The impact of project portfolio management practices on the relationship between organizational ambidexterity and project performance success

Abstract Some studies suggest that organizational ambidexterity is best orchestrated through individual projects. However, stand-alone individual projects are relatively limited in scope, while suffering from susceptibilities to horizontal and vertical segmentation. This may render them poorly suited to serve as conduits for organizational ambidexterity. By contrast, organizations which deliver projects in portfolios, often in order to maximize resource utilization, may discern that these also provide better conduits for organizational ambidexterity. This study examines not only the extent to which project portfolio management (PPM) practices impact orchestrations of organizational ambidexterity, but also whether these orchestrated PPM practices impact further lead to superior project performance. Data were collected from one hundred and sixty PPM stakeholders spread across eight countries in the Middle East November 2016 to January 2017. The study finds portfolios performance to be strongly and highly correlated with organizational ambidexterity. Furthermore, the more organizations exhibited efficient project-portfolio-management practice, the more they were found to develop ambidextrous capabilities.


Context
Academic scholarship has traditionally suggested that in globalized operations, independent 'stand-alone' projects are key to orchestrating the strategic visions of the parent organizations (Chipulu et al. 2016;Kopmann et al. 2017;Maylor et al. 2018;Midler, Maniak, and de Campigneulles 2019). This view recognizes that projects involve non-routine, ad-hoc and one-off decisions. These decisions may be further laden with potentially helpful 'paradoxes', 'contradictions' and 'dualities' that have traditionally spoken to perceived needs for trade-offs between competing strategic expectations and positions (Roscoe and Blome 2019). Generally, 'Organisational ambidexterity' is an approach that organizations can follow in order to accommodate spontaneous and dynamic tensions arising between management practices by simultaneously engaging in contradictory explorative and exploitative behaviour (Tushman and O'Reilly 1996;O'Reilly and Tushman 2011).

The problem setting
It is commonly accepted that orchestrations of ambidexterity are challenging because of the distinct structural and competency requirements required to support the multiple distinct and contradictory constructs and related practices at issue (Tushman and O'Reilly 1996;O'Reilly and Tushman 2011;Luger, Raisch, and Schimmer 2018;Martin, Keller, and Fortwengel 2019). These differences mean that finding appropriate organizational conduits for organizational ambidexterity is critical. Often individual projects are viewed as able to provide such conduits (Eriksson 2013;Turner and Lee-Kelley 2013;Turner, Maylor, and Swart 2015;Petro et al. 2019). As our literature review will explain, a number of descriptive and normative perspectives underlie this view (e.g. commonplace views of individual projects as seamlessly engaged in continuous iteration and adjustment).
Contrary perspectives express concern that individual projects are susceptible to high fragmentation, and so they look to alternative forms of project orchestration for organizational ambidexterity. One particular form, sometimes preferred, involves combined use of projects and programmes (Pellegrinelli, Murray-Webster, and Turner 2015). Midler, Maniak, and de Campigneulles (2019) prefer programmes over projects. In this context, drawing from both Stal-Le Cardinal and Marle (2006, 226) and Pellegrinelli (2011, 236), we define a 'project' as a ' … temporary, but deliberate manifestation of an intent to utilise structured techniques and approaches to efficiently deliver tangible deliverables'. More specifically, the management of programmes is concerned with projects that maintain similar objectives and resources.

The research aim and question
This research therefore sets out to examine the extent to which PPM practices impact orchestrations of organizational ambidexterity, and whether these PPM practices further impact the abilities of organizations to channel their ambidextrous capabilities towards superior project performance. At the crux of our enquiry is whether 'Project portfolio management' enables/increases the effect of 'Organisational ambidexterity' on 'Project performance successes'. We therefore present our research question as: RQ: To what extent do Project Portfolio Management (PPM) practices impact upon the orchestration of ambidexterity, and do these practices impact upon an organization's ability to channel its ambidextrous capabilities towards superior project performance?
This research question is framed for consistency with the practice of 'gap spotting' in the formulation of research questions (see Sandberg and Alvesson 2011;H€ allgren 2012). Notably, the gap which our research question addresses appears to exist despite literature asserting that PPM is optimal for managing and actualizing projects/programmes (Pellegrinelli, Murray-Webster, and Turner 2015), and other literature which has primarily focussed on understanding the challenges that organizations face which may be resolvable through PPM (Petro et al. 2020). Recognizing that such research is obliquely related to our own, we explore territory (Handfield and Melnyk 1998, 324) in the interplay between (i) Organizational PPM, (ii) Organizational ambidexterity and (iii) Project performance (project success); one secondary outcome of exploring such territory will be our extension of the above project-focussed ambidexterity literatures.

