Can Project Team Members’ Willingness to Disclose Past Performance During Procurement Improve Organizational Business Process Success?

1. Introduction

Although global political and environmental tensions, cybersecurity threats, and the COVID-19 pandemic have fundamentally altered business processes worldwide, particularly through reduced business travel and increased reliance on virtual collaboration—we know that the likelihood of project success remains unchanged at only 50% [1,2]. We recently examined the project failure factors [3], but the results were inconclusive. Other researchers identified significant causal factors of project success versus failure as prior experience [4], leadership skills [5], certification/education [6], quality [7], contextual elements like project type, organizational maturity, and risk management [8,9,10,11]. In this study, we distinguish between the project team—those individuals actively working on project deliverables—and the end user community, which represents the broader population of users who will be impacted by, trained on, or utilizing the project outputs. This distinction is important as these two metrics measure different aspects of project scope and complexity. Even the most rigorous quantitative studies of project success produced marginal effect sizes, commonly below 5% [3]. Consequently, this study aims to close the literature gap by identifying an additional causal factor that impacts project success, namely team member willingness to share past performance evaluations.

A significant factor impacting project success is unknown yet obvious: team member capability. It is difficult to know a team member’s capability without testing them or reviewing past performance evaluations. Aptitude testing is not a commonly accepted method for project team member selection, and aptitude tests may not indicate team member capability in the dynamic project context. Project managers can determine team members’ hard and soft skill capability by reviewing past performance evaluations [12].

There were some studies showing that the combined skillsets of team members at the group level of analysis impacted project success [13,14]. However, there were no studies about how team members’ individual past performance evaluations influences project success, which makes it a relatively unknown causal factor. It is logical to hypothesize that team member capability would significantly impact project success. Rarely does one person solely control a project’s success, so we argue that this is why no research has determined why projects are not always succeeding. Most prior studies focused on project managers or organizational context rather than the individual team members executing the project tasks.

We take a different perspective when studying team members’ impact on project success compared to most human resource management (HRM) literature. Studies of team member impact in the HRM literature commonly use individual dependent variables like performance rating evaluated by line managers rather than project managers. Additionally, we argue that it is already clear from the psychology and HRM literature that past performance predicts future performance, so studying past performance indicators as predictors of project success will not add new meaning to the body of knowledge.

Our perspective is that given today’s high data privacy standards, performance evaluations of new team members may not be openly available to project managers (PM). Since PMs are not formal line supervisors, they may not have authorization to access team members’ past performance records. We argue that PMs ought to have access to team members past performance. We posit that if team members volunteer to share their past performance evaluations with PMs, this will eliminate one of the severe unknowns in project management internal risk management. Subsequently, we posit that when team members refuse to disclose their past performance evaluations to PMs, this will create an unknown and subsequently, in the long run, reduce project success because employees are more likely to withhold unfavorable past performance evaluations, and this disclose/withhold behavior is an indicator of future behavior and indirectly project success. In simple terms, team members will want to hide bad behavior but be willing to disclose good prior performance evaluations voluntarily, and this disclosed/undisclosed indicator will impact project success. Our research question (RQ) is whether a team member’s willingness to disclose past performance impacts project success in the context of already known causal factors such as project size, complexity, and experience.

Organizational decision-makers and PMs need to know if a team member’s willingness to disclose their prior performance evaluations impacts project success. PMs can use that to inform their team member selection decision if they can choose team composition. On the other hand, if the PM does not have the authority to choose team members, they may still be able to use this indicator for work breakdown structure allocations. If the PM knows a proposed resource is unwilling to share and disclose their prior performance evaluation, it could assume the resource is high risk and allocate them to a less critical task where other team members can absorb weaker peer performance. While one indicator may not fix the high project failure rate, knowing such a significant predictor could influence future research and organizational culture to ensure PMs are given copies of team members’ prior performance evaluations. Access to a team member’s historical performance record would allow the PM to observe the employees’ apparent strengths and weaknesses and inform allocation and mentoring decisions in a project.

Since our RQ is a cause-effect type construct and references past performance in project management, we may assume this will be a quantitative study where historical or retrospective data are sought. We can also assume quantitative techniques will be used to test hypotheses to support the RQ. Accordingly, this study aims to determine whether team members’ willingness to disclose past performance evaluations improves project success, after accounting for known causal factors such as project size, complexity, and experience. The remaining sections of this manuscript review the literature, outline the methods, and present results followed by implications and conclusions.

