Measuring Airline Performance: An Integrated Balanced Scorecard-Based MEREC-CoCoSo Model

Abstract

The assessment of company performance requires a holistic approach, encompassing both financial and non-financial metrics. Accordingly, we develop a comprehensive airline performance evaluation model utilizing the Balanced Scorecard (BSC)-based multi-criteria decision-making (MCDM) framework. Based on contingency theory, we use 30 Key Performance Indicators (KPIs) derived from the literature and develop a novel performance model by combining the BSC framework with the Method based on the Removal Effects of Criteria (MEREC) for KPI weighting and the Combined Compromise Solution (CoCoSo) for ranking. The focus on Turkish Airlines, serving as a comparative benchmark, over the period 2020–2023 reveals that while financial KPIs hold the greatest weight, non-financial KPIs have the most significant impact on performance. The lowest performance is recorded in 2020, most probably attributable to the COVID-19 pandemic, followed by a remarkable recovery in 2021. We offer a methodological contribution for managers, decision-makers, and scholars—an objective, data-driven tool to assess airline performance. Furthermore, we furnish policymakers with tangible data for more effective industrial incentives and convenient regulatory strategies. In contrast to most of the literature emphasizing financial indicators and subjective weighting approaches that might yield biased rankings, we suggest a novel integrated performance evaluation model tailored for the airline industry.

1. Introduction

The airline industry plays a pivotal role in global economies as it helps to create employment, facilitates trade, and enables tourism [1]. As derived from Oxford Economics estimations by [1], this industry supports nearly 4% of the global Gross Domestic Product (GDP) by supporting more than 86 million jobs, 11.6 million of which are direct employment, in 2023. However, these statistics are different for different geographies or country classifications. For instance, in 2023, this industry supports nearly 4.7% of all employment and 4.7% of GDP in North America, while these figures are, respectively, 1.5% and 2.6% in Africa [1]. We observe similar sharp differences between the least-developed economies, developing economies, and OECD countries. From that perspective, it can be inferred that this industry is crucial for economic growth and development [2], and it is not surprising that it is considered to be one of the most strategic industries for nations [3]. Given its substantial resource consumption and contribution to environmental pollution, the airline industry is a key factor in achieving the global Sustainable Development Goals [4,5]. From this perspective, the efficient utilization of financial, human, and natural resources, as well as their interaction with each other, should be addressed for this industry [6].

Although the airline industry is the focus of many jurisdictions, the development of this industry is highly dependent on infrastructure and aircraft investments [7]. Despite the huge number of investments, the profitability of the industry remains low [8]. According to the industry statistics provided by the International Air Transport Association [IATA] [9], the net profit margin of commercial airlines was 3.1% in 2019, −35.80% in 2020, −7.9% in 2021, −0.5% in 2022, and 3.9% in 2023. These data show that the pandemic affected profitability for a long time and this margin turned positive three years after the pandemic. Moreover, the Earnings Before Interest, Taxes, Depreciation, Amortization, and Rent (EBITDAR) margin was reported as the highest in 2019 over these five years [9]. On the one hand, the global airline industry has become extremely competitive [10], which means that the price is almost a given. Hence, cost competitiveness is the most apparent option to keep profitability in this industry. On the other hand, input prices, which are one of the most significant ingredients of cost competitiveness, are determined by global markets. Overall, long-term profitability and sustainability are critical, which highlights the significance of efficiency.

The concept of efficiency varies from utilizing the existing capacity to effectively employing resources. Therefore, identifying and improving potential ways to boost efficiency is essential to maintain performance. As performance measures, not only financial statement-based indicators but also special non-financial metrics that are specifically designed for this industry should be considered [11,12,13], since the former may not be sufficient to observe the performance as a whole. Although the latter may include several indicators ranging from service quality [14,15] and sustainability-related factors [16,17,18] to industry-specific factors [8,19,20], most studies generally consider a very limited number of indicators [21,22].

As underlined by [8], even basic indicators such as the number of passengers are not available in annual reports; thus, the literature must use the available indicators (such as the load factor, available seat km, and revenue passenger km). This data availability problem, however, induces the likelihood of a spurious relationship which may yield incorrect inferences in research findings [22]. Therefore, a more complete and accurate set of value-generation indicators should be considered [23]. A systematic approach is followed to determine the airline business models and two frameworks are developed by the literature [24]. However, as these models aim to understand business models, they may miss several valuable financial and non-financial performance indicators. For instance, although the model developed by [25] catches the labor index using several indicators including passenger per employee, personnel cost per Available Seat Kilometer (ASK), and ASKs per employee, it does miss several significant indicators including ethical scandals and the ratio of female employees in income-generating positions. We again underline that such models are developed to understand business models and it is worth emphasizing them as they pursue a systematic approach.

All in all, to our knowledge, the literature has a limited understanding of the indicators boosting the efficiency of the industry. Given the fact that this industry is considered to produce a social and public good [26], and that the aviation value chain suffered the biggest economic losses during the COVID-19 pandemic with economic losses of USD 175 billion in 2020 and USD 104 billion in 2021 [27], the performance management of this industry is unsurprisingly attracting not only academic and professional but also public interest. Therefore, with the motivation of supporting several parties from airline managers to policymakers and academicians with further insight into the indicators affecting airline company performance, we aim to contribute to the literature by providing a comprehensive picture of the topic. Hence, this raises the following questions:

RQ1: 

What model can be developed and implemented to systematically evaluate both the financial and non-financial performance of the airline industry?

RQ2: 

Which KPIs capture the overall performance adequately?

For this purpose, first, we derive both financial and non-financial KPIs from the literature related to the four pillars of BSC, named financial, customer, internal processes, and learning and growth. We underline that knowing the exact KPIs of each airline company is not possible [8]. Hence, we classify each KPI under its corresponding BSC pillar by undertaking a comprehensive literature review. After finalizing the KPIs, we check the annual reports, the sustainability reports, and the financial reports of several global airline companies to see whether our comprehensive set of KPIs is fully available over 2020 and 2023. Note that, as [8] highlights, many KPIs are not available in the annual reports of most airlines. In the end, we recognize that Turkish Airlines (THY) provides a convenient research setting for our objective as it provides a very wide range of KPIs. Note that this period of analysis contains the adverse outcomes (amid lockdowns and travel restrictions) of the pandemic and the economic expansion just after the pandemic.

To our knowledge, the literature analyzes the topic by employing several techniques including parametric and non-parametric methods: the former (latter) majorly includes regression analyses (Data Envelopment Analysis) [28]. In our research, we determine the overall performance by employing the recently developed MEREC-CoCoSo model which offers distinct benefits in multiple ways. First, both financial and non-financial KPIs constitute the criteria of the decision matrix, while the company performance in the years 2020–2023 constitutes the alternatives. The use of MEREC enables a systematic and objective weight assignment of the criteria which reflects how the significance of each criterion has evolved. Specifically, real-time data-driven analysis, rather than expert opinion-based techniques, captures the influence exerted by factors such as macroeconomic fluctuations, changes in operational conditions, and trends in customer preferences, etc., on performance. Upon determining the criteria weights through MEREC, ranking through the CoCoSo method provides a robust method for selecting the optimal performing alternative. Notably, this integrated model presents an objective methodology grounded in empirical data, facilitating comparative performance analysis across various periods, including both crisis and recovery phases. By leveraging this approach, our model contributes to the identification of the KPIs that are most adversely affected by shocks, such as the outbreak of the pandemic or hyperinflationary periods. Understanding which criteria exhibit heightened sensitivity, particularly during such disruptions, is critical for top management’s decision-making processes, as it might enable the formulation of accelerated recovery strategies. Furthermore, the proposed model extends beyond short-term performance evaluation, allowing for the comprehensive exploration of long-term performance trends. This methodological novelty is our further contribution to the literature. For policymakers, as our model provides benchmarking across airline companies, it can serve as a useful tool to design incentive schemes or monitor performance for sectoral development.