Structure of the article
The next section (Section 2) gives an overview of the relevant literature in organizational ambidexterity. This commences with a review of relevant concepts. We also examine PPM literature, and the relationship between PPM and organizational ambidexterity. In Section 3, we present the research methodology, which is based on data obtained between November 2016 and January 2017 from 160 practitioners. It is explained that the questionnaire was improvised from organizational scale measures developed earlier by Lubatkin et al. (2006). In Section 4, we report the results of the data analysis, which was undertaken using IBM SPSS Statistics software version 22. Discussions of implications from findings are presented in Section 5. The article concludes in Section 6 by attesting to the importance of the portfolio approach while calling attention to its advantages with some critical reservations.

What is organizational ambidexterity?
At its very basic form, 'organisational ambidexterity' is a concept that seeks to explain the competencies and capabilities which organizations need to develop and maintain in order to successfully engage in the simultaneous co-existence, trade-off and switching between 'exploration of new possibilities' and 'exploitation of old certainties' (March 1991;Tushman and O'Reilly 1996;O'Reilly and Tushman 2011). 'Organisational ambidexterity' refers to an organization's ability to ' … simultaneously (i.e., concurrently) pursue both incremental and discontinuous innovation and change … ' (Tushman and O'Reilly 1996, 24). Two relevant terms of action are important here. First, the 'exploration of new possibilities' focuses on search and discovery actions (March 1991). This construct is both strategic and experimental in character, being focussed on 'distant' knowledge (Petro et al. 2019(Petro et al. , 2020. Second, the 'exploitation of old certainties' focuses on deriving advantage from current and evolved management initiatives and other practices, not only recognizing but also actively regulating and cultivating the organizational efficiencies and internal organizational consistencies with which these activities are associated (March 1991). Exploitation is therefore more tactical than strategic in character, being associated with 'near' knowledge that is more likely to emerge from existing management information (Petro et al. 2020).
Generally, organizational ambidexterity literature assumes tensions not just between management practices in general, but more specifically, and thinking more reflexively here, it also considers tensions associated with trade-offs which organizational ambidexterity effort generates itself. In effect, it assumes such tensions will be greater where there is more organizational capacity to engage in the necessary structural separation effort (Simsek 2009).
The literature posits that exploration and exploitation require different but complementary organizational structures and designs. For example, exploration generally thrives under decentralized structures while exploitation requires more centralization. The literature also posits that organizations simultaneously pursuing and switching between exploration and exploitation will outperform those that do not, as they develop complementary competencies (Luger, Raisch, and Schimmer 2018;Wolf et al. 2019). This view of 'complementary' suggests that an organization's ability to pursue exploration is more likely to be successful when undertaken in the presence of a simultaneous pursuit of exploitation and vice versa (Levinthal and March 1993).

Projects as conduits for organizational ambidexterity
Three reasons predominate within literature, for regarding individual projects as the most appropriate conduit for organizational ambidexterity. First, operational routines can be implemented and managed (Hayes 2009). Second, project management as a discipline is very familiar with demands associated with balancing the need for certainty (via the use of formalized project management methods: Milosevic and Patanakul 2005) against the need to create the space for unanticipated challenges (Gross 2014;Klein, Biesenthal, and Dehlin 2015). Individual project structures also allow for easy switching of project resources between 'exploitation' and 'exploration' activities. Third, management of individual projects can readily oversee continuous adjustments without major disruptions to team structures, particularly since projects are inherently ephemeral (S€ oderlund 2013; Tryggestad, Justesen, and Mouritsen 2013) and temporal (S€ oderlund 2013). Furthermore, most project teams are organized for delivering one-off tasks with pre-designated (usually) completion dates. They are also structured amorphously (Scott-Young and Samson 2009), allowing for flexible selforganizing (Manning 2017) and also for professionals to be rotated in and out of the project for designated tasks (Riis and Pedersen 2003) with minimal disruption (Scott-Young and Samson 2009). Furthermore, project teams are usually subjected to flexible role-balancing between 'reserved legitimacy', which comes with appointment to their role, and 'status hierarchy', which comes from participation within project governance (Clegg and Courpasson 2004). The extent to which individual projects facilitate individual manager adjustment is therefore also key. Lee, DeLone, and Espinosa (2006) found software development project managers engaging in ambidextrous coping strategies. Aubry and Li evre (2010) examined the experience of duality associated with varying action modes over project lifecycles. Lin and McDonough (2011) found a correlation between project leadership and innovation ambidexterity. Hoang and Rothaermel (2010) looked at organizations (at the strategic and operations level) and their managerial participation in different alliances and external partnerships, suggesting that the 'exploration' construct of ambidexterity, within this frame, negatively impacted project outcomes, while the 'exploitation' construct exerted positive impact. In summary, then, the relationship between organizational ambidexterity and project-related performance suggests that stronger manifestations of ambidexterity (i.e. ambidextrous capabilities) in individual project management is likely to lead to stronger project performance (Jansen, Simsek, and Cao 2012; Cegarra-Navarro, Jim enez-Jim enez, and Garcia-Perez 2019; Dezi et al. 2019).
Reservations expressed, concerning preferences for individual projects as conduits for organizational ambidexterity, include the following. Projects are increasingly complex, particularly by reliance on distributed and networked teams. This can lead to high levels of both horizontal and vertical fragmentation (Fellows and Liu 2012;Alashwal and Fong 2015). Here, while horizontal fragmentation implies that projects increasingly involve multiple and heterogeneous specialist-functional actors, vertical fragmentation recognizes project delivery processes as disruptively phase-managed by actors drawn from multiple levels of organizational hierarchies. Reasons given for high internal fragmentation include heterogeneous stakeholder objectives (Chipulu et al. 2019), diverse cultural perspectives on projects (Chipulu et al. , 2016Ojiako, Chipulu, et al. 2015, Ojiako, Papadopoulos, et al. 2015, functional and role specialization  and preference for projects to be monitored and controlled within distinct stage or phase boundaries (Parvan, Rahmandad, and Haghani 2015). Sometimes such fragmentation has led to projects being characterized by not only poor flow of information, knowledge and learning, but also limited sharing of resources. This has led to calls for more inter-project coordination ('No project is an island', Engwall 2003, 789) recognizing that projects are increasingly interdependent with other projects and are likely to enhance their performance by leveraging this interdependence. Accordingly, in the next section we consider the project portfolio perspective.