2. Review of Literature

Several researchers have examined honesty in sharing performance appraisals, which implies willingness to disclose past performance, whether that may be good or bad. Past behavior predicts future project performance behavior [4,5]. Therefore, a PM could leverage past behavior in resource allocation decisions. Since projects are temporary endeavors, a PM may not have access to team members past performance evaluations due to organizational culture or company policy. Certainly, if past performance evaluations are available to the PM, they can be leveraged for allocation choices. Regardless of whether the PM has access to prior performance reports of team members, the PM may strategically ask the resource for a copy and base allocation decisions on that answer.

Trust and transparency are central to effective performance evaluation. Fair and open evaluation processes strengthen employee confidence and willingness to share feedback, fostering continuous improvement. Employees may sometimes undergo additional training or read materials to overcome the negative aspects identified in their performance evaluations. These elements carry through to the employees’ honesty and transparency to the new PM.

Motivation and morale are fundamental in performance evaluations. If the organization values integrity and if it is committed to helping employees grow, employees can be motivated by that and spread positive morale to other team members, even helping them to improve their performance in projects. Accountability is a well-known factor in performance and the evaluation process. Honest appraisals done by supervisors can hold team members accountable for their performance (good resulting in organizational praise or bad resulting in being overlooked for promotions). When employees know their performance is being evaluated truthfully, they are more likely to take responsibility for their work and strive for better project performance. Likewise, past performance sharing has a carry-forward effect on future projects. When team members know that a PM may request their past evaluations, they tend to perceive project performance as part of their long-term professional record, encouraging accountability and consistent effort across projects. In other words, this concept of accountability in project performance evaluation would increase project success if all team members knew their performance evaluation would be shared with future PMs.

Improvement and development are underlying reasons for conducting and sharing performance valuations across projects. Honest feedback identifies improvement areas, providing a clear path for promotions. This can lead to enhanced skills and better performance on future projects. When team members realize that sharing past performance evaluations could lead to skill improvement and promotions, they are more likely to be willing to disclose past performance evaluations so that a PM will gain an objective view of strengths and weaknesses.

Organizational culture impacts projects. If the organization has a work environment where honesty and transparency for performance evaluations are highly regarded, team members will see this as a benefit for improvement, not a hindrance. The organization can use open communication and promotional techniques to ensure team members are exposed to this culture. Once team members see the organization is committed to honesty and transparency, they may feel more willing to share their past performance evaluations even with new PMs they do not yet know. This increased transparency in performance evaluation will improve project success because knowledge is better than the unknown. This supports the notion that transparent performance information is preferable to uncertainty when managing project risk.

2.1. Factors That Impact Project Success

The literature is clear that several factors impact project success in most projects. The better-known factor impacting project success is the size of the project in terms of budget, which reflects complexity, meaning the as complexity and budget increases, this tends to have an inverse relationship with project success [8,15,16]. The size of the organization and end user community impacted by the project have also been commonly found to impact project outcome - generally, projects affecting larger end user communities experience more complexity and coordination challenges, which can lower overall performance [9,15,16]. More years of experience, sometimes described as seniority, has also been found to improve project success for leaders and team members [4,9]. Many researchers have found that gender does not predict project outcome, mainly because the discipline is dominated by males [4]. Therefore, we assert it will be of no value to measure gender unless the sample is approximately equally balanced. Certification has been found to increase project success while more education tends to have the opposite effect [4]. We argue the common factors discussed above are a priori meaning that we expect a sample population to follow these generally accepted characteristics.

To control established project characteristics documented in prior research with this dataset [17], we included project budget, end user community size, and professional certification as control variables. Previous analysis of this sample confirmed that higher budgets (r = -0.041, p < .01) and larger end user communities (r = -0.241, p < .001) negatively predict success, while certification shows modest positive effects (r = 0.099, p < .001) [17].

Leadership and project management skills are equally important for the success of projects [6,18]. For instance, effective leadership can ensure that projects are planned adequately and that risks are mitigated, helping to keep teams motivated [14]. Effective project leadership can ensure that clear direction is provided to the team members and conflicts among team members are resolved immediately ensuring collaborative work atmosphere in the team [19]. The composition and skillsets of the team members also contribute to the success of projects and having the right mix of skills and experience can ensure that the team can handle technical challenges as well as interpersonal dynamics that may come up during the project lifecycle [13,14]. We developed this hypothesis to assess if PM leadership skills positively increased project success:

H1: Better project leadership improves project success.

