INTRODUCTION

Key performance indicators (KPIs) are used to evaluate the performance of a sports team, department, staff, and/or athletes themselves (¹²). They are commonplace in business (⁵²) and sports (^30,35) on account of the many benefits they offer. Some examples include enabling objective evaluation of an organization, team, or individual in meeting performance goals (²⁸), informing training and practice decisions (²⁵), and directly being able to predict competition performance from KPIs (^44,60).

Despite their ubiquity and copious benefits, some issues still exist with their use in sports performance. These include potential detrimental effects of measurement itself on athlete behaviors (^24,41), and concern with how well KPIs can truly approximate performance in complex dynamic events such as team sports (³⁵). Furthermore, the increasing attention on sports performance support and abundance of available data may induce pressure to use KPIs to quantify the value of performance support and try to demonstrate return on investment (^12,31). Conversely, practitioners are keen to demonstrate how the data amassed and the subsequent interventions used have a positive impact on performance. Therefore, the purpose of this article is to discuss the issues with using KPIs in team sports performance support. Based on these issues, we propose a new framework, ongoing process trackers (OPTs). This approach, which could be used in conjunction with KPIs, is intended to help sports performance practitioners better select and track indicators of the processes that ultimately underpin sports performance.

CURRENT LANDSCAPE OF KEY PERFORMANCE INDICATORS IN SPORTS PERFORMANCE

Striving for peak performance is a central tenet of sporting endeavor. Thus, it follows that sport staff will seek indicators that underpin performance on an organizational, departmental, and individual level. Indeed, a recent organizational staff structure for team sports proposed that performance staff, including practitioners across sports science, strength and conditioning, nutrition, and psychology, and team sport staff, made up of coaching and scouting staff among others, each provide expertise related to specific KPIs (¹¹). Although the use of KPIs is relevant for team sport staff, in this article, we focus our discussion predominantly as it relates to performance staff.

One recent framework (¹²) has outlined the following steps for performance staff to use KPIs in a sporting setting.

• • Identify what it takes to succeed
• • Define the performance model
• • Determine KPIs
• • Assess the athlete
• • Plan and deploy the program of interventions
• • Review

A performance model can be dissected into the following determinants: physiologic demands, technical requirements, tactical requirements, psychological skills, equipment characteristics, health aspects, and rules and regulations (¹²). In individual time–distance-based sports, the determinants of performance may be more straightforward. For instance, structural equation modeling in swimming explained 79% of performance in young male athletes based on biomechanical and energetic profiles (⁴). Only variables that can be assessed by performance staff were included in the model and given the high prediction of swim performance demonstrated, it follows that the variables identified can therefore be used for training control and evaluation (⁴).

Such division of determinants offers performance staff an understanding of the underpinning qualities to performance, such as the physical capacities that can be developed to enable the athlete to meet, and potentially surpass, the physical demands of the sport. Associations between jump power and heading success, and strength (predicted 1 repetition maximum from a 3 repetition maximum test) and tackle success in elite youth soccer may warrant development of such physical capacities, for example (⁵⁹). In addition, increases in reactive strength (measured via drop jump performance) have been associated with reductions in sprint times, whereas increases in power (via countermovement jump performance) were associated with improvements in change of direction abilities in elite female soccer players (²⁰). However, team sports performance offers greater complexity than can often be broken down into linear determinants of success (¹). As such, simple deterministic approaches, such as those used in individual time–distance-based sports, may be less suitable. Therefore, the current approach to applying KPIs in the team sport environment warrants a critical review.

ISSUES AND CHALLENGES WITH KEY PERFORMANCE INDICATORS

In reviewing KPIs through a critical lens, we have identified 4 areas of concern for their current application in team sports. They are as follows.

• • Definition: Lack of a universal definition
• • Complexity: Isolated metrics overlook the complexity of performance
• • Goodhart's law: The threat of drifting from on-field performance with off-field targets
• • Agency issues: Different stakeholders often have competing interests

In this section, we will delve into each of these issue areas in further detail.