Portfolios as an alternative conduit for organizational ambidexterity
Unlike programmes which are interested in projects that maintain similar objectives and resources, PPM is more focussed on maintaining a centralized perspective of its constituent projects. The primary aim, here, is to ensure that there is limited or no competition among specific projects for resources (Sohani and Singh 2017). Through portfolios, organizations can provide platforms for resource sharing between projects, leveraging not only information, knowledge and learning, but also competencies and resource expertise from partner projects (Hoang and Rothaermel 2010;Jansen, Simsek, and Cao 2012;Stettner and Lavie 2014;Wassmer, Li, and Madhok 2017). In effect, portfolios serve as boundary-spanning platforms that facilitate bridging of sometimes highly diverse projects (Montibeller et al. 2009;Kornfeld and Kara 2011). Portfolios can also provide some level of predictable institutional regulatory and cultural context for managing and coordinating resources across projects, and they can further be used to optimize risk ownership distributions. Using portfolios in this way, an organization can increase the performance of individual projects within the portfolio. Portfolios vary by type (see Bresciani, Ferraris, and Del Giudice 2018;Luger, Raisch, and Schimmer 2018;Wolf et al. 2019). One type is where constituent individual projects are relatively dis-similar, perhaps lacking overall networked coordination. This type may be more suitable in circumstances where the organization is particularly interested in 'explorative' operations strategies (and individual supporting projects). In such portfolios, there is likely to be a focus on 'upstream' projects requiring new knowledge creation. Examples include New Product Development (NPD) projects. Conversely, an organization that is more interested in 'exploitative' operations strategies may focus its attention on portfolios comprising projects which are either part of the same coordinating network(s) or which are in varying ways homophilic with each other. Here, the emphasis is likely to be on collaborating for 'downstream' endeavours. Examples include projects focussed on product commercialization.
A number of salient factors discussed in alliance theory provide insight into the ambidextrous capabilities of portfolios (Teller et al. 2012;Sun and Lo 2014;Stettner and Lavie 2014;Wassmer, Li, and Madhok 2017;Degener, Maurer, and Bort 2018). These factors include portfolio maturity and coordinating experience. Portfolio maturity and co-ordinating experience involve particular aligning mechanistic tendencies. As portfolios become more mature and their constituent projects become embedded within their information and knowledge networks, and also more reliant on complementary resources, they become more likely to emphasize exploitation. Similarly, portfolios with substantial coordinating experience (and associated mature/efficient mechanisms) are more likely to focus towards downstream endeavours, thus also preferring exploitation.
The use of portfolios serves to mitigate against the challenges of dual pulls experienced by single projects seeking to pursue contradictory strategic objectives. In this vein, project portfolios are likely to serve as better conduits of ambidexterity in that they represent a more appropriate platform to deal with the tensions that arise from exploration and exploitation. This is because portfolios allow for the facilitation of permeability across project boundaries, thus leveraging more varied competencies for ambidextrous management practice. It must be noted, drawing from Tiwana (2008), that while project portfolios are associated with various ambidexterity-related and project performance benefits, these can be mediated by the ties (relationships) between the participating projects. Thus, projects exhibiting strong inter (homogeneous) portfolio relationships may be much better able to implement exploitative initiatives, while at the same time being less able to develop these initiatives. Conversely, projects within the same portfolios which exhibit strong heterogeneous portfolio relationships are arguably very often less likely able to implement exploitative initiatives, while being more likely to develop these initiatives. In effect, then, project portfolios with individual projects that are more able to forge relationships with other projects outside the same portfolio, and which therefore become connected to other projects with varying skills, expertise and competencies, are more able to generate new ideas. Nonetheless, they are arguably very often less able to successfully implement these initiatives. Hence, our first hypothesis considers that the performance PPM practices impact upon the orchestration of ambidexterity.
H1a. Portfolio performance is correlated with the orchestration of organizational ambidexterity