Resource availability, including time, money, and technology can determine the success of a project as well-funded projects are likely to have access to modern tools and technology [20]. Resource constraints can lead to delays or scope reductions negatively impacting the outcomes of projects [21,22]. Organizational culture can also influence the outcomes of projects as an organizational culture promoting innovation, flexibility, and accountability can create a conducive environment contributing to project success [23]. We developed this hypothesis to determine if a supportive organizational culture impacted and in fact improved project success:

H2: Supportive organizational culture adds to project success.

Risk management is crucial to the success of projects because projects are inherently risk and the ability of the project manager to identify, assess, and mitigate risks will determine the success of a project [24]-. Effective risk management will ensure that the risks are identified in time before they affect a project and that risks are handled in time [11]. Stakeholder engagement is also essential as this will ensure that projects are active and supportive stakeholders can ensure that projects face minimal roadblocks as stakeholders will help secure resources, facilitate communication and advocate for projects within the organization [25]. When combined, risk management and stakeholder management require regular communication. Therefore, stakeholder communication management is essential for project success. Stakeholder communication management requires regular updates, clear instructions, and open channels for feedback to ensure that minimal misunderstandings occur and are also required to ensure that teams can work effectively [26,27]. We developed this hypothesis to measure if using stakeholder communication management within a project would contribute to project success:

H3: Effectively applying stakeholder communication management helps a project to succeed.

Change management is also an important factor as projects undergo changes in scope or objectives. Hence, the ability to manage these changes without disrupting the project timeline or budget is essential for the project success [24]. Change management can be effectively managed through efficient monitoring and evaluation mechanisms that ensure that progress is tracked and deviations from the plan are identified early [28]. When deviations are identified early, project managers will have the opportunity to assess the performance against key metrics and make necessary adjustments increasing the chances of the project success. We developed this hypothesis to assess whether effectively managing change was critical for project success:

H4: Effectively managing change is critical for project success.

2.2. Team Members Past Performance Impact on Project Success

Experienced team members have a significant positive influence on the success of projects [15,29]. Team members with a history of strong performance bring with them valuable experience, skills, and problem-solving abilities contributing to better project outcomes [30,31]. The prior experience of the team members helps the teams’ meet deadlines and also ensures effective collaboration among the team members. Team member experience, in terms of years, has been proven to be a predictor of project success so this is a necessary control parameter. To test this a priori theory that a team member’s years of experience impact project success, as more experience will signify a resource with more capability, we developed this hypothesis:

H5: Higher team members’ years of experience increase project success.

Team members with a history of poor performance can become a liability to the project as past failures may be an indication of underlying issues, including lack of motivation, poor time management, and inability to handle pressure [12]. Such individuals may struggle to meet the expectations and could cause delays in the project timeline and ultimately lower the quality of the project deliverables. Experienced team members with a prior history of success can be role models for the new members of the team and can inspire confidence and boost the morale of the team [32]. These high-performing team members set standards for other members to follow by contributing innovative ideas and finding creative solutions to problems [33]. These factors ultimately contribute to the success of projects.

However, the challenge for project managers would be ensuring the willingness of team members to disclose their past project performance. Transparency is essential in performance evaluations to ensure that there is trust between the team members and the project members [34]. When there is trust within the teams, the members will voluntarily share their performance history which will allow project managers to make appropriate decisions about task assignments [35]. In this context the team dynamics are quite critical as the trust within the teams also depends on the internal team dynamics and culture [36]. Project managers can build trust among team members who openly acknowledge past weaknesses and demonstrate commitment to improvement by coaching and mentoring the team members. This approach will ensure that team members will be open to self-improvement and will be better equipped to handle the challenges of complex projects.

Trust, transparency, and willingness to share a performance evaluation can be tested as a single factor because they are interlocked. It would be difficult to test these three elements separately. Additionally, the PM may or may not have access to all the team members’ performance evaluations, so we propose that the PM can indicate if most of the team members consented to provide their last performance evaluation, which we argue represents trust, transparency, and willingness to share performance evaluations.

