Widely Used Multi-Criteria Decision Analysis Methods—A Comprehensive Ranking

Abstract

This article presents a comprehensive ranking of popular Multi-Criteria Decision Analysis (MCDA) methods, assessed across six dimensions of visibility and usage: general web presence, academic mentions, number of academic publications, citation impact, reported academic applications, and real-world (industrial and public sector) use cases. Rather than evaluating theoretical performance or methodological robustness, this study focuses on visibility and adoption, the extent to which an MCDA method has achieved prominence through name recognition, dissemination, academic use, and documented practical application. The observed ranking patterns are consistent with the possibility that method popularity is partly shaped by branding, dissemination dynamics, and cumulative visibility, rather than by demonstrated methodological superiority alone. AHP emerges as the most dominant method across all six criteria, followed by TOPSIS, PROMÉTHÉE, and VIKOR. Newer or more recently visible methods such as BWM demonstrate rapid academic growth, while software-supported methods such as PAPRIKA illustrate how dedicated tools can strengthen practical visibility. By contrast, established methods like ÉLECTRE and MACBETH, while academically respected, exhibit more niche or regionally concentrated use. The study also touches on the meta-cognitive irony that MCDA methods, designed to reduce bias and promote rational deliberation, may themselves be adopted in ways consistent with mechanisms such as availability, familiarity, and brand recognition. This raises concerns about the criteria by which methodological choices are made, particularly in academic modelling and industrial decision-support contexts. Ultimately, the findings suggest that while branding plays a legitimate role in method dissemination, it should not overshadow considerations of methodological transparency, theoretical coherence, and decision-context fit. By distinguishing between methodological performance and visibility/adoption, this article contributes to a more reflective discourse on the selection, teaching and development of MCDA methods and techniques. The ranking should therefore be interpreted as a comparative map of visibility and adoption, not as an assessment of methodological quality, theoretical soundness, or decision-analytic performance.

1. Introduction

Multi-Criteria Decision Analysis (MCDA) methods help decision-makers evaluate options against multiple criteria. This study presents a comprehensive ranking of popular MCDA methods, grouped by method families (variants consolidated), and ranks them based on various metrics. We consider six criteria: general web mentions, academic mentions, number of academic articles, total academic citations, reported academic use cases, and reported industrial use cases. Each method’s placement is justified with data from the recent literature (e.g., systematic reviews, bibliometric studies) and known applications.

The MCDA methods included in this article were identified through a structured search and screening procedure described in Section 3.1. The selection was based on documented visibility and adoption across academic, professional, and real-world decision-support contexts. Together, the selected method families represent the most visible and commonly reported MCDA methods across the evidence categories considered in this study. Their usage and citation patterns suggest that method popularity may partly reflect branding success, dissemination dynamics, and institutional familiarity, rather than demonstrated methodological superiority alone. The prominence of certain methods appears to be associated not only with their methodological characteristics but also with factors such as memorable acronyms, compelling narratives, software availability, and academic network effects. This was noted already by Belton and Stewart [1]. Additionally, being early to the methods scene has afforded some approaches a lasting advantage, allowing them to establish a dominant position before competing approaches emerged, further reinforced by cumulative citation effects. In classic marketing theory, branding and narratives can create switching barriers and reinforce familiarity with a product or service. The proliferation of some MCDA methods is consistent with network-effect mechanisms, where name recognition and earlier citations may influence uptake, sometimes independently of fresh comparative validation or theoretical assessment. While these methods promote structured decision-making, their dissemination might have benefitted from mechanisms related to availability, familiarity, and affective association, although the present study does not directly measure such cognitive processes. A recent book [2] discusses the impact of brand names and perceived USPs (unique selling points) on methodological choices within MCDA, both for researchers and practitioners.

2. Background

Multi-Criteria Decision Analysis (MCDA) encompasses a family of formal methodologies designed to support decision-making in contexts involving multiple, often conflicting, criteria. Originating in response to the inadequacy of single-criterion optimisation frameworks for real-world problems, MCDA spans a broad range of applications [1,3]. While differing in structure and computational approach, all MCDA methods share the central objective of enabling a decision-maker to navigate trade-offs among incommensurate values in a transparent, structured, and rationally defensible manner. The fundamental challenge addressed by MCDA is that preferences over alternatives cannot be represented as a function of a single attribute without loss of crucial evaluative nuance. Alternatives may differ in cost, risk, efficiency, equity, sustainability, and qualitative factors such as political acceptability, none of which admit to aggregation under a single cardinal scale without normative assumptions. MCDA therefore constructs procedures that allow such dimensions to be weighted, synthesised, and compared in a way that reflects the decision-maker’s priorities [4].