Organizational ambidexterity and superior project performance
The literature opines that organizations engaged in exploration-focussed projects are likely to end up in endless 'search and innovation' that remains underexploited (Junni et al. 2015;Petro et al. 2019) unless they develop strategic capabilities that are 'rare' and 'inimitable' (Barney 1991(Barney , 1996(Barney , 2001Barney, Wright, and Ketchen 2001). Moreover, investing or promoting projects focussed primarily or solely on exploration is an inherently risky strategy. This is because advantages to be gleaned from such projects can take years to be realized, if at all (Lubatkin et al. 2006). In fact, organizations investing or promoting projects predominantly focussed on exploitation may become increasingly less willing and able to search for new opportunities; hence, their existing skill base may become obsolete (Levinthal and March 1993). Focussing equally on exploration and exploitation, on the other hand, is likely to place an organization in a position where it both enshrines its current core competencies and becomes willing and able to search for and develop future capabilities. Generally, the high rate of uncertainty associated with project environments raises questions relating to how portfolios perform (as ambidexterity conduits) in terms of effectiveness and efficiency (see Lin, Yang, and Demirkan 2007). Here, we have drawn upon the earlier works of Neely, Gregory, and Platts (1995, p. 80) to define 'performance' as ' … the effectiveness and/or efficiency of action'. From this literature, 'effectiveness' refers to the degree to which a project meets stakeholder specifications and expectations, while 'efficiency' refers to resources deployed to meet stakeholder specifications and expectations. Drawing from Cooke-Davies (2002; p. 188), project performance, thus conceived, further predicts project success.
Drawing from the literature, it is widely proposed that when organizations simultaneously engage in both exploration and exploitation, they are likely to outperform organizations that emphasize just one of these (Jansen, Simsek, and Cao 2012;Junni et al. 2013;O'Reilly and Tushman 2011;Hughes 2018;Luger, Raisch, and Schimmer 2018). Thus, the literature acknowledges that ambidextrous competencies can enhance performance across all levels of an organization (Junni et al. 2013). This same view is shared in project management literature where scholars have found that the existence of ambidextrous capabilities within project teams significantly contributes to project performance (De Visser et al. 2010;Liu and Leitner 2012;Turner, Maylor, and Swart 2015). More often than not, co-dependency between the iterative clarification of objectives, continuous management review and continuous striving towards more efficient cycles of feedback, will lead to a key competency whereby the project is better able to avoid various process and learning traps that threaten project performance (Chandrasekaran, Linderman, and Schroeder 2012). Sohani and Singh (2017) note that ambidexterity plays a critical role in achieving project success. Hence, our second hypothesis considers that an organization's ability to channel its ambidextrous capabilities will drive superior project performance.
H2a. Ambidexterity is correlated with superior project performance.

The study and hypotheses
We wish to examine the relationship between PPM, organizational ambidexterity and (individual) project performance success. Study of the literature aligned to our own experience indicates possible relationships, but clearly a quantitative study was needed. We, therefore, undertook a survey relating these three areas together. While this raises the possibility of Common Methods bias (Podsakoff et al. 2003), it does nonetheless provide an initial study of these relationships.

Description of the survey instrument
For the study, we employed a questionnaire (Appendix B) consisting of five sections structured as follows. The first section (questions 1-5) focussed on respondent demographics, while the second section tapped evaluations of overall organizational success relative to competitor organizations. The third section focussed on more specific evaluations of project performance, construing 'PPM performance' in simple terms as a combination of 'PPM effectiveness' and 'PPM success'. Here, 'PPM effectiveness' denotes the degree to which use of PPM is perceived to meet stakeholder specifications or expectations (see Neely, Gregory, and Platts 1995). On the other hand, our 'PPM success' construct denotes perceptions that the intended goals of every project within the portfolio were actually delivered (see Chipulu et al. 2019Chipulu et al. , p. 1074. PPM effectiveness comprises: (i) 'Future preparedness' (Meskendahl 2010;Petro and Gardiner 2015), (ii) 'Strategic fit' (Jonas 2010;Petro and Gardiner 2015), (iii) 'Project portfolio balance' (Jonas 2010;Petro and Gardiner 2015) and (iv) 'Synergies/collaboration' between business units (Jonas 2010). While 'Future preparedness', 'Strategic fit' and 'Synergies/collaboration' employed three measures of assessment, 'Project portfolio balance' employed five measures. Four measures were used for 'PPM success' (schedule, budget, quality and customer satisfaction). These associations are diagrammatically represented in Figure 1. The fourth section focussed on perceived ambidexterity (see the following section). The fifth (final) section of the questionnaire focussed on perceived synergies and collaboration between projects (within a portfolio).