We also believe willingness to show a performance evaluation may be an indirect mediator or moderator of other factors’ impact on project success. First. we argue that if a team member is willing to supply their last evaluation to inform project team member selection, this will allow the PM to choose only resources with past behavior most conducive to the context, thus decreasing the end user community size and improving project success - that would be a mediator of the end user community size factor. On the other hand, we argue willingness to share a past performance evaluation during the team selection process would divide the staffing pool into those with good past performance (those saying yes) and those with unfavorable items in their past (those saying no). In that situation we argue willingness is a moderator, a conditional factor representing resource capability, but if the performance evaluation is not being shared this may cause the PM to select too many resources in a guessing game of which is a better fit. In that moderator situation, unwillingness may not directly impact performance, but it would increase the end user community size when the answer is no but decrease end user community size when the answer is yes, thereby a yes answer would cause a lower end user community size to improve project outcome, and vice versa. However, it is not clear in the literature if any such relationship may exist. Therefore, we propose the following hypotheses to test these propositions:

H6a: Team members’ willingness to share performance evaluations positively and directly impacts project success.

H6b: Team members’ willingness to share performance evaluations lowers (mediates) end user community size, improving project success (but willingness has weak correlation with project success).

H6c: Team members’ willingness to share performance evaluations negatively interacts with (moderates) end user community size impact on project success (with no direct effects).

3. Methods

A post-positivist ideology was adopted which means the researcher will focus on quantitative data facts in the collection/analysis phase, putting emphasis on developing/testing hypotheses and reporting effect sizes with robust statistical techniques in the results phase [37]. The RQ was whether a team member’s willingness to disclose past performance impacts project success in the context of already known causal factors such as project size, complexity, and experience. This indicates a correlational and predictive design, depending on which factors are tested. Since one factor is a mediator or moderator, correlation must be used to first check the expected indirect relationship between the mediator and the dependent variable (project outcome). We also expect to not see correlation between a moderator and the dependent variable (willingness and outcome), which will require a partial correlation or regression with partial effects type of test. Once we have tested our hypotheses using correlation, we will develop a predictive model using and partial least squares effect size technique to test moderation and mediation impacts.

As is common practice, we start the analysis by describing the sample using descriptive statistics. We then use correlation and regression to test the hypotheses, reporting probability values (using a 95% confidence level) and effect sizes. Regarding effect sizes, we applied the suggestion of Pierce, Block [38] to report partial ETA estimates rather than ETA alone, because the former accounts for variances of the effect on the total after removing the intercept. Following Pierce, Block, and Aguinis [38], effect sizes were reported using partial eta², where 2% indicates a small, 5% a medium, and 10% or higher a strong regression effect. In other words, they argue that the effect size should not include the basic variance of the slope. An analogy is that comparing household electricity consumption across design models must remove the basic differences in obtaining service and the basic load on the wire when there is power, but every appliance is turned off, since the wires and capacitors in the various devices will initially absorb some current. To that end, following Pierce, Block [38]’s reasoning, we will claim an effect size of 2% is minimal, 5% is medium, and 10% or higher represents a strong regression model.

3.1. Measurement Instrument

The first author served as the principal investigator (PI), being responsible for research design, instrument development, and data collection. The PI developed the instrument to be administered to project team members of a large financial and insurance technology industry (Fintech) company. The purpose of the survey was to add additional data to existing longitudinal project records. The existing human resource information system (HRIS) contained records of common attributes associated with FinTech’s projects over the last three years since COVID-19 (with records roughly beginning on January 1, 2021). The following fields were available in the project database: employee ID, user email, project ID and name, project type, project role (e.g., member or project manager), end-user community size, project budget, education level, PMP certification (yes/no), other certifications (listed by type), and employment date (used to calculate years of experience in the company), company goal alignment (for project), outcome, and date information fields including project start and stop. The end user community size represented the number of users impacted by the project deliverables (e.g., staff receiving training, users of new systems), not the number of team members working on the project. Additional data were available but not used in this study.

Records were extracted for team members of projects valued at least $1,000,000 USD. Years of experience were calculated using the employment date or date on the first project, whichever was more recent since some resources were recently hired. Therefore, experience reflected years as a team member on projects at the current company. The yes or no fields in certification were converted to ordinals (yes = 1, and no = 0). The outcome was a proprietary field representing a score of 50 points, with 25/50 considered the pass/fail point at the company.