LACK OF A UNIVERSAL DEFINITION

Although many researchers have aimed to formalize the definition of KPIs for different sectors (^18,36,40), there remains a lack of universal agreement, especially in sport. Given the widespread familiarity with the term, it may be believed that a universal definition is unnecessary. Yet, potential misuse of terminology is a long-time cause for discussion in sports science (^34,54). One recent definition in sports science is:

“a quantifiable measure used to evaluate the success of an organization or employee in meeting a performance objective” (¹²).

A performance objective is subsequently illustrated as winning a league, tournament, or other championship, achieving a specific time, distance, or mass lifted in centimeter-gram-second sports, or beating the opposition in tactical events (¹²). Clearly in this instance, clarity of the collective performance objective is required for KPIs to be effective. Yet, this definition of KPIs is firmly linked to a single-outcome measure. Meanwhile, other definitions use the properties of the measure itself, asserting that KPIs should be a valid measure of performance, an objective measurement using a known scale of measurement, and provide a valid way of interpretation (⁴²).

To add further confusion, similar but different terminology, such as “performance indicators” (³⁰) also exists. A performance indicator is “a selection, or combination, of action variables that aims to define some or all aspects of a performance” (³⁰). In team sports, the relationship between performance indicators and match outcome has been explored in Australian rules football (⁴⁸), soccer (¹³) and rugby union (³³). In this context, performance indicators are statistical actions that can be used to evaluate teams and individual athletes, including which indicators are likely to result in a winning outcome.

It is clear from these various definitions that a useful performance indicator should relate to successful performance or outcome. These applications traditionally sit within the performance analysis realm of team sports. However, these definitions lack guidance for how such KPIs can be used in practical terms. The focus on statistical actions during match play also limits their use to the wider multidisciplinary performance staff. It is therefore unclear how current definitions of KPIs relate to sports science data collection. Such ambiguity may fuel agency issues with how different stakeholders consider, interpret, and use data in relation to their interpretation of the term KPI.

ISOLATED METRICS OVERLOOK THE COMPLEXITY OF PERFORMANCE

As performance staff increasingly seek to relate sports science data to performance, it seems (at least to the authors) that KPIs may have become synonymous with broader terms such as metrics or variables, confusing the issue further. This is perhaps because of the abundance of metrics, thanks to the expansion of data collection in today's sporting environments. Indeed, a recent article described how the increase in data streams within the contemporary training environment may cloud parsimonious and valid applications if not used appropriately (³¹). It may also be because of the breaking down of sports performance into silos in traditional fields such as physiology, biomechanics, and performance analysis (⁵³), in which the term “performance indicator” may have a specific context.

Previous literature has categorized performance indicators into match classification (e.g., score, number of shots on targets), biomechanical (e.g., optimal release angle for javelin throw (⁷)), technical (e.g., passes to oppositions (³²)), and tactical (e.g., passes and possessions (³⁰)). Yet, consideration is warranted as to whether each of these categories and individual indicators truly reflect performance, and if so, how they combine to do so. One of the reasons why individual indicators may not truly reflect performance is because sports performance, particularly in team sports, is a complex entity (³). As such, team performance is not only achieved by qualities of the individual athletes (e.g., technical skills or physical abilities), but also the emerging pattern from the dynamic interaction between individuals, their opponents, and environments (⁵¹). Therefore, the impact of using isolated or specific measures as KPIs, such as physical running output captured by in-game tracking technologies, is often limited (^10,15).

Furthermore, different leagues may require different physical or technical KPIs to best approximate performance (¹⁶). As well as overarching consideration for context, such as league or position, more nuanced contextual factors are often lacking in developing KPIs. Phatak et al. (⁴⁶) recently demonstrated the need to account for phase of play and other contextual factors such as ball possession. For example, accounting for ball possession changes the interpretation of fouls as a KPI in soccer (⁴⁶). Caution is therefore warranted when it comes to labeling data as KPIs, particularly isolated metrics that lack contextual narratives.