Scale measures for ambidexterity
As with Lubatkin et al. (2006) study, the 12 ambidexterity items used for the present research were measured by a 5point Likert scale, ranging from 1 ('strongly disagree') to 5 ('strongly agree'). The scale measures for 'exploration' and 'exploitation' employed in this study are shown in Appendix Table A1 (adapted from Lubatkin et al. 2006), which was based on earlier measures developed by He and Wong (2004)). The original He and Wong (2004) scale measured views concerning how well organizations shift their product design attention and resources between explorative and exploitative objectives. The Lubatkin et al. (2006) scale was innovative by employing twelve items balanced evenly between 'exploration' and 'exploitation'. These two constructs were included as distinct, but not separate, within one unitary measure designed to recognize their complementary contributions to organizational ambidexterity. This 'additive' measurement approach has also since been followed by other scholars (see Jansen, Simsek, andCao 2012, Jansen et al. 2009;Yu, Patterson, and de Ruyter 2013).
As the questionnaire (Appendix B) shows, we made the following changes to the original Lubatkin et al. (2006)

Hypotheses
The hypotheses stated above now become more precisely: H1b. Portfolio performance (PPM effectiveness and PPM success) is correlated with the organizational ambidexterity (using Lubatikin's measure).

Administration of the survey instrument
The survey instrument (in English) was administered online to respondents predominantly based in the Middle East. Each of the researchers approached practitioners within their individual networks on the basis of likely relevant (current or very recent) experience. As Chipulu et al. (2014) noted, although arguably a form of non-probability sampling technique, this means of identifying survey respondents is generally employed in research settings such as this study, where there is no comprehensive frame for sampling respondents deemed to possess the appropriate experience from which relevant information, knowledge and learning can be gleaned. Data collection was undertaken between November 2016 and January 2017. Of the 248 returned responses, we discarded 88 responses because they were either (i) incomplete or (ii) there were key critical items relating to ambidexterity which were not answered. This left 160 usable responses.

Brief descriptive overview of demographics
The overall industry distribution of the respondents suggests a combination of Construction (25.6%), Engineering (16.2%), Allied consultancy (9.4%) and Manufacturing (10%), amounting to 61.2% of the total respondents. About 41.9% of the respondents worked for organizations with more than 300 employees, 26.9% worked for organizations employing between 75 and 300 staff. More than 65% of the respondents (107 respondents) were drawn from senior to top management levels of their organizations.

Factor analysis
The survey data were analysed using a combination of approaches including (i) Factor, (ii) Correlation and (iii) Regression analysis. Adopting an approach similar to that of Chipulu et al. (2014), we opted to test both hypotheses simultaneously.
Factor analysis was carried out on the selected variables in order to establish the three central constructs comprising the research model: 'PPM effectiveness', 'Ambidexterity' and 'Project performance'. An exploratory factor analysis (CFA) was conducted using the factor reduction function in IBM SPSS Statistics software version 22. To test the factor loadings per construct, a Varimax rotation with a minimum of 25 iterations was used. In summary, the four variables used to describe PPM effectiveness ('future preparedness', 'strategic fit', 'project portfolio balance', and 'synergies/collaboration') loaded successfully on the four different factors, as shown in Table 1. A few items, as shown, were now dropped for not contributing.
The two variables for ambidexterity ('Exploration' and 'Exploitation') loaded successfully, with their 12 items loading  on two corresponding factor domains in line with expectation. This is shown in Table 2. A few items designed to tap ambidexterity were, however, jettisoned due to their relatively poor contributions. The two variables ('Performance' and 'Success') used to describe project performance and success also loaded successfully on two different factors as shown in Table 3. All items for this construct loaded successfully and were retained.
The relationship of ambidexterity with both was first of all tested, thereafter testing mediation with project performance separately due to its greater involvement with ambidexterity. All factors were then tested for reliability and internal consistency. These tests used all items that had loaded successfully into their relevant factors by Cronbach Alpha (Nunnally 1978). In taking this decision it was noted that while Cronbach's alpha has traditionally been used for this purpose, composite reliability tests are becoming more popular; nonetheless, according to Peterson and Kim (2013) influential study, the differences are 'relatively inconsequential'. The test showed all factors to be reliable. Almost all alphas were 0.7 and above, as Table 4 shows.
Although the 'PPM performance' items loaded successfully on two separate factors, its two constituent factors, 'Performance' and 'Success' exhibited reliabilities below 0.7. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy was carried out to ascertain the appropriateness of using factor analysis and to further confirm the reliability of those measures selected for field investigations (Kaiser 1970). This measure represents the ratio between variables' squared correlation and their squared partial correlation. A KMO value closer to 0 represents a diversion between those variables, and hence, factor analysis may not be appropriate for testing. Good KMO results are those with a value that is higher than 0.7 (Hutcheson and Sofroniou 1999). KMO was measured using SPSS and presented in Table 5, along with the reliability results.