A short survey poll was designed with help from the Fintech company using a Likert 1-5 scale for each question where 1 = strongly disagree, 2 = disagree, 3 = unsure, 4 = agree, and 5 = strongly agree. Questions were designed to capture the team members’ perception of how well the project manager’s leadership (leadership), company culture (culture), stakeholder management, and change management positively impacted project success in each specific project. A question was designed to ask the team members if they would be willing to share their previous performance evaluations with the project manager for the selection decision, should they be contacted to potentially participate in a similar project (willingness). The response scale was 1=yes or 0=no.

This study extends our previous analysis of this dataset [17], which examined budget, end user community size, goal clarity, and certification as predictors of fintech project success (R² = 0.096). The current study introduces team members’ willingness to share performance evaluations as a novel predictor, hypothesizing this factor would substantially improve explained variance beyond established project characteristics. The survey instrument was specifically designed to capture willingness to disclose past performance, along with perceptions of leadership, culture, stakeholder management, and change management factors.

3.2. Sample and Data Collection

For the selected projects, the team member ID was queried for unique records to extract the email addresses of each team member. An electronic survey was designed to incorporate the above instrument and to obtain informed consent. The Fintech company had revenues exceeding $1 billion USD, more than 500 employees, and they were based in the USA. The survey was sent to each member, and responses were collected online within two weeks during 2024. Once the data were cleaned up to remove incomplete or duplicated response records from the same team member on the same project, there were 518 unique valid responses. This represented a 41.5% response rate from the 1,248 team members originally contacted.

4. Discussion

4.1. Preliminary Analysis

Wave analysis was performed to ensure the data collected in the first week matched dependent variable variance with the second week. The week-by-week results showed no statistically significant distributional difference. In other words, there was no obvious situation where keen project team members responded first, and lackluster members were late responding.

Table 1 lists the sample’s descriptive statistics, starting with the mean and then the standard deviation (SD), followed by the correlations flagged for significance. A few variables in Table 1 can be interpreted by looking at the frequencies rather than sample estimates (only mean and SD are shown). Descriptive statistics for control variables replicated our previous analysis [17]: budget (M = $3.03M, SD = $4.87M), end user community size (M = 164, SD = 142.1), and correlations with success showed expected patterns. Overall, 19% of the participants had a graduate degree, 20% had a bachelor’s degree, 20% (rounded) had an associate, 19% had an associate, and the remaining had only grade school education. Only 4% had the PMP professional certification, so the mean is zero in Table 1. The Fintech project experience was 2.5 (rounded, SD = 0.7). Approximately 30% of the team members indicated they were willing to share their performance evaluations for project selection, leaving 70% who were not. We could say roughly a third of the members were willing to disclose their past performance evaluation. Most means for the 1-5 scale survey factors were at or slightly above 3 with deviations from 1-2 (SD = 1.4).

Another interesting observation from Table 1 was the project success variable. Outcome had an average of 30.7 (SD = 14.2). We also calculated the median of 37. When considering the Fintech pass/fail break-even point for project success was 20/40 (50% score), the larger median of 37 against the mean of 31 (rounded) may be interpreted that most of the Fintech projects were successful because there are more values in the data above the mean of 31. To determine the overall success rate, we calculated the percentage of projects with outcome scores ≥25 (the company’s pass/fail threshold). Based on this criterion, 71% of projects were successful. The median outcome score of 37 exceeds both the mean of 30.7 and the threshold of 25, further indicating that the majority of projects in this sample achieved success.

4.2. Hypothesis Test Interpretations

Before examining our focal hypotheses on willingness to share performance evaluations, we tested the exploratory hypotheses on leadership (H1), organizational culture (H2), stakeholder communication management (H3), change management (H4), and team member experience (H5). Consistent with prior findings from our initial analysis of this dataset [17], none of these factors showed significant correlations with project success. Specifically, leadership had a correlation of r = -0.01 (ns), culture showed r = -0.01 (ns), stakeholder management had r = 0.009 (ns), change management showed r = 0.005 (ns), and years of experience had r = 0.107 (ns). Therefore, we rejected hypotheses H1 through H5. These null findings replicate our previous research [17], which found that in this fintech organization, these traditional project management factors did not significantly predict project outcomes when controlling for budget, end user community size, and certification.

Control variables performed consistently with prior published findings using this dataset [17]. Budget (r = -0.041, p < .01), end user community size (r = -0.241, p < .001), and certification (r = 0.099, p < .001) showed expected negative and positive patterns respectively. Leadership, culture, stakeholder management, and change management showed no significant correlations with project success (all r < .02, ns), consistent with our previous analysis [17]. Given these replications, we focus our hypothesis testing on the novel willingness factor.

