firms. National Science Foundation 69(17).

Nahapiet, J., Ghoshal, S. (1998). Social Capital, Intellectual Capital, and the Organizational Advantage. The Academy of Management review, 23(2), 242-266.

Shalley, C.E. (1991). Effects of productivity goals, creativity goals, and personal discretion on individual creativity. Journal of applied psychology, 76, 179-185.

Sy, T., Annunzio, L. (2005). Challenges and Strategies of Matrix Organizations: Top-Level and Mid-Level Managers. Human resource planning, 28(1), 39-48.

Tushman, M.L. (1977). Special Boundary Roles in the Innovation Process. Administrative Science Quarterly, 22(4), 587-605.

Young, T.M. (2007). Aircraft Design Innovation: Creating an Environment for Creativity.

Proceedings of the Institution of Mechanical Engineers; Part G; Journal of aerospace engineering, 221(2), 165-174.

(25)

APPENDIX A: sample questionnaire 1

In the following section, you are asked questions about your current employment at the company. Please check the corresponding box.

In the next section, you are asked questions about your participation in certain innovation projects. Note that there are three distinct project ‘phases’ for each project that you can be involved in.

For each project, we ask you to show your participation (regardless of the importance or level of participation) in any of the three phases by checking the corresponding box. If you f.i. participated in all phases you should check all three boxes.

End of questionnaire.

My functional area is best described

as:

○

Commercial

○

Engineering

○

Production and support

staff

Manage-

○

ment

The division that I

am working for is:

○

A

○

B

○

C

○

D

○

E

○

F

○

G

Project name I participated in

phase 1

I participated in phase 2

I participated in phase 3

“Project A” (this line was repeated

for each project at the company)

○ ○ ○

(26)

APPENDIX B: sample questionnaire 2

In the following section, four statements are shown. Can you tell us on which level you agree or disagree with these statements? Please check the box that corresponds the most with your answer.

1. The result of this part of the project is very original.

I disagree very much

I disagree much

I disagree a little

I do not agree and do not

disagree

I agree a little A agree much I agree very much

○ ○ ○ ○ ○ ○ ○

2. In this part of the project we used as little time as possible.

I disagree much

I disagree a little

disagree

○ ○ ○ ○ ○ ○ ○

3. In this part of the project we used as little resources as possible.

I disagree much

I disagree a little

disagree

○ ○ ○ ○ ○ ○ ○

4. I consider this part of the project as very efficient.

I disagree much

I disagree a little

disagree

○ ○ ○ ○ ○ ○ ○

5. Who did you communicate with during this part of the project, regarding the following:

new ideas

new ways of seeing things,

giving solutions to problems about the project

asking for advice about the project

(27)

The method of communication is not relevant: any means of communication (including telephone and e-mail) is included.

Can you also tell us how often you communicated with this person or group? Please check the corresponding box (only one answer per row is allowed).

5.1 Groups outside the company (select only 1 answer per row)

Name of the group Never Less than 5 times a month

5 to 10 times a month

More than 15 times a month Customers

○ ○ ○ ○ ○

Suppliers

○ ○ ○ ○ ○

Research institutes

○ ○ ○ ○ ○

Other:_______

○ ○ ○ ○ ○

5.2 People inside the company (select only 1 answer per row)

Name of employee Never Less than 5 times a month

More than 15 times a month

<for each employee this line

was repeated>

○ ○ ○ ○ ○

Other:_______

End of questionnaire.