Correlation analysis
A correlation analysis for factors constituting the research model was carried out using the bivariate Pearson correlation test (Table 6). Too enable this, a data normality test was conducted using two methods. First, normality was checked  visually by inspecting the graphs and figures generated through the descriptive analysis. A normal bell curve was detected for almost all the variables. This was slightly skewed to the right-hand side, indicating some participant optimism in responding to the field questionnaire. Second, the Kolmogorov-Smirnov statistical non-parametric test was carried out to confirm the visual findings for normality. A twotailed correlation test was therefore performed to capture all possibilities of correlation.
The outcome of our study showed support moderately strong and highly significant (r¼ þ0.501 at sig ¼ 0.01) for the relationship between ambidexterity and project performance, and moderately strong and highly significant (r¼ þ0.593 at sig ¼ 0.01) for the relationship between ambidexterity and PPM success.
Ambidexterity was found highly correlated with the combined/clustered global factor 'Performance' and 'Success' at r¼ þ0.641 and sig ¼ 0.01 also. Moreover, taking exploration and exploitation factors separately, these correlated with the factors for project performance and project success, both separately and clustered. 'PPM effectiveness' was found highly correlated with ambidexterity at r¼ þ0.778 with sig ¼ 0.01, suggesting that the more an organization or its business units exhibit an effective application of its PPM practices (measured through its 'PPM effectiveness'), the higher the level of ambidexterity, which contributes to the explanation of H1b (Portfolio performance being PPM effectiveness and PPM success together). In particular, the model showed 'PPM effectiveness' highly and strongly correlated with 'Performance' at r¼ þ0.539, average projects 'success' at r¼ þ0.687, and the cluster of 'Performance' and 'Success' at r¼ þ0.710.
As a number of factors were significantly and highly intercorrelated within the model, a check for collinearity was necessary prior to moving on to regression analysis. Correlation values of 0.8 were taken as the threshold for collinearity (Field 2009). Collinearity here refers to the close dependence, and the exhibition of 'non-independence', of the predictor variables (Dormann et al. 2013). This may occur, for example, upon measuring two very similar variables. For instance, one could be the 'age' of a respondent and another could be their 'year of birth'. Collinearity, in such cases, can easily inflate regression values (Dormann et al. 2013).
The correlation model shows 'Exploration', 'Exploitation' and 'Ambidexterity' as highly correlated. This, however, has already been assumed from widespread literature, reviewed earlier, which identified 'Exploration' and 'Exploitation' as the two ambidexterity constructs. The same applies for the other constructs 'PPM effectiveness', 'PPM performance' and 'PPM success'. However, since their correlation was close to 0.8, our study was particularly interested in the relationship between 'PPM effectiveness' and 'Ambidexterity'.
Collinearity between those factors was checked using the Variance Inflation Factor (VIF) measure (Field 2009). This measure shows whether a variable or a predictor has a strong linear relationship with the other variables(s) or predictor(s). It is based on Myers (1990) (see also Dormann et al. 2013), who suggests that VIF values above 10.0 may be problematic. However, it is worth noting that the guide for  Bivariate correlation analysis using the Pearson correlation method.

ÃÃ
Correlation is significant at the 0.01 level (two-tailed).
Bold font refers to relationships directly related to the quantitative research model. Bold italic font represents a check for collinearity -that is, a correlation that exceeds 0.8 may trigger a case of collinearity (similar items) as per Field (2009). IBM SPSS Statistics software version 22 (IBM SPSS Statistics, Armonk, NY) suggests that any value above 2.0 is of concern. Based on this, several scenarios for the two constructs and their constituent variables were performed. The VIF for all variables can be seen in Table 7 to be less than 2.0, indicating that collinearity or multi-collinearity was not an issue.