We found support that team members’ willingness to share performance evaluations positively impacted project success. To initially test if team member’s willingness to share performance evaluations positively and directly impacted project success (H6a) we applied a student t-test on outcome grouping by willingness. The result was acceptable (T [DF = 516] = -28.099, p < .001). This result can be interpreted as group 0 (the lowest value in the data) had a negative mean difference, it was lower than the second group 1. We calculated Cohen’s D as the effect size (the equivalent of partial ETA squared), which was -0.869, indicating a large effect size according to conventional standards (Cohen’s D > 0.8). That is very high and when considering this with the 28% or participants who indicated they would be willing to disclose their past performance evaluations for new project selection decisions, we can say yes, willingness results in better project outcomes in our sample. We will accept hypothesis H6a while we continue examining the mediation and moderation hypotheses.

We did not find support for our hypothesis that team members’ willingness to share performance evaluations mediates the relationship between end user community size and project success (H6b). For mediation to occur, willingness must correlate with the mediating variable (end user community size), but Table 1 shows no significant relationship (r = -0.003, ns). Additionally, willingness showed a strong direct correlation with outcome (r = 0.364, p < .001), contradicting the H6b prediction of weak direct effects. Therefore, we reject H6b.

We did not detect obvious support for our hypothesis H6c that team member’s willingness to share performance evaluations negatively interacts with (moderates) end user community size impact on project success (with no direct effects). End user community size initially had a significant negative correlation of -0.241 (p < 0.001) with project success, consistent with prior research indicating that larger projects face greater complexity. When end user community size was moderated (conditioned) by willingness, the correlation between end user community size and outcome changed from negative to a positive but significant 0.037 (p < .01), suggesting that willingness reverses the typical negative relationship between project size and success.

For this to be a moderator, we would have expected the correlation to improve in certain conditions and statistically that the willingness factor was known prior to project outcome, even though in our study it was a survey question completed after the project. Therefore, we needed to examine this factor willingness in a partial least squares model, using linear regression, on the dependent variable outcome. We developed three models. The first model was a basic one with only the slope.

The second model included all measured variables as predictors: end user community size, budget, education level, certification, goal alignment, leadership, organizational culture, stakeholder management, change management, team member experience, and willingness to share performance evaluations (11 predictors), in order to determine the maximum potential impact on project success. The key statistical estimates indicated the second model was useful. The coefficient of determination (r² = 0.096, adjusted r² = 0.093, RMSE = 13.530, F [25,514] = 21.919, p < .001 (with minimal variable inflation or collinearity intolerance). We can state that model 2 with all the factors, have a moderately strong effect size of 9.3% (adjusted for all factors entered).

The third model contained only the significant factors of end user community size, certification, and willingness in stepwise fashion, regressed on outcome, by entering and removing any factor with an insignificant p value. This third model was significant as can be seen from the key statistical estimates of r² = 0.188, adjusted r² = 0.188, RMSE = 12.79, F [4,514] = 299.684, p < .001 (with minimal variable inflation or collinearity intolerance). Now we can state this third model is better, with a very strong 18.8% effect size, double the statistical power of model 2, and using only three factors to predict project outcome.

Table 2 lists the key statistical estimates of the third model according to the factor level of analysis, namely the beta or standardized coefficient (B), the standard error (SE), t estimate, and probability (p) value, while remembering that our confidence level was set at 95%. The linear regression coefficients are shown for categorical factors willingness and certification, while the standardized beta coefficients were calculated for the continuous ratio data types of size and budget in Table 2. The beta coefficients indicate the relative impact of each factor on the dependent variable project outcome. A negative value suggests higher levels decrease project success. Control variables performed as expected based on prior research [17]: budget (B = -0.224, p < .001), end user community size (B = -0.043, p < .001), and certification (B = 6.534, p < .001). Most importantly, willingness to share performance evaluations emerged as the strongest predictor with the greatest positive impact on outcome (B = 10.834, p < .001), representing the novel contribution of this extended analysis.

To further examine the strategic impact of willingness on project success we computed the partial least squares effect sizes from the third regression model, which are listed in Table 3. We could assert that the partial correlations in the second column of Table 3 is a conditional correlation suggesting moderation. Partial correlation between two variables in Table 3 was under the assumption that we know and consider these values to predict or show association on project outcome. These could be considered as beta coefficients, relative to one another, so higher betas mean more influence with the possibility of a change in direction if negative. The next column in Table 3 is the part correlation, the more detailed specific impact of the factor on project success, with the other factor effects removed. It could be considered a lower level of measurement, more granular than partial, a pure unique factor beta regressed on project success.