The rest of this paper is structured as follows. The second section begins by laying out the theoretical discussion related to the performance measurement techniques, focusing on the KPIs of the airline industry, and the decision criteria that constitute the basis of our research are retrieved from prior studies. The following section outlines the methodology in which our BSC-based MEREC CoCoSo model is introduced, together with the sample selection and data selection processes. The results and discussion are presented in the fourth section. We then present the concluding remarks and give the limitations and suggestions for further research in the last section.

2. Theoretical Discussion

Performance measurement represents a structured process designed to evaluate a company’s efficacy in utilizing resources to produce outputs, which are vital for internal and external stakeholders. The authors of [29] portrayed organizations as flexible and even unpredictable systems whose functioning depends on managerial preferences and decisions.

Beginning in the second half of the 20th century, diverse performance measurement systems have been developed to accurately reflect the dynamic environment of a company which has been directly or indirectly influenced by internal and external factors [30]. Several theories have been and still are applied to performance measurement. Defined as “a theoretical perspective of organizational behavior that emphasizes how contingent factors such as technology and the task environment affected the design and functioning of organizations” (p. 4) [31], contingency theory lies at the heart of the performance measurement literature, and it is based on the premise that organizations are open systems that interact with the environment in which they operate and therefore must adapt to changing market conditions [32].

Indeed, many works have attempted to explore contingencies affecting company performance, including heterogeneity in country systems [33], organizational structures [34], leadership characteristics [35], management control systems [36], and technological shifts [32], and the airline industry is no exception. In fact, the airline industry is the most susceptible industry impacted by contingencies in which any economic, social, environmental, or political disruption can rapidly spread from one country to the entire world [37]. Despite this industry’s prominence and strategic importance, there remains a paucity of work addressing this field of research. For instance, the authors of [38] examine the impact of financial distress on airfares in the United States (U.S.) airline industry under contingencies such as operating expenses, firm size, and market share. The authors of [39] conclude that operational performance, capacity utilization, and profitability are susceptible to the operating model—focused or full-service—for major carriers in the U.S.

In recent years, contemporary performance measurement methods have emerged in response to the constraints of longstanding traditional evaluation methods that rely predominantly on financial metrics [40]. Developed by [41,42], the Balanced Scorecard (BSC) offers a modern multidimensional framework based on both financial and non-financial indicators that enable the assessment of company performance and strategic management [43]. The fundamental aspects of BSC include (i) financial perspective, which encompasses enhancing revenue generation while maintaining cost control and driving value maximization, (ii) customer perspective, which is centered on customer satisfaction, loyalty, and market share, (iii) internal perspective, which assesses the efficiency of operational processes and managerial practices within the organization, and (iv) learning and growth perspective, which focuses on the retention of qualified personnel and the promotion of organizational development initiatives [42]. The causal path of the KPIs initiates with organizational learning and growth, leading to the extent of the efficacy of internal processes, subsequently shaping the quality of customer services, and ultimately culminating in financial performance. In essence, the financial aspect (lagging factor) reflects the historical performance, while the remaining non-financial aspects (leading factors) are determinants of the future performance [42].

The BSC has received considerable interest in the performance assessment of airline companies. For instance, Ref. [44] examines the role of digitalization on the performance of Emirates Airlines before and during the COVID-19 pandemic by using the BSC model, which yielded superior financial performance over non-financial performance in the pre-COVID-19 period. Conducting a comprehensive study based on airline companies, Ref. [45] reveals the validity of the use of BSC through confirmatory factor analysis.

Hence, a significant body of research (see, among others, Refs. [46,47,48]) asserts that performance evaluations based on the BSC framework are fundamentally influenced by contingency theory, given that its interactive structure enables a thorough approach to performance assessment. Therefore, the selection of KPIs and the structural design of our performance measurement model grounded in contingency theory incorporate the inherent characteristics of the airline sector, where performance outcomes are continuously shaped by exogenous factors such as regulatory changes and geopolitical and socioeconomic conditions. In other words, contingency theory shed light on our KPI selection and structural design.

As discussed in the first section, knowing the exact KPIs of each airline company is not possible [8]. Therefore, we retrieve the relevant KPIs meticulously tailored for the performance measurement of the airline industry based on prior studies, as detailed in

Table 1. Decision criteria and definitions.

BSC Dimension	Criteria (Ci)	KPIs	Type	Description and Formula	Reference
Financial	C1	Share of Passenger Revenue	max	${\displaystyle \frac{\mathrm{Passenger} \mathrm{Revenue}}{\mathrm{Total} \mathrm{Revenue}}}$	[10]
	C2	Share of Cargo Revenue	max	${\displaystyle \frac{\mathrm{Cargo} \mathrm{Revenue}}{\mathrm{Total} \mathrm{Revenue}}}$	[10]
	C3	Operational Costs Ratio	min	${\displaystyle \frac{\mathrm{Operational} \mathrm{Costs} }{\mathrm{Total} \mathrm{Revenue}}}$	[38]
	C4	Cash Flow Ratio	max	${\displaystyle \frac{\mathrm{Cash} \mathrm{Flow} }{\mathrm{Total} \mathrm{Revenue}}}$	[49,50]
	C5	Net Profit Margin	max	${\displaystyle \frac{\mathrm{Net} \mathrm{Income}}{\mathrm{Total} \mathrm{Revenue}}}$	[51]
	C6	EBITDAR Margin	max	${\displaystyle \frac{\mathrm{EBITDAR}}{\mathrm{Total} \mathrm{Revenue}}}$	[52]
	C7	ROA	max	${\displaystyle \frac{\mathrm{Net} \mathrm{Income}}{\mathrm{Total} \mathrm{Assets}}}$	[40,49,51]
	C8	CASK	min	${\displaystyle \frac{\mathrm{Total} \mathrm{Operating} \mathrm{Costs}}{\mathrm{ASK}}}$	[53]
	C9	Non-Fuel CASK	min	$\mathrm{CASK}-{\displaystyle \frac{\mathrm{Fuel} \mathrm{Costs}}{\mathrm{ASK}}}$	[53]
	C10	Passenger Revenue per ASK	max	${\displaystyle \frac{\mathrm{Passenger} \mathrm{Revenue}}{\mathrm{ASK}}}$	[53,54]
	C11	Revenue Yield	max	${\displaystyle \frac{\mathrm{Total} \mathrm{Revenue}}{\mathrm{RPK}}}$	[55,56]
	C12	Aircraft Ownership Cost per Block Hour	min	${\displaystyle \frac{\mathrm{Total} \mathrm{Aircraft} \mathrm{Own}. \mathrm{Costs}}{\mathrm{Total} \mathrm{Block} \mathrm{Hours}}}$	[56]
	C13	Maintenance Cost per Block Hour	min	${\displaystyle \frac{\mathrm{Total} \mathrm{Maintenance} \mathrm{Costs}}{\mathrm{Total} \mathrm{Block} \mathrm{Hours}}}$	[57,58]
	C14	Handling Cost per Landing	min	${\displaystyle \frac{\mathrm{Total} \mathrm{Handling} \mathrm{Costs}}{\# \mathrm{of} \mathrm{Landings}}}$	[57,58]
	C15	PCSE Ratio	min	${\displaystyle \frac{\mathrm{Total} \mathrm{PCSE}}{\# \mathrm{of} \mathrm{Passengers}}}$	[59]
Customer	C16	Number of Landings	max	Annual # of Aircraft Landings	[60]
	C17	Passengers Carried	max	Annual # of Passengers Carried	[38]
	C18	Average Response Days to Customer Complaints	min	${\displaystyle \frac{\mathrm{Total} \mathrm{Response} \mathrm{Time} \mathrm{to} \mathrm{Complaints}}{\# \mathrm{of} \mathrm{Complaints}}}$	[59]
	C19	Passenger Load Factor	max	${\displaystyle \frac{\mathrm{RPK}}{\mathrm{ASK}}}$	[38]
	C20	Passenger Satisfaction Rate	max	${\displaystyle \frac{\# \mathrm{of} \mathrm{Satisfied} \mathrm{Passengers}}{\# \mathrm{of} \mathrm{Surveyed} \mathrm{Passengers}}}$	[59]
Internal	C21	Average Fleet Age	min	${\displaystyle \frac{\mathrm{Total} \mathrm{Age} \mathrm{of} \mathrm{Aircraft} \mathrm{in} \mathrm{Fleet} }{\# \mathrm{of} \mathrm{Aircraft}}}$	[61]
	C22	Available Seat Kilometers (ASK)	max	Total Seats Available × km Flown	[54]
	C23	Number of Aircraft	max	Annual # of Active Fleets	[57]
Learning and Growth	C24	Total Workforce	max	# of Total Employees	[10]
	C25	Women in Workforce Ratio	max	${\displaystyle \frac{\mathrm{Male} \mathrm{Employees}}{\mathrm{Female} \mathrm{Employees}}}$	[62]
	C26	Employee Retention Rate	max	(1 − Employee Turnover Ratio)	[63]
	C27	Ratio of Female-to-Male in MLM	max	${\displaystyle \frac{\# \mathrm{of} \mathrm{Females} \mathrm{in} \mathrm{MLM}}{\# \mathrm{of} \mathrm{Males} \mathrm{in} \mathrm{MLM}}}$	[64]
	C28	Discrimination Cases	min	$\#$ of Discrimination Complaints	[62]
	C29	Ethical Scandals	min	$\#$ of Ethical Line Complaints	[65]
	C30	Share of Female Employees in IGP	max	${\displaystyle \frac{\# \mathrm{of} \mathrm{Females} \mathrm{in} \mathrm{IGP}}{\mathrm{Total} \# \mathrm{of} \mathrm{Employees} \mathrm{in} \mathrm{IGP}}}$	[66]