Regression analysis and test for mediation
The previous section presented a correlation analysis between all the model variables. This section takes the analysis one step further using the Ordinary Least Square (OLS) regression method. OLS regression tends to be used to predict the behaviour of relationships and how they could act in future events (Field 2009). Regression was carried out to explore whether PPM practices mediate the relationship between ambidexterity and an organization's performance (Table 7). In effect, our findings confirmed the general hypothesis that not only does the existence of ambidextrous capabilities significantly contribute to the performance of projects (H2b), but also that PPM capabilities do moderate the relationship between organizational ambidexterity and PPM success. Our findings thus support earlier literature suggesting that more ambidextrous capabilities will lead to higher levels of business and operational performance. As shown in Table 5, the coefficient b (i.e. the slope of the regressed line) for ambidexterity in Model 1 was positive and highly significant (b ¼0.525 at sig < 0.01), therefore, supporting this finding. Our proposition that the more an organization or its business units exhibit an effective application of PPM practice (measured by PPM effectiveness), the higher the level of ambidexterity (as presented in Model 2), is supported by the highly significant positive b coefficient (b ¼ 0.674 at sig < 0.01) (H1b).
To further explore whether PPM mediates the relationship between ambidexterity and an organization's performance, three steps were followed. First, we established that the independent variable (i.e. 'PPM effectiveness') influences the mediator (i.e. 'Ambidexterity'). This was established and supported in Model 2, where b ¼0.674 at sig < 0.01. Second, we established that the independent variable ('PPM effectiveness') influences the dependent variable (performance). This was established in Model 3 at b i0.480 and sig < 0.05, as shown in Table 5. In the last step, the challenge was to demonstrate that the mediator (ambidexterity) influences the dependent variable (performance) with the independent variable ('PPM effectiveness') controlled. If, in this last step, the effect of PPM effectiveness on performance is found to be no longer significant when the mediator 'ambidexterity' is in the model, then full mediation can be indicateddepending on how significant the model was.
Recognizing the above, and as shown in Model 4, a multiple regression analysis was performed using the two variables as independent variables -'PPM effectiveness' and 'Ambidexterity'. The b coefficient for ambidexterity was found to be positive and significant at b ¼ 0.445 and sig < 0.05. The VIF for all variables in the four models was measured and found to be less than 2.0, indicating that collinearity or multicollinearity did not exist in these models, entailing their results are acceptable (Myers 1990;Field 2009).

Discussions
The study set out to explore to what extent PPM practices impact orchestrations of ambidexterity, and whether these practices further impact the abilities of organizations to channel their ambidextrous capabilities towards superior project performance. Generally speaking, these relationships were found to exist. However, our findings do also bring to the fore a number of discussion points which we theorize below in the light of concepts drawn from organizational ambidexterity theory.

PPM practices and the successful orchestration of organizational ambidexterity
As relates to the first part of the research question (To what extent do PPM practices impact upon the orchestration of ambidexterity?), we found higher levels of organizational ambidexterity strongly and highly correlated with higher levels of PPM performance (PPM effectiveness and PPM success). Also pertinent here is literature espousing strong ties between collaborating projects within a portfolio and enhanced ambidexterity (Tiwana 2008). On this view, 'portfolios', because of their boundary-spanning leveraging competencies, are more able to cater for those inconsistent routines that are a hallmark of organizational ambidexterity (see Stettner and Lavie 2014). In other words, by using a portfolio of projects, organizations may sidestep tensions which would be harder to overlook should they arise within a single project; moreover, portfolios allow for concurrent pursuit of project delivery strategies that employ both exploitative competencies (focussed on increasing existing market share) and explorative competencies (focussed on new revenue streams). When ambidexterity is orchestrated through portfolios, the organization can better ensure that with each project representing an alternative implementation platform, each project maintains consistent and focussed processes.

PPM's ability to drive project performance through enhanced ambidextrous capabilities
As relates to the second part of the research question (Do PPM practices impact upon an organization's ability to channel its ambidextrous capabilities towards superior project performance?), we found that greater (both effective and efficient) use and application of PPM practices led to enhanced orchestration of higher levels of ambidexterity (further entailing that the more an organization would be able to channel ambidextrous capabilities towards positive project performance). The literature suggests that project portfolios which ensure tight resource-based, contextual and structural synergies between their individual constituent projects are, because of their improved process and practices integration and complementarity (see Sun and Lo 2014), better able to develop expertise for delivering overall PPM effectiveness (Tiwana 2008). This suggests that an 'exploitative' perspective of PPM may outperform 'explorative' perspectives. The underlying explanation, here, is that an organization's PPM expertise may become more efficient and effective with iterative use for the same PPM delivery processes. In suggesting this, Bierly and Daly (2007) noted that despite its relatively low optimal level, exploitation, compared to exploration, tends to be the stronger performance driver. Furthermore, we also reflect critically on our general finding that ambidextrous competencies and capabilities may increase the performance of an individual operating entity. This may happen, Cao, Gedajlovic, and Zhang (2009) contend, because at the core of ambidexterity is the capability developed by an operating entity to innovate while retaining long developed efficiency competencies. There is, however, a weakness to this argument. The weakness stems from our recognition that in seeking to increase performance, projects may draw on complementary benefits from being part of a portfolio, in that they are able to draw upon and develop expertise which is either outside their pre-existing project boundary, or more specifically, outside pre-existing portfolio boundaries (as in the case of organizational participants in strategic alliances). Under such circumstances, project performance (assessed as PPM effectiveness and PPM success) will depend on how well projects balance resource constraints against the ability to draw on complementary strengths from outside traditional portfolio networks. This is because during intense competition within internal environments, it is unlikely that other portfolios will be particularly willing to share scarce resources (Tsai 2002). To conclude, it may be surmised that project portfolios are only able to serve as mechanisms for orchestrating ambidexterity and at the same time, for enhancing performance, when their activities are robustly coordinated at project level.