Control variables showed expected patterns [17]. The critical finding is that willingness to share performance evaluations was clearly the strongest factor in our final model, with a part beta coefficient of positive 0.341 regressed on project outcome. This is almost 50% stronger than budget, four times larger than certification, and eight times more influential than end user community size. When we consider that certification is a binary variable (1 = yes, 0 = no), the positive beta of 0.081 shows it has twice the power in the model compared to end user community size, but only a quarter of the impact that budget has on project success. Nonetheless, being certified is associated with better project outcomes, though willingness remains the dominant predictor.

What does this tell us in terms of implications? It shows that selecting project team members who are willing to disclose their past performance evaluations will dramatically increase project success, and the decision maker may also consider the budget as well as the end user community size. Certification ought to also be a factor in the decision-making process since we can see it has a positive although small impact on project success. To put all this in perspective, the model shows that it is possible to predict project success with lower budgets, lower end user community sizes, selecting members with certification, and especially those employees who are willing to share their past performance evaluations, which captured approximately 19% of the variance, or chance, towards a failing versus a successful project in our sample. The big question is how widely this model generalizes. We address that below in the next section.

5. Conclusion

This study’s primary contribution demonstrates that team members’ willingness to share past performance evaluations substantially improves project success prediction, increasing explained variance from 9.6% (control variables only [17]) to 18.8% when willingness is included. We argue that a decision maker, a project sponsor or project authority would want their project to be successful and therefore they would now want to facilitate giving the PM access to team members past performance evaluation. However, that is not always possible due to bureaucracy or political constraints. The next best option would be to forecast this scenario by asking employees (proposed team members) if they would be willing to share their past performance evaluations just for the single purpose of project selection. The team members may not realize the strategic significance of their answer—but an informed PM would. The answer to that single question could improve project outcome prediction, with the model explaining 18.8% of the variance in project success—nearly double the 9.6% explained by traditional factors alone. This suggests that incorporating willingness as a selection criterion could substantially improve project team allocation decisions.

This study makes a theoretical contribution by demonstrating that transparency and trust—operationalized as willingness to disclose past performance—substantially improve project outcomes beyond traditional resource allocation factors. While prior research [17] established that budget and end user community size negatively predict success in fintech contexts, the current findings reveal that human factors related to performance transparency explain twice the variance (18.8% vs 9.6%). This suggests that ‘who is willing to be transparent’ matters more than ‘how many’ or ‘how much’ in predicting fintech project success.

The main limitation of our study was that we essentially used sequential mixed methods. We already have a company database of project performance records from the HRIS. We used the survey method to collect additional data from project team members. Naturally, there will be auto-selection based on behavior. If a member knew they had good performance, they were more likely to participate in the mandatory survey issued by the company before the deadline with truthful answers. Any self-reported answers are always subject to validity. In the future we recommend using retrospective data as much as possible and ensuring the company records useful metrics of projects to facilitate academic studies.

The second limitation in our study concerns the timing of the data capture for the willingness factor which was done after the projects ended. In a perfect environment we would capture this first, and compare factually to past performance, then use that as part of the project team member selection criteria. However, on our study, the project came first, and the answer to sharing the past performance evaluation came later. This was not a true cause-effect order, that is, it was a self-reported perception by the team members that they would disclose their past performance evaluations to the PM for a future project selection exercise like their previous projects. We do not know 100% if that was true. In the future, a controlled environment or an experiment could be constructed to better control that before-after time sequence between a factor (willingness to share performance evaluation) and the dependent variable (project success outcome).

A third limitation concerns the sample context. This study examined only one fintech company, despite capturing numerous projects over several years. While this provided consistency in organizational culture and processes, it limits external validity and generalizability to other industries or organizational contexts. Additionally, the 41.5% response rate, while reasonable for organizational research, may have introduced non-response bias. Groups of participants may not have responded due to external factors such as poor timing, holding negative sentiments about the project (and not wanting further involvement even for research), or simply being too busy. This last constraint could have had a more pronounced statistical impact if those busy individuals were also more knowledgeable and experienced compared to those who responded, potentially biasing our results.