DRIFTING FROM ON-FIELD PERFORMANCE WITH OFF-FIELD TARGETS

Performance staff attempt to translate their understanding of what it takes to succeed into performance goals (¹²). Drifting from these on-field performance goals is a threat, because it may result in time and attention being placed on areas less meaningful to performance. This may be driven by the different motivations of various stakeholders within team sports, as we shall discuss in the next section. In addition, using targets from objective measurements may underpin a drift from performance goals. This is exemplified by Goodhart's law. Named after British economist Charles Goodhart, the common adaptation of the decree is by Marilyn Strathern:

“When a measure becomes a target, it ceases to be a good measure.”

When outcomes are complex, as demonstrated by sport performance, a single measure cannot represent them. This is why there is rarely (if ever) a single KPI; the complexity of performance warrants multiple. Even within specific disciplines, physical development for example, a single measurement is often underpinned by a multitude of factors. This is illustrated by low predictive relationships between acceleration and lower limb strength and power, for example, indicating a complex interaction between sprint technique and leg muscle performance (³⁷). Therefore, if each indicator is imperfectly correlated with the goal of performance, concentrating on them in isolation may inadvertently create unwanted distractions and inefficient application of resources. This has been exemplified elsewhere in areas such as academic publishing (²¹), higher education (¹⁹), and government spending (²⁹) to name a few.

Although using KPIs as targets can drive processes that have positive effects on performance, Goodhart's law serves as a reminder that targets may distract from the goal of on-field sporting performance. Given the limited time and resources available to performance staff, these must be spent on the most impactful contributors to performance. Lower limb strength and power measures may frequently be converted into targets for athletic development, but although meaningful, they do not solely account for performance. Although strength and maximal power were the best discriminators of playing level in rugby league, they still only accounted for 12–17% of the variation in playing level (¹). Similarly, KPIs that focus on vertical force production, such as those derived from countermovement jump performance, may overlook the importance of horizontal force production and therefore, limit transference of gym-based strength to on-field performance (⁴⁷).

Beyond the issue of time efficiency, performance staff are in danger of “naive interventionism” if potential harmful effects of so-called performance targets are not considered (⁵⁵). Indeed, iatrogenics (“caused by the healer” in Greek) represents a treatment that causes more harm than good (⁵⁵). Goodhart's law may underpin iatrogenics in sports performance that result from turning measures into evaluations. In some cases, this may have been witnessed through the assessment of total distance covered as measured by tracking technology, which has been reported to have led to some athletes running around during breaks in play simply to increase their distance covered (²³). Similarly, velocity based training provides objective feedback on bar velocity numbers during strength work, which can improve training intent, but with poor implementation could encourage athletes to chase targets to the detriment of technique, and potentially safety (²³).

Another example is performance staff may set targets to address asymmetries assessed through strength and power testing, such as <10% interlimb asymmetry for return to sport from injury (^8,38). However, clinicians have also acknowledged that the unaffected limb is rarely normal (⁴³) and research has yet to demonstrate a clear influence of asymmetry on performance (^8,39). Further research is warranted to understand whether intervention planning to target asymmetry reduction is beneficial (⁴³). A final example may be setting bodyweight targets, which could lead to athletes engaging in unhealthy weight loss behaviors. For example, 22% of youth American football players given bodyweight targets in order to be selected to compete were deemed to be at risk of abnormal eating behaviors, with eating binges, excessive exercise and drastic weight loss methods such as sauna use reported (⁶¹). Such body-mass manipulation can lead to dehydration that may affect performance in subsequent training activities (⁶), and similar thermal weight loss techniques have been described as “concerning” in combat sports (⁵). Despite good intentions, such emphasis on metrics has the potential to adversely affect the quality of output and can distract from, or even replace the original purpose (²³). Therefore, performance staff should use continuous reflection and intention to ensure any targets introduced serve their central goal of enhancing sports performance.