Conclusions
The study undertook to ascertain whether effects of organizational ambidexterity on project performance are stronger for orchestration of ambidexterity at portfolio level. Its strong positive finding, specifically that PPM practice does indeed enhance prospects for ambidextrous management practice to contribute to project success, clearly underscores the need for project portfolios to be designed and operated with such enhancement in mind. And yet it also clarifies that future studies require more theoretical nuance to tease out the specific portfolio effects at issue, being mindful in particular to recognize that positive and negative effects may occasionally be bundled together. This further entails, we would suggest, that practitioners conceive of portfolio design with some openness to the possibility that there is always a calculus of positives and negatives to consider for purposes of design optimization.
In enabling us to conclude with the above recommendation, our research has been deeply concerned with the actuality of practice in both project and operations management. We designed our studies in a manner consistent with quantitative studies, recognizing in particular persistent and ongoing calls for more empirical research in operations management (Fisher, Olivares, and Staats 2020; Terwiesch 2019; Terwiesch et al. 2020) and project management S€ oderlund 2016, 2018). We also acknowledge that there is also, currently, very little empirical research that has so far revealed core project-related interdependencies engaged in orchestrating organizational ambidexterity, or, indeed, processes (mechanisms) by which PPM practices can mediate the relationship between organizational ambidexterity and project performance. Petro et al. (2020) observe that despite the growth in ambidexterity research over the years, a number of practical questions about the precise mechanisms involved remain unanswered. Therefore, our study construed PPM as a standalone construct measured against a combination of scales. Ambidexterity was operationalized as a measurable scale construct, although construed as a distinctive variable encompassing the two paradoxical (but not necessarily, polarized) constructs. Our study paid particular attention to foundational characteristics of ambidexterity mechanisms earlier identified as likely to impact orchestrations of ambidexterity through PPM practices, and we recommend that future studies pay more attention to the specifics.
As expected for studies of this nature, ours did have limitations, which provide a platform for future studies. First, despite our recognition in the findings that management of the contradictions associated with ambidexterity could be impacted by salient organizational-wide attributes, our study specifically did not explore these attributes. Our second limitation relates to the recognition within literature that the relationship between organizational ambidexterity and performance may vary not only across levels, but also under different functional, structural and resource conditions. Accordingly, it can be argued that more detailed research is necessary, pertaining to dynamics of inter-project variation and integration and how these may impact upon designing organizations for optimal interaction (between different projects being pursued by the organization), and horizontal and vertical interaction with other areas of the organization. We think this would be a good approach for improving on present research through more granular studies of portfolio effects and optimization. Understanding the nature of interproject variation and integration should enable more value creation for the organization through enhanced understanding of how best to combine insights from not only other projects but also the wider organization, as a means to create the conditions for enhanced performance.

Disclosure statement
No potential conflict of interest was by reported the author(s). Terry Williams is Director of the Risk Institute, University of Hull. He previously served as Dean of Hull University Business School and before that, Director of the School of Management, University of Southampton. Terry has worked in both the public and private sector, beginning as a Lecturer in Operational Research at the University of Strathclyde, while completing his PhD research. He later became Professor and Department Head, where he worked in research and consultancy, modelling project behaviour and supporting $1.5bn þ post-project claims. Terry is a Fellow of the Institute of Mathematics and its Applications, a Chartered Mathematician, a Fellow of the OR Society and a certified Project Manager (PMP).

Item
Exploration items taken from Lubatkin et al. (2006) Exploration items (amended for our study) 1 Looks for novel technological ideas by thinking "outside the box".
Look for novel technological ideas by thinking outside boundaries. 2 Bases its success on its ability to explore new technologies. Base success on its ability to explore new technologies. 3 Creates products or services that are innovative to the firm. Create products or services that are innovative to the firm. 4 Looks for creative ways to satisfy its customers' needs. Look for creative ways to satisfy its customers' needs. 5 Aggressively ventures into new market segments. Aggressively ventures into new market segments. 6 Actively targets new customer groups. Actively targets new customer groups. Exploitation items taken from Lubatkin et al. (2006) Exploitation items (amended for our study) 7 Commits to improve quality and lower cost. Commitment to improve quality. 8 Continuously improves the reliability of its products and services.
Commitment to lower cost.

9
Increases the levels of automation in its operations.
Continuously improves the reliability of its products and services. 10 Constantly surveys existing customers' satisfaction.
Increases the levels of automation in our operations. 11 Fine-tunes what it offers to keep its current customers satisfied.
Constantly survey existing customers for satisfaction.

12
Penetrates more deeply into its existing customer base.
Enhance and improved what we offer to keep its current customers satisfied.