Note: The table above represents our decision criteria and relevant information. All abbreviations are illustrated with parentheses just after them as follows: EBITDAR (Earnings Before Interest, Taxes, Depreciation, Amortization, and Rent), ROA (Return on Assets), CASK (Cost of Available Seat Kilometer), PCSE (Passenger Catering and Service Expenses), ASK (Available Seat Kilometer), RPK (Revenue Passenger Kilometers), MLM (Mid-Level Management), IGP (Income Generating Positions). # stands for the word number.

. To recap, (i) the financial dimension consists of the first fifteen KPIs, which represent cost, revenue, and earnings ratios; (ii) the customer dimension comprises five criteria (16–20), covering a range of metrics varying from passenger indicators to flight operations; and (iii) while there are three indicators (21–23) related to the core activities of airline companies in the internal dimension, the learning and growth dimension (24–30) is assessed through seven criteria related to workforce characteristics.

3. Materials and Methods

3.1. Research Model

Despite its significance in performance measurement, the BSC is criticized due to the linear cause-and-effect relationship assumption among leading indicators and lagging indicators [67] that eventually influence the company’s overall success [68]. A survey conducted by [69] with a sample of 104 airline companies shows the ambiguity of importance among the BSC pillars which causes a failure in performance assessment. The work in [70] highlights the complications arising due to the interconnectedness among the KPIs utilized in four aspects of the BSC.

Given that MCDM ensures robust models for decision-making, given that it accounts for not only quantitative but also qualitative data which reflects the shifts in industries [71], these techniques have been merged with the BSC approach for performance evaluation in multiple industries, i.e., banking [72,73,74], hospitality [75], oil production [76] and glass manufacturing [77].

However, despite the complexity of the inherent nature and the contingency of externalities of the airline industry, few scholars have put effort into incorporating MCDM methods into the use of the BSC in the literature [40,78,79] to alleviate the so-called tradeoff concerns embedded in the framework [80,81,82]. For instance, by sampling three Taiwanese airports, the authors of [43] proposed a model based on the use of the BSC together with Decision-Making Trial and Evaluation, a laboratory-based Analytical Network Process (DEMATEL—DANP), and VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) to investigate the key influences of airport performance. Another study conducted by [40] applied the BSC to the Fuzzy DEMATEL, Fuzzy ANP, and Multi-objective Optimization By Ratio Analysis (MOORA) techniques within the context of the European airline industry and found the priority of the customer dimension.

Drawing upon the prior literature, we developed the integrated BSC-based MEREC CoCoSo model which comprises three main stages, as depicted in Figure 1. In the first stage, following the determination of the research objective, criteria and alternatives are defined in line with the BSC dimensions. This stage is critical in terms of structuring the decision matrix. In the second stage, the MEREC method is used for weighing the criteria (KPIs) derived from the pillars of the BSC. For this purpose, this step includes normalizing the values in the matrix (N), calculating the overall performances of the alternatives (years) (S_i), analyzing the variations in the alternative performance by removing each criterion (S_ij’), determining the removal effects of each criterion, and ultimately obtaining the criterion weights. In the third stage, the CoCoSo ranking method is employed based on the following steps. Initially, the criteria values are normalized based on the initial decision matrix, the aggregated scores of the alternatives are calculated (S_i and P_i), and the relative evaluation scores of the alternatives are computed, culminating in the overall utility score (k_i), which informs the final decision. As indicated by [83,84], the logarithmic transformation of large values is not required since initial normalization processes embedded in MCDM methods inherently prevent disproportionate distributions among disparities. The use of the BSC together with the MEREC and CoCoSo methods is comprehensively elaborated in the next section.

3.1.1. MEREC Criteria-Weighting Method

MCDM studies heavily depend on criteria weights as these directly affect the research outcomes [85]. Criteria-weighting methods are classified as subjective, objective, or a combination of both. Subjective weighting methods rely on the preferences of experts within the relevant field. However, expert opinion-based weighting methods often struggle to assign appropriate weights to the criteria for several reasons. First, concerns may arise regarding the sufficiency of experts’ knowledge regarding the accuracy of weights [86]. Furthermore, as the number of criteria increases, the complexity of preference assessments also escalates [87], leading to time-consuming workloads [88]. In contrast, objective methods provide a statistical and systematical approach to assigning values based on the decision matrix, thereby minimizing human error [89]. These methods yield reproducible results and are commonly favored in cases where obtaining expert opinions is challenging or even impossible [86].

Among various objective methods, MEREC, developed by [89], analyzes the effect of each criterion on the overall performance of the alternatives. Specifically, it relies on the evaluation of equally weighted logarithmic metrics and the variations in the performance across alternatives when a criterion is removed from the analysis [90]. If the exclusion of a criterion from the decision matrix significantly affects the overall performance of the alternatives, this criterion is evaluated with a higher weight. In other words, criteria with higher variability are assigned higher weights [91]. Since MEREC evaluates the impact of removing each criterion from the decision matrix, rather than presuming their isolated effects, this method captures interdependencies among criteria by assessing the impact of excluding each criterion on the outcome and provides a more nuanced assessment that is often overlooked by other weighting methods [92]. For instance, Ref. [93] found evidence that the results provided by MEREC are more efficient when compared to other objective methods such as Entropy and CRITIC. This also removes the multicollinearity problem among KPIs [79].