DIFFERENT STAKEHOLDERS OFTEN HAVE COMPETING INTERESTS

KPIs can be different for different stakeholders. In business, organizations have multiple stakeholders reflecting multiple functions (⁵²). Similarly in sport, different stakeholders can be seen as competitors with different motivations and means to achieve them (²²). For example, physical preparation staff may seek to set high training load targets to maximize performance, whereas medical staff seek to control training load to minimize risk of injury (²²). However, agency problems and conflict may arise if performance staff select these KPIs in relative isolation, without considering the interests of all involved stakeholders, even if well-intentioned. For example, an injured athlete may be concerned about their place in the team and an upcoming contract extension, and thereby wanting to play as soon as possible, whereas the medical staff are predominantly concerned about the rehabilitation and subsequent re-injury risk.

Nevertheless, the team sport staff (e.g., coaches) would like the player to compete in an imminent important game, because they are concerned about the team's winning percentage. When the above stakeholders' KPIs do not align, it may lead to tension within the sports organization. To resolve these kinds of agency problems, performance staff and team sport staff may adopt a shared decision-making model when making decisions regarding rehabilitation (⁶²). Expansion of this model, however, is needed to better support them in how to go about selecting their KPIs or performance-focused measures.

Indeed, the rehabilitation setting may warrant greater attention to KPIs, given the more controlled nature of returning an athlete from injury, and potential tension between physical performance and medical staff during such an interdisciplinary process (²²). KPIs may cover medical (e.g., palpation pain and range of motion), physical (e.g., high-speed running and accelerations), and technical aspects (e.g., passing and tackling) (⁵⁶). Time and/or clinical markers for return to play may also be seen as a KPI (⁹). Although these markers may not necessarily be named KPIs per se, they serve similar purposes in evaluating the player's performance and gauging their progression in RTP.

When performance is placed as the primary goal, competition between key stakeholders should disappear. Performance-focused measures determined in an interdisciplinary manner can bring stakeholders together as teammates (²²). For instance, greater preseason participation has been associated with lower in-season injury risk (⁵⁸); therefore, the so-called training load-injury paradox is actually in the best interest of physical preparation and the medical staff. Targets for injury availability may also be used, given the relationship between injuries and chance of success demonstrated across a variety of sports (¹²). It remains clear, however, that differing responsibilities within the performance staff, in addition to the wider team sport staff, can threaten agency issues when incorporating KPIs in applied practice.

A NEW FRAMEWORK FOR PERFORMANCE INDICATORS: ONGOING PROCESS TRACKERS

One of the main reasons KPIs are used so broadly is that they provide many well-established benefits in supporting individuals, teams, and organizations in monitoring progress toward desired performance goals. However, given some of the issues highlighted above, complementing KPIs with other tools may offer a more complete approach to evaluating performance progress over the long-term. Here, we propose a new approach to KPIs that may offer a solution to any unintended consequences associated with traditional KPIs and a complementary means to help more holistically track and support the progress of athletes or teams toward performance goals.

Ongoing process trackers are quantifiable measures of the execution of behaviors and processes that create the environments, cultures, and conditions for successful performance outcomes. Where traditional KPIs are often based on metrics that directly serve outcome goals, OPTs focus on measuring how well an athlete or team are adhering to the pathway toward achieving specific outcomes. The rationale for the development of OPTs thus stems heavily from literature describing the developmental pathway to peak performance of top athletes, and literature related to the interplay between process and outcome for supporting successful sporting performance. Specifically, one of the key factors distinguishing “super-elite” athletes (Olympic and/or World Champions) from “elite” athletes (international representatives but non-medalists) in a multidisciplinary developmental biography study (²⁷) was the joint focus of super-elites on mastery and outcome goals, whereas elite athletes demonstrated a strong outcome focus only. In addition, although super-elite athletes reported similar training volumes and training characteristics in adulthood, their continued performance improvement compared with elite athletes was attributed in significant part to the positive psychosocial characteristics of their training environment related to their mastery orientation (²⁶). Furthermore, in team sports, a perceived mastery climate created by coaches and performance staff is a stronger predictor of player performance than individual psychological factors such as mental toughness and grit (⁴⁵).

Indeed, focusing on the process over the outcome has long been a strategy favored by top coaches to help athletes manage the challenges and stressors associated with performance under pressure (¹⁴). Equally, a recent meta-analysis concluded that using process goals had a much larger positive effect on athlete performance (d = 1.36) compared with performance goals (d = 0.44). Process goals can be effective when set in isolation (⁵⁷). That said, other research has shown that performance is best served by focusing on process goals for areas of learning and development, with a switch to outcome focuses for areas of mastery (⁶³). Practically speaking, we suggest that practitioners may wish to integrate OPTs and KPIs into their measurement practice, potentially with individualized approaches for different athletes.

Borrowing from the business world, many top companies such as Google, Amazon, and Uber have successfully used a goal system called objectives and key results (OKRs) to help set goals and track progress toward them (¹⁷). Where traditional KPIs are often fixed and act to monitor the day-to-day health of a system (e.g., 80% squad availability), OKRs are agile and designed to also hold teams to bold and ambitious long-term goals (e.g., become the top injury management program in the league). Thus, we have adapted several elements key to the success of OKRs in this framework, in the hope that this will help practitioners effectively implement OPTs within their program. Table 1 presents the characteristics of OPTs with corresponding actions for practitioners, along with applied examples.

By helping to focus athletes, departments, and/or entire teams and organizations on the process over the outcome, complementing KPIs with OPTs may offer several potential advantages over traditional KPIs alone. First, by their design, they aim to reinforce behaviors that contribute to performance success, helping to avoid the negative consequences of Goodhart's law—these positive behaviors function as measures and targets that are helpful for performance (²⁴). Conversely, they can include “antigoals,” behaviors or approaches that should be avoided to meet the OPTs. This feature would ensure that performance is never inadvertently compromised to achieve targets.

Second, given the complexity of human performance, even our best predictive models populated by reams of data still struggle to help us understand how the multiple variables traditionally assigned as KPIs combine to contribute to team sport performance, particularly in open-skill sports such as soccer, basketball, and Australian rules football (³⁵). Instead, OPTs would help acknowledge these unknowns and would be intended to help tip an athlete's or team's odds in favor of success. Third, they offer the capacity to harness the experience, intuition, and context of practitioner and coaching teams in defining the processes for success in ways that traditional KPIs based solely on data-driven targets often neglect (²⁴). Even in a data-abundant sports performance environment, it warrants remembering that not everything that can be measured matters, and not everything that matters can be measured (⁴⁹). In addition, OPTs may also help drive talent development in a team, program, or organization by creating a culture that encourages younger or less-experienced athletes to model the positive behaviors of more experienced athletes, and helping to incentivize the value of long-term athlete development, and short-term performance outcomes (²). Finally, by encouraging athletes to focus on controllable factors, OPTs may offer the ancillary benefit of fostering athlete autonomy and intrinsic motivation (⁵⁰), helping prevent burnout and supporting athlete well-being while facilitating long-term performance success.

Multiple strategies may support the effective implementation of OPTs in sports performance environments. Given the multitude of data now collected, combined with the nonlinear relationship to performance, using OPTs may also help practitioners demonstrate the buy-in and success of a program with some distance from on-field outcomes. Figure 1 displays a checklist with key questions performance staff practitioners can ask themselves and their team when starting to select and use OPTs in their program.

CONCLUSION

Despite widespread adoption, using KPIs can be problematic given the complex and nonlinear nature of team sports performance. This concern is compounded by the growing data abundance and multidisciplinary practitioners within the performance staff, which can cause a drift from on-field performance goals and agency issues across different stakeholders. Given these issues, we propose a complementary framework, OPTs, that shifts attention to the processes that underpin performance. These are quantifiable measures of the behaviors and processes that create the environment for success, underpinned by mastery goals rather than a sole focus on outcome measures. Using a basket of OPTs allows different key stakeholders to engage and align to support the behaviors that can lead to long-term goals of an organization. Practitioners can use the OPT checklist (Figure 1) to track and demonstrate how their program supports process-driven goals.