Integrated with the BSC framework, as discussed above in detail, MEREC is deemed to be the most relevant and suitable criteria-weighting method for evaluating airline performance encompassing both financial and non-financial KPIs since this method eliminates bias and personal judgments [85] especially when considering the airline industry which is sensitive to dynamic macroeconomic determinants [94]. This permits the comprehension of the influence of individual contingency on performance, which eventually facilitates achieving organizational goals through the optimization of strategic decision-making [95].

In addition, the removal of each criterion to methodically evaluate the overall performance illuminates deviations in performances across different years, contributing a valuable comparative understanding especially during economic extremities.

The calculation of weights of each criterion based on MEREC is conducted as follows.

Step 1. Constructing the decision matrix. As defined in Equation (1), the decision matrix X, which serves to evaluate the alternatives in the performance evaluation, consists of m alternatives and n criteria, where x_ij denotes the elements of the matrix X, representing the performance of the alternative I under criterion j.

$$\mathrm{X}=\left[\begin{array}{ccc}{\mathrm{x}}_{11}& \dots & {\mathrm{x}}_{1\mathrm{m}}\\ ⋮& \ddots & ⋮\\ {\mathrm{x}}_{\mathrm{n}1}& \dots & {\mathrm{x}}_{\mathrm{nm}}\end{array}\right]$$

(1)

Step 2. Normalizing the decision matrix (N). As specified in Equation (2), a simple linear normalization is used to scale all values of x_ij in the decision matrix and the elements of the normalized matrix are denoted by n^x_ij where J⁺ denotes beneficial criteria, while J⁻ denotes non-beneficial criteria.

$${\mathrm{n}}_{\mathrm{ij}}^{\mathrm{x}}=\left\{\begin{array}{c}{\displaystyle \frac{{\mathrm{minx}}_{\mathrm{kj}}}{{\mathrm{x}}_{\mathrm{ij}}}} \mathrm{i}\mathrm{f} \mathrm{j}\in {\mathrm{J}}^{+}\\ {\displaystyle \frac{{\mathrm{x}}_{\mathrm{ij}}}{{\mathrm{maxx}}_{\mathrm{kj}}}} \mathrm{i}\mathrm{f} \mathrm{j} \in {\mathrm{J}}^{-}\end{array}\right.$$

(2)

Step 3. Calculating the overall performance of the alternatives (S_i). A non-linear logarithmic function is applied, as shown in Equation (3) below to compute overall performances. This step ensures that lower values obtained from the normalization process result in greater performance values.

$$S_i=\mathit{ln}\left(1+\left({\displaystyle \frac{1}{\mathit{m}}}\sum _{\mathit{j}}\left|\mathit{ln}\left({\mathit{n}}_{\mathit{i}\mathit{j}}^{\mathit{x}}\right)\right|\right)\right)$$

(3)

Step 4. Calculating the performance values of alternatives by removing each criterion ${(\mathrm{S}}_{\mathrm{i}\mathrm{j}}^{\prime })$. This step calculates the variations of the alternatives by removing each criterion separately. This process aims to obtain ${\mathrm{S}}_{\mathrm{i}\mathrm{j}}^{\prime }$ which symbolizes the overall performance of the ith alternative when the jth criterion is removed. Equation (4) describes the computation.

$${\mathrm{S}}_{\mathrm{ij}}^{\prime }=\ln \left(1+\left({\displaystyle \frac{1}{\mathrm{n}}}\sum _{\mathrm{k},\mathrm{k}\ne \mathrm{j}}\left|\ln ({\mathrm{n}}_{\mathrm{ij}})\right|\right)\right)$$

(4)

Step 5. Calculating the removal effect of the j th criterion. Based on previous steps, the removal effect of the j^th criterion, E_j, is obtained by using Equation (5).

$${\mathrm{E}}_{\mathrm{j}}=\sum _{\mathrm{i}}\left|{\mathrm{S}}_{\mathrm{ij}}^{\prime }-{\mathrm{S}}_{\mathrm{i}}\right|$$

(5)

Step 6. Determining the criteria weights. The objective weights (w_j) of each criterion are calculated based on the removal effects E_j of the previous step. Equation (6) shows the computation of each weight (w_j) associated with each criterion j^th.

$${\mathrm{w}}_{\mathrm{j}}={\displaystyle \frac{{\mathrm{E}}_{\mathrm{j}}}{\sum _{\mathrm{k}}{\mathrm{E}}_{\mathrm{k}}}}$$

(6)

3.1.2. CoCoSo Ranking Method

Introduced by [96], CoCoSo is an MCDM ranking method that incorporates simple additive weighting and exponentially weighted product models. This method is differentiated from its counterparts due to its simplified complexity which enhances its understandability and applicability across multiple research domains [97]. CoCoSo is attracting significant interest in the literature on the evaluation of financial and non-financial performance thanks to its reliability in computing optimal compromise scores through an integrated framework [98]. For instance, Ref. [99] uses a hybrid integrated data-driven weighting system and CoCoSo approach to measure the financial performance of Fortune 500 companies. CoCoSo is generally preferred to be integrated with MEREC due to the latter’s outperformance of criteria weighting over its counterparts [98,99]. For instance, Ref [100] applies CoCoSo in airport service quality evaluation while Ref. [101] uses CoCoSo in choosing sustainable suppliers. Ref. [102] analyzes the financial performance of the insurance industry based on six criteria which are weighed by MEREC and Logarithmic Percentage Change-driven Objective Weighting (LOPCOW) and ranked by CoCoSo and Evaluation based on Distance from Average Solution (EDAS) methods. Ref. [90] combined an MEREC-CoCoSo model for assessing the changes in the performances of companies operating in the tourism sector due to the COVID-19 pandemic. Ref. [79] developed a MEREC-CoCoSo/Borda model tailored for evaluating the performance of the airline industry. Therefore, pipelining MEREC with CoCoSo offers a robust performance evaluation methodology to track deviations in performances under turbulent periods including the COVID-19 outbreak which should be considered to be a significant corner for the airline industry.

The steps of the CoCoSo method are as follows:

Step 1. Determining the Initial Decision Matrix. This matrix is constructed based on Equation (1).

Step 2. Normalizing the criteria values. Equation (7) shows the compromised normalization equations for beneficial and non-beneficial (cost) criteria.

$$\begin{gathered} {\mathrm{r}}_{\mathrm{ij}}={\displaystyle \frac{{\mathrm{x}}_{\mathrm{ij}}-{\mathrm{minx}}_{\mathrm{ij}}}{{\max }_{\mathrm{ij}}-{\mathrm{minx}}_{\mathrm{ij}}}} ; \mathrm{j}\in {\mathrm{J}}^{+} \\ {\mathrm{r}}_{\mathrm{ij}}={\displaystyle \frac{{\mathrm{maxx}}_{\mathrm{ij}}-{\mathrm{x}}_{\mathrm{ij}}}{{\mathrm{maxx}}_{\mathrm{ij}}-{\mathrm{minx}}_{\mathrm{ij}}}} ; \mathrm{j}\in {\mathrm{J}}^{-} \end{gathered}$$

(7)

Step 3. Calculating the aggregated scores Si and Pi. The sum of the weighted comparability sequence and the power weight of comparability sequences are computed, where S_i is derived based on the grey relational generation approach as shown in Equation (8), and P_i represents the multiplicative comparison values, as depicted in Equation (9).

$${\mathrm{S}}_{\mathrm{i}}=\sum _{\mathrm{j}=1}^{\mathrm{n}}\left({\mathrm{w}}_{\mathrm{j}}{\mathrm{r}}_{\mathrm{ij}}\right)$$

(8)

$${\mathrm{P}}_{\mathrm{i}}=\sum _{\mathrm{j}=1}^{\mathrm{n}}{\left({\mathrm{r}}_{\mathrm{ij}}\right)}^{{\mathrm{w}}_{\mathrm{j}}}$$

(9)

Step 4. Obtaining relative assessment scores of alternatives. Based on Equations (10)–(12), three appraisal scores are calculated.

$${\mathrm{K}}_{\mathrm{ia}}={\displaystyle \frac{{\mathrm{P}}_{\mathrm{i}}+{\mathrm{S}}_{\mathrm{i}}}{\sum _{\mathrm{i}=1}^{\mathrm{m}}({\mathrm{P}}_{\mathrm{i}}+{\mathrm{S}}_{\mathrm{i}})}}$$

(10)

$${\mathrm{k}}_{\mathrm{ib}}={\displaystyle \frac{{\mathrm{S}}_{\mathrm{i}}}{\min {\mathrm{S}}_{\mathrm{i}}}}+{\displaystyle \frac{{\mathrm{P}}_{\mathrm{i}}}{\min {\mathrm{P}}_{\mathrm{i}}}}$$

(11)

$${\mathrm{k}}_{\mathrm{ic}}={\displaystyle \frac{\mathsf{\lambda }\left({\mathrm{S}}_{\mathrm{i}})+(1-\mathsf{\lambda }\right)\left({\mathrm{P}}_{\mathrm{i}}\right)}{\left(\mathsf{\lambda }\max {\mathrm{S}}_{\mathrm{I}}+(1-\mathsf{\lambda }\right)\max {\mathrm{P}}_{\mathrm{i}})}}$$

(12)

where $k_{ia}$ represents the arithmetic mean of the S_i and P_i values in Equation (10), $k_{ib}$ denotes the relative scores compared to the best alternative in Equation (11), and $k_{ic}$ depicts the balanced compromise between S_i and P_i values. For equal balance, we take $\mathsf{\lambda }$ as 0.5 in Equation (12).

Step 5. Calculating the overall utility score ${(\mathrm{k}}_{\mathrm{i}}).$ The overall score is obtained by combining the values both by the arithmetic and the geometric mean of ${\mathrm{k}}_{\mathrm{i}\mathrm{a}}, {\mathrm{k}}_{\mathrm{i}\mathrm{b}} \mathrm{and}$ ${\mathrm{k}}_{\mathrm{i}\mathrm{c}}$ as shown in Equation (13). The problem is solved with the final ranking which is determined based on the descending order of ${\mathrm{k}}_{\mathrm{i}}$.

$${\mathrm{k}}_{\mathrm{i}}={{(\mathrm{k}}_{\mathrm{ia}}.{\mathrm{k}}_{\mathrm{ib}}.{\mathrm{k}}_{\mathrm{ic}})}^{{\displaystyle \frac{1}{3}}}+{\displaystyle \frac{1}{3}}({\mathrm{k}}_{\mathrm{ia}}.{\mathrm{k}}_{\mathrm{ib}}.{\mathrm{k}}_{\mathrm{ic}})$$

(13)

3.2. Sample and Data Collection

Studying the Kenyan airline industry, Ref. [95] argues the unavailability of a sole common optimal strategy for all airline companies since each of them is contingent on distinct and unique factors and, therefore, performance measurement should be conducted by taking the individual circumstances of each company into account. Single-company-based studies give valuable intrinsic insights regarding not only organizations’ strengths and weaknesses but also enhance the generalizability of a given model through comparisons with industry peers and the identification of best practices. Furthermore, they foster methodological consistency in subsequent studies by standardizing data collection and analytical processes [103].

To perform our analyses, the availability of KPIs is essential. As highlighted by [8], many KPIs are not fully available in annual reports. Since we do need a comprehensive set of KPIs, we identify that THY is the most suitable company with no missing KPIs. THY is certified by Skytrax, a leading global airline and airport rating agency, as the Best Airline in Europe for the ninth time, together with the Best Business Class Catering and the Best Airline in Southern Europe at the 2024 World Airline Awards, which are widely recognized as the Oscars of the Aviation Industry [104]. As of the 21st of February 2025, THY is the tenth largest airline with the largest market capitalization at the global level (for a detailed list, we refer the reader to the following link: https://companiesmarketcap.com/airlines/largest-airlines-by-market-cap/ (accessed on 24 February 2025)) and it is a member of the Star Alliance which is the first global airline alliance focusing on enhancing customer value globally [105] and has a network exceeding one thousand airports as of the end of 2024 [106]. With the help of such global collaborations and its potent growth vision, THY broke the Guinness Record for Most Countries Flown to by an Airline as of November 2024 [107]. Expectedly, the percentage of passengers carried on domestic flights in total was around 36% in 2022 and 2023 [108]. According to the Turkish Exporters Assembly’s data, this unique business environment makes THY the highest exporting company (including manufacturing giants) in Turkey in 2022 and 2023 (for detailed reports, we refer the reader to the following link: https://tim.org.tr/tr/turkiyenin-ilk-1000-ihracatci-arastirmasi (accessed on 24 February 2025)).

In our study, necessary data are manually collected from the annual reports, sustainability reports, and audited financial reports of THY which are publicly available on its website. The period of analysis begins in 2020, aiming to analyze the company’s resilience and performance against the effects of the COVID-19 outbreak, and extends through 2023, the most recent year for financial and non-financial reports. Although the functional currency is Turkish Lira in the reports, we rely on the data in US dollars reported by THY for comparability and standardization purposes on an international scale. The reason behind this is that Turkish Lira-reported measures are expected to be majorly influenced by the adverse impacts of the high level of inflation and the drastic depreciation of the Turkish Lira taking place after the pandemic. According to the Ref [109] Turkish Statistical Institute (n.d.) (the annual inflation rate was 14.60% in 2020, 36.08% in 2021, 64.27% in 2022, and 64.77% in 2023. Moreover, based on the monthly average USD and EUR exchange rates data of the Ref. [110] we first calculate the average annual USD and EUR exchange rate figures and construct a mixed exchange rate by taking half from USD and half from EUR for each year. Our calculation reveals that the Turkish Lira depreciated by 25% in 2020, 28.2% in 2021, 76.1% in 2022, and 45.5% in 2023. These nominal depreciation figures show the severity of depreciation which directly refers to the inconvenience of Turkish Lira-based reported indicators.

4. Discussion

In this section, we present our empirical findings obtained from the use of our proposed integrated BSC-MEREC-CoCoSo model. As the period of analysis is 2020–2023, we consider 2020 to be the crisis period for the industry due to the pandemic, and the following year to be the recovery period. We further compare and discuss the results.

4.1. MEREC Results

In this study, the observation of the thirty criteria over the four-year period of 2019–2023 yields a 4 × 30 decision matrix. The initial decision matrix consisted of both positive and negative values.

However, as the MEREC method requires only positive values due to the natural logarithm of values, negative x_ij values must be transformed into positive values by an appropriate method [89]. Following [111] and [112], we transform all negative elements in our decision matrix by means of the Z-score. The Z-score serves as a statistical measurement of the distance a data point is, in terms of standard deviations, from the mean. The Z-score is calculated as

$$Z={\displaystyle \frac{x-\mu }{\mathsf{\sigma }}}$$

where

• $x$ = individual data point;
• $\mu$ = mean of the population;
• $\mathsf{\sigma }$ = standard deviation of the population.

After converting negative values into positive by means of the Z-score, the final decision matrix is shown in

Table 2. Initial and final decision matrices.

		Panel A. Initial Decision Matrix				Panel B. Final Decision Matrix
Ci	Type	2020	2021	2022	2023	2020	2021	2022	2023
C1	max	0.5631	0.5980	0.7695	0.8465	0.5631	0.5980	0.7695	0.8465
C2	max	0.4042	0.3757	0.2027	0.1240	0.4042	0.3757	0.2027	0.1240
C3	min	0.1397	0.1052	0.0999	0.1209	0.1397	0.1052	0.0999	0.1209
C4	max	0.2689	0.2505	0.2573	−0.1611	3.3629	3.2617	3.2990	1.0000
C5	max	−0.1241	0.0897	0.1479	0.2875	1.0000	2.4437	2.8362	3.7785
C6	max	0.2199	0.3500	0.2920	0.2900	0.2199	0.3500	0.2920	0.2900
C7	max	−0.0327	0.0361	0.0881	0.1688	1.0000	1.0689	1.1208	1.2015
C8	min	0.0969	0.0731	0.079	0.0778	0.0969	0.0731	0.0790	0.0778
C9	min	0.075	0.0521	0.0458	0.0513	0.0750	0.0521	0.0458	0.0513
C10	max	0.0506	0.05	0.0708	0.0755	0.0506	0.0500	0.0708	0.0755
C11	max	0.0712	0.0737	0.0887	0.0914	0.0712	0.0737	0.0887	0.0914
C12	min	2.711	1.817	1.342	1.328	2.711	1.817	1.342	1.328
C13	min	778	537	574	576	778	537	574	576
C14	min	2.314	2.087	2.061	2.388	2.314	2.087	2.061	2.388
C15	min	7.75	6.05	8.6	10.4	7.75	6.05	8.6	10.4
C16	max	240,354	357,207	472,724	539,743	240,354	357,207	472,724	539,743
C17	max	28	44.8	71.82	83.38	28	44,8	71.818	83.378
C18	min	4.9	4.3	6.7	5.1	4.9	4.3	6.7	5.1
C19	max	0.71	0.679	0.806	0.826	0.71	0.679	0.806	0.826
C20	max	0.77	0.83	0.83	0.81	0.77	0.83	0.83	0.81
C21	min	8.4	8.5	8.7	9.3	8.4	8.5	8.7	9.3
C22	max	75	127.793	201.735	234.839	75	127.793	201.735	234.839
C23	max	363	370	394	440	363	370	394	440
C24	max	33,583	33,191	37,379	35,013	33,583	33,191	37,379	35,013
C25	max	0.4600	0.4173	0.4018	0.4373	0.46	0.4173	0.4018	0.4373
C26	max	0.954	0.947	0.97	0.95	0.954	0.947	0.968	0.95
C27	max	0.4653	0.4548	0.4664	0.4860	0.4653	0.4548	0.4664	0.4860
C28	min	2	4	1	14	2	4	1	14
C29	min	134	179	436	355	134	179	436	355
C30	max	0.07	0.06	0.07	0.047	0.07	0.06	0.07	0.047

Note: Criteria C16 and C22 are expressed in billions.

.

Figure 2 depicts the distribution of weights derived from the MEREC method among the financial, customer, internal, and learning and growth dimensions. The findings reveal that financial indicators (45.42%) represent—in total—the largest portion in evaluating airline performance, indicating that airline companies are primarily assessed based on factors including but not limited to cost-management and revenue-generation KPIs.

Figure 2 further shows that, following the financial dimension, the learning and growth dimension (25.07%) holds the second highest weight, underscoring the significant role of human resources within the industry. While the financial dimension constitutes—as a whole—the largest portion of the weight distribution, the learning and growth and customer KPIs rank higher on an individual basis, as illustrated in

Table 3. Criteria weights obtained from MEREC.

Rank	Criteria	Ci	wi	BSC Dimension
1	Employee Retention Rate	C26	4.1202	Learning and Growth
2	Ratio of Female-to-Male in MLM	C27	4.0570	Learning and Growth
3	Total Workforce	C24	4.0034	Learning and Growth
4	Passenger Satisfaction Rate	C20	3.9905	Customer
5	Women in Workforce Ratio	C25	3.9465	Learning and Growth
6	Average Fleet Age	C21	3.9460	Internal
7	Number of Aircraft	C23	3.9190	Internal
8	Handling Cost per Landing	C14	3.9036	Financial
9	ROA	C7	3.8644	Financial
10	Passenger Load Factor	C19	3.8276	Customer
11	Revenue Yield	C11	3.7567	Financial
12	CASK	C8	3.6014	Financial
13	Operational Costs Ratio	C3	3.5561	Financial
14	Passenger Revenue per ASK	C10	3.5503	Financial
15	Share of Passenger Revenue	C1	3.5441	Financial
16	Maintenance Cost per BH	C13	3.3934	Financial
17	EBITDAR Margin	C6	3.3555	Financial
18	PCSE Ratio	C15	3.3545	Customer
19	Average Response Days to Customer Complaints	C18	3.3519	Customer
20	Share of Female Employees in IGP	C30	3.3429	Learning and Growth
21	Non-Fuel CASK	C9	3.1915	Financial
22	Ethical Scandals	C29	3.1490	Learning and Growth
23	Share of Cargo Revenue	C2	2.7616	Financial
24	Aircraft Ownership Cost per BH	C12	2.7448	Financial
25	Number of Landings	C16	2.6950	Customer
26	Discrimination Cases	C28	2.4509	Learning and Growth
27	Passengers Carried	C17	2.2337	Customer
28	Cash Flow Ratio	C4	2.1879	Financial
29	ASK	C22	2.1872	Internal
30	NPM	C5	2.0133	Financial

. This outcome emphasizes that employee commitment and intellectual capital substantially influence the overall company success. Furthermore, the customer dimension, which ranks third in the total distribution (19.45%), surpasses financial KPIs in the individual weight ranking, highlighting the crucial importance of service quality. Conversely, the internal dimension accounts for the lowest total weight (10.05%). While this finding does not imply that operational efficiency is inconsequential to company performance; rather, it suggests that the impact of operational processes on performance tends to manifest over the long term and is often overshadowed by the more immediate influences of financial, learning and growth, and customer factors. On an individual basis, the top ten significant KPIs include two internal factors which are listed ahead of financial factors.

Albeit representing a smaller portion—as a whole—of the overall weight distribution, non-financial KPIs carry a strategic significance on an individual basis. This distinction between cumulative and individual weight distributions highlights valuable insights regarding the drivers of airline performance. Our findings indicate that the most significant criterion is the employee retention rate (w₂₆ = 4.1202) within the learning and growth pillar. This suggests that, given the airline industry’s dependence on a highly specialized workforce (where for the total workforce, w₂₄ = 4.0034)—including pilots, cabin crew, and operational and maintenance staff—the long-term retention of employees plays a crucial role in ensuring operational continuity and service quality. In other words, qualified human capital emerged as the top-ranking criterion since high employee turnover not only increases training costs but also diminishes operational efficiency [113]. This finding is consistent with the prior literature (see [114]) as well as the status quo of the airline market. As stated by the ref. [115], 37% of ground handling professionals anticipate staff shortages through 2023 and beyond, while 60% are concerned about not having enough skilled labor for operations, whereas 27% fear the potential departure of current employees in the following periods. Accordingly, the IATA has outlined plans to increase online competency-based training initiatives, invest in automation to reduce manual workloads for employees, and develop rewarding systems for career advancement aimed at addressing the workforce shortage [115]. These findings underscore the imperative and superior role of Human Resources Management in strategic management and competitive positioning, driven by regulatory changes and cost pressures, as well as employee safety and competency in the sector [116].

Moreover, the following top criteria with the highest weight also belong to the criteria related to female employment ratios (where the female/male ratio in middle management ratio with w₂₇ = 4.0570, and female employee ratio with w₂₅ = 3.9465), which should be interpreted as an increasing significance of gender equality. As indicated by [117], the talents and competencies of female employees contribute notably to company performance by driving productivity and generating added value within the workforce. Although the presence of women in revenue-generating positions (w₃₀ = 3.3429) contributes to the enhancement of company performance, the MEREC results show that it has a lower impact on operational success since this ratio encompasses a narrow spectrum of job positions.

When passengers perceive that their needs and expectations are adequately met, they tend to award airline companies with increased passenger satisfaction rates in acknowledgment of the high-quality service provided, which plays a pivotal role in survival within the industry [118]. The prioritization of the passenger satisfaction rate (w₂₀ = 3.9905) over the passenger load factor (w₁₉ = 3.8276), number of landings (w₁₆ = 2.6950), ASK (w₂₂ = 2.1872), and passengers carried (w₁₇ = 2.2337) reflects the significance of maximizing the passenger experience rather than passenger and flight numbers since the last two indicators are contingent upon externalities such as demand elasticity and market conditions. Younger fleet age (w₂₁ = 3.9460) is assumed to improve operational efficiency, attributed to lower fuel consumption and technical maintenance needs [119], with superior interior designs (e.g., comfortable seating arrangements) together with advanced technologies (e.g., entertainment systems) that enhance customer satisfaction and minimize penalties regarding non-compliance with security regulations. As the airline industry requires heavy capital expenditure, airline companies with a larger fleet (w₂₃ = 3.9190) can provide service to more destinations and thus increase customer volume, gain competitive advantage, manage capacity in response to demand fluctuations, and even increase market share, including during peak seasons or times of crisis [120].

The greater weight associated with the handling cost per landing (w₁₄ = 3.9036) over ROA (w₇ = 3.8644) can be attributed to several factors. While the handling cost per landing represents a direct cost component that varies based on the daily operational activities of airline companies, including factors such as landing and takeoff frequency and flight distance, as well as costs associated with airport fees, ground services, and landing/takeoff operations, ROA reflects the efficiency of long-term asset utilization. Given the substantial fixed-asset investments in the industry, ROA may not effectively capture the nuances of operational costs and could present comparatively lower values when benchmarked against other industries. Furthermore, the last positioning of NPM (w₅ = 2.0133) highlights the importance of emphasizing operational continuity in the long term rather than short-term profitability, since the latter is contingent upon multiple factors including economic shocks, natural diseases, and governmental restrictions [121]. The lower ranking of NPM relative to the EBITDAR margin (w₆ = 3.3555) indicates NPM’s sensitivity to factors such as fuel prices, exchange-rate volatility, and debt financing. In other words, as the EBITDAR margin reflects performance without being influenced by capital structure and tax-related issues, it provides more stable inferences and conclusions.

The mixed ranking of cost and revenue KPIs between 8 and 18th places in our results reveals that, contrary to the traditional airline performance evaluation approach as obtained by, for instance [8], cost-based indicators do not take precedence over revenue-based indicators. In other words, our findings shed light on the necessity of the equilibrium between both cost and revenue metrics for achieving performance optimization in the airline industry.

Although CASK and the operational cost to total revenue are two primary indicators of cost-efficiency assessments, the greater weight of CASK (w₈ = 3.6014) compared to the weight of the operational cost to total revenue (w₃ = 3.5561) underscores the critical proportion of unit costs for each seat for every km. Notably, while CASK is positioned in 12th place, non-fuel CASK (w₉ = 3.1915) is ranked in 21st place, which underscores the substantial surge in fuel costs, which are anticipated to escalate by 53% in 2024 relative to 2019, constituting approximately 32% of total airline expenditures, up from 25% five years prior [122]. In addition, this finding is consistent with the projections of [123], which estimated a slight decrease in non-fuel unit costs per ATK from 39.2 cents to 39 cents in 2024, returning to pre-pandemic levels.

Additionally, the higher weight of the revenue yield (w₁₁ = 3.7567) compared to the passenger RASK (w₁₀ = 3.5503) suggests the ever-increasing importance of the diversification of ancillary revenue streams rather than a sole emphasis on ticket sales, e.g., cargo services and premium offerings, especially in the post-pandemic era [124]. Nevertheless, the share of passenger revenue (w₁ = 3.5441) remains more critical than the share of cargo revenue (w₂ = 2.7616). This finding is consistent with the optimistic outlook of [125] for passenger revenues in 2025, when average airfare is projected to decrease by 1.8% compared to last year, representing a 44% reduction compared to the previous decade, and an 8% increase is anticipated in RPK. Furthermore, the survey data of [125] indicates that 53% of respondents intend to maintain their current travel frequency, while 41% plan to travel more often, supporting a positive forecast for passenger revenue.

Moreover, the maintenance cost per BH (w₁₃ = 3.3934), which directly affects daily flight operations, draws attention to the importance of operational costs over leasing agreements, external financing, and taxation. Especially, as leasing becomes increasingly preferred over ownership in the airline industry—accounting for 58% of the total fleet by 2023 [126]—the aircraft ownership cost per BH (w₁₄ = 2.7448) ranks even lower in the ranking.

The greater value of catering and service expenses per passenger (w₁₅ = 3.3545) than the average number of days to respond to customer complaints (w₁₈ = 3.3519) suggests that proactive service quality is more effective than reactive service quality in customer satisfaction.

Both ethical scandals (w₂₉ = 3.1490) and discrimination cases (w₂₈ = 2.4509) are assigned relatively lower weights in the evaluation of performance metrics for airline companies. This is primarily due to the emphasis on cost management, operational processes, and customer experience, which are considered more critical than other KPIs.

The cash flow ratio (w₄ = 2.1879) is assigned to have the lowest weight with the NPM among all financial KPIs. The logic behind adding this KPI is the significance of cash flow in capital-intensive industries like the airline industry, as well as the severe cash flow problems the industry faced during the pandemic [127]. Hence, this result is also consistent with the status quo of the industry as approximately USD 77 billion in cash flew away during the second half of 2020 which corresponds to USD 13 billion/month, while this loss decreased by USD 6 billion/month in the following year, reported by [128].

The MEREC results provide an objective weighing of the performance criteria, highlighting the cumulative dominance of financial indicators within the BSC pillars, while emphasizing the critical role of individual non-financial KPIs.

To provide a ranking of the airline’s performance in different periods based on the selected KPIs, the third step of our model involves the application of the CoCoSo method to integrate the weights obtained with MEREC. This allows a comparative assessment of the airline’s annual performance, especially during the pandemic and in the post-pandemic period.

4.2. CoCoSo Results

After deriving the criteria weights from the MEREC method, the performance ranking for the period spanning from 2020 to 2023 was evaluated as an alternative timeframe to identify when the company demonstrated optimal performance. It should be noted that these years represent turbulent periods springing from the COVID-19 pandemic and, therefore, this analysis aims to observe the ability to adapt to ever-changing contingencies. We also underline that the performance ranking across the years implicitly reflects the impact of annual industrial and economic developments, including pandemic-related shocks, on our outcomes.

Table 4. CoCoSo results and rankings.

Year	CoCoSo Score	Ranking
2020	33.1452	4
2021	46.0975	1
2022	44.4806	2
2023	39.9762	3

illustrates the final performance scores and annual ranking outcomes utilizing the CoCoSo method. This ranking not only indicates the performance shifts of the company from during the pandemic to the post-pandemic recovery phase but also serves as a foundation for a comprehensive analysis. The results reveal that the lowest performance occurred in 2020, with a score of 33.1452 points, whereas the best performance was recorded in the following year, increasing by 38.08% to 46.0975 points. The significant decline in passenger transportation resulting from domestic and international flight restrictions during the COVID-19 pandemic had a profound detrimental impact on the aviation industry. However, a surge in performance was observed in 2021, aligning with the global post-pandemic travel boom, as airline companies reported a global ASK of 51.2% and a global passenger load factor of 67.2%, as indicated by [129]. The improved performance was characterized by a favorable balance, wherein both financial and non-financial revenue-generating KPIs outweighed cost-based metrics. This yielded a relevantly more stable performance in 2022 and 2023.

4.3. Sensitivity Analyses

To test the robustness of our model, we perform two-stepped sensitivity analyses. While our model reflects performance trends based on current KPIs, the sensitivity analyses reflect methodological consistency rather than the company’s responsiveness to dynamic industry circumstances beyond the observation period. The steps of the sensitivity analyses are described as follows.

4.3.1. Sensitivity to Weight Variations

Assessing the stability of weights in response to minor perturbations and determining whether the performance is sensitive to specific criteria is crucial for ascertaining the reliability of the model. For this purpose, we change the weights obtained by MEREC by ±5%, ±10%, ±15%, ±20%, and ±30%. Then, we recalculate the CoCoSo scores to test whether the rankings of the decision alternatives change.

Our results generate the same orders related to the performance, as shown in

Table 5. Sensitivity to weight variations.

	2020	2021	2022	2023
Score (w)	33.1452	46.0974	44.4806	39.9762
Score (w + 5%)	33.1094	46.0514	44.4536	39.9590
Score (w − 5%)	33.1811	46.1437	44.5078	39.9932
Score (w + 10%)	33.0738	46.0055	44.4266	39.9415
Score (w − 10%)	33.1452	46.0974	44.4806	39.9762
Score (w + 15%)	33.0383	45.9597	44.3996	39.9238
Score (w − 15%)	33.2535	46.2366	44.5624	40.0263
Score (w + 20%)	33.0030	45.9141	44.3728	39.9060
Score (w − 20%)	33.2900	46.2832	44.5898	40.0423

. The integrated BSC-MEREC-CoCoSo model can be considered robust as small weight variations do not cause significant deviations in the rankings.

4.3.2. Sensitivity to Aggregation Parameter (λ) Variations

In the second step, the λ values were assigned from 0 to 1 in 0.1 increases to measure the stability of the performance rankings obtained with the CoCoSo method. The value of λ = 0 calculates only the Power Weighting Score (PWS) and the performance becomes vulnerable to low values of the criteria weights, while the value of λ = 1, only the Simple Additive Weighting Score (SAW) is used, and the score of the alternatives from all criteria is taken into account.

As being evident in

Table 6. Sensitivity to λ variations.

λ	2020	2021	2022	2023
λ = 0	33.1614	46.1005	44.4746	39.3436
λ = 0.1	33.1596	46.1001	44.4753	39.6415
λ = 0.2	33.1573	46.0997	44.4762	39.7360
λ = 0.3	33.1544	46.0992	44.4772	39.8155
λ = 0.4	33.1381	46.0850	44.4661	39.8916
λ = 0.5 (baseline)	33.1452	46.0974	44.4806	39.9762
λ = 0.6	33.1374	46.0960	44.4835	40.0726
λ = 0.7	33.1247	46.0936	44.4882	40.1954
λ = 0.8	33.1007	46.0891	44.4970	40.3742
λ = 0.9	33.0378	46.0774	44.5199	40.7115
λ = 1	32.4330	45.9710	44.7215	42.4127

, for all different values of λ, the best performing years consistently remained 2021 and 2022, consecutively, which signals the stability of the performance assessment regardless of the effect of PWS and SAW. The performance of 2023 increased by 7.8% when λ shifted from 0 to 1, while the performance of 2020 decreased from 33.1614 to 32.4330 (−2.2%). Although no drastic deviation is observed in the rankings, these findings reflect the sensitivity in years with external shocks and economic constraints in the airline industry. Nevertheless, both different aggregation methods used for weights and λ values support the stability and the reliability of our proposed integrated BSC-based MEREC-CoCoSo model.

5. Conclusions

The lockdowns during the COVID-19 pandemic deeply affected the aviation industry due to domestic and international transportation restrictions. As these restrictions were eased, airline companies initiated great efforts to recover pandemic-related losses, leading to severe competitiveness in the industry. Therefore, performance measurement has become more crucial than ever.

This paper attempts to introduce a new model to capture the performance of the airline industry by leveraging contingency theory and both the financial and non-financial pillars of the BSC. The integrated BSC-based MEREC CoCoSo model offers an innovative approach to performance measurement based on objective MCDM methods. The weights of 30 KPIs of our comprehensive BSC framework are calculated through the MEREC by systematically omitting each of them. These KPIs are then ranked by the CoCoSo method, which yields the performance scores of each alternative.

Applying the model to THY, one of the incumbent companies in the aviation industry, revealed the worst performance in 2020, underscoring the devastating effect of the COVID-19 pandemic. The subsequent year demonstrated the effectiveness of the company’s proactive actions for the following year as the best performance was recorded. Although performance dropped in the following years, the decline was not comparable to 2020, coinciding with the economic downturns in the global aviation industry. The consistency in performance ranking observed across variations in both MEREC weights and CoCoSo aggregation parameters confirmed the reliability and robustness of our integrated BSC-based MEREC CoCoSo model.

Consequently, this model offers valuable insights for decision-makers in managerial positions seeking to enhance competitive advantage, especially in the learning and growth and internal dimensions of the BSC. Additionally, given that the aviation industry is highly capital-intensive and sensitive to external fluctuations, it serves as a comprehensive tool for financial institutions to assess a company’s financial and non-financial stability in a multidimensional framework. Furthermore, the easy replicability of this model establishes a benchmark for investors, facilitating performance comparisons across different companies within the same industry or across industries.

Last, our study addresses certain limitations. Our study can be extended by the incorporation of a larger number of companies to improve the generalizability of outcomes for the airline industry. Note that finding even a limited range of common KPIs may not be possible [8], which is a limitation faced in our study. We underscore that CoCoSo generates accurate performance scores even when certain KPIs are unavailable since MEREC takes the marginal contribution of each available criterion into account and dynamically re-optimizes the weights according to the remaining dataset rather than relying on predetermined weights. It enhances the flexibility of our model, which aims to integrate a relatively novel MCDM method with the BSC framework to highlight the complexity of performance evaluation in this sector. This generalizability limitation acknowledges the absence of a comparative benchmark. In any case, we underline that our integrated BSC-based MEREC-CoCoSo model is designed to be replicable across different airline companies in other jurisdictions, as the scoring and ranking structure provides a consistent foundation for future benchmarking studies [33] and facilitates the reproducibility of analyses. This shows the practical adaptability of our integrated model. Also, the aforementioned flexibility makes the use of our model convenient for low-cost carriers. We expect that our outcomes provide fruitful insights for future research despite these limitations. We encourage future research to employ supplementary MCDM techniques and alternative KPIs to compare changes in performance.