The intellectual foundations of MCDA draw from disparate traditions, notably decision theory, operational research, and utility theory. The axiomatic structure of rational choice, as formalised in the von Neumann–Morgenstern and Savage frameworks, provides the baseline for normative coherence in many MCDA methods, especially those that aim to elicit preferences in a cardinal and compensatory manner. Yet, MCDA departs from strict Bayesian or utility-maximising frameworks by allowing for partial comparability, non-compensatory judgments, and procedures that do not necessarily presuppose completeness or transitivity in the preference relation [2]. In particular, methods such as ÉLECTRE [5] and PROMÉTHÉE [6] introduced the notion of outranking, in which alternatives may be compared on the basis of concordance and discordance rather than strict utility dominance. These methods allow non-transitivities and non-comparabilities by design, in an effort to reflect the epistemic or normative limitations of real-world judgments. Other approaches, notably the Analytic Hierarchy Process (AHP; [7]), rely on pairwise comparisons to derive ratio-scale priorities from qualitative judgments, albeit at the cost of imposing completeness and encouraging full comparability, even in contexts where such judgments may be context-dependent or unstable.

A central distinction within MCDA methodology lies between compensatory and non-compensatory approaches. Compensatory methods, such as value function models and weighted summation techniques, allow trade-offs between criteria, thereby granting high performance on one criterion the capacity to offset poor performance on another [4]. This assumption aligns with classical economic models of rational choice but has been criticised in domains where moral, ethical, or operational constraints render certain thresholds non-negotiable. Non-compensatory methods, on the other hand, often reflect the view that some criteria represent hard constraints or categorical imperatives; such models are frequently employed in public sector decision-making or in participatory frameworks where stakeholder values are diverse and not easily commensurable [1]. For example, ÉLECTRE employs veto thresholds that can block an alternative from being selected despite high performance in most other respects. The theoretical divergence between these paradigms reflects a common misunderstanding in decision analysis: whether decision support tools should model an idealised rational agent or not. The actual question is whether the computational core should be conceptually and theoretically sound or be confounded with behavioural process details. The answer is that an outer MCDM layer (‘M’ as in decision-making, a wider domain than ‘A’ as in analysis) should address the cognitive and institutional complexity of actual decision environments.

The differences in philosophy and the different brandings of methods should influence the elicitation processes, the presentation formats, group decision mechanisms, and much more—as long as the methods stand on established scientific ground. Substituting a long-established and sound axiomatic computational core for homemade calculi only leads to questionable results and opaqueness, as does mixing descriptive and psychological factors with an axiomatically grounded computational core; the former should belong only to an outer MCDM layer. The need to stand out by branding and perceived uniqueness should be satisfied in other ways.

MCDA is not a monolithic field; rather, it is defined by a heterogeneous landscape of methods, each characterised by their own assumptions about preferences, aggregation, and the nature of acceptable trade-offs. In the deterministic domain, key methods include weighted sum (SAW) models such as SMART, as well as TOPSIS, VIKOR, ÉLECTRE, PROMÉTHÉE, and AHP. Each of these methods introduces a unique mechanism for combining scores across criteria and resolving conflicts between alternatives [3,8]. For instance, TOPSIS constructs a geometric compromise between an ideal and an anti-ideal solution in normalised criterion space, while VIKOR employs a regret-based ranking by evaluating the closeness of alternatives to the best values for each criterion, adjusted by maximum and average deviations [3]. The PROMÉTHÉE methods generate positive and negative outranking flows by aggregating preference functions defined for each criterion, facilitating both partial (PROMÉTHÉE I) and complete (PROMÉTHÉE II) rankings [6]. These methods are typically applied under conditions of certainty, where performance evaluations are treated as known or deterministically estimated.

Probabilistic or fuzzy extensions of MCDA methods have been developed to address the reality that information is often uncertain and incomplete. Fuzzy MCDA approaches use linguistic variables and membership functions to express vague preferences, whereas stochastic MCDA introduces probability distributions over performance outcomes [3,8]. Such variants complicate both the computation and interpretation of results, raising questions about elicitation procedures and the cognitive load placed on decision makers. Nevertheless, their development reflects a broader commitment within MCDA to capture the inherent uncertainty and ambiguity of complex decision problems.

From a prescriptive standpoint, MCDA aims not merely to rank alternatives but also to support decision-makers in structuring their problem, clarifying values, and justifying outcomes [9]. This involves not only computational output but also the design of decision processes that are inclusive and communicative. Several MCDA methodologies have therefore been embedded into participatory decision support systems [3]. The adoption of MCDA in these settings is driven not only by its analytical rigour but also by its ability to foster dialogue among stakeholders with divergent perspectives and conflicting objectives. In this respect, MCDA operates as both a normative and procedural framework.

The formal properties of MCDA methods such as monotonicity, invariance, dominance preservation, and rank reversibility have been the subject of extensive theoretical scrutiny [2,8]. Desiderata for rationality and consistency have been proposed to evaluate and compare the normative adequacy of competing methods. For example, the independence of irrelevant alternatives (IIA) is violated by most ranking-based methods under certain conditions. However, strict adherence to IIA may itself be questionable in contexts where the presence of alternatives changes the interpretive frame of the decision. Similarly, weight sensitivity and scale dependence are active concerns: methods that rely on normalisation or ratio scales may produce outcomes that shift dramatically under rescaling.

3. Ranking of MCDA Methods

The methodological literature has increasingly emphasised the need for meta-criteria for method selection. These include considerations of computational tractability, transparency, ease of use, stakeholder acceptability, and theoretical coherence. Indeed, there is no universally preferred MCDA method. Rather, method selection must be grounded in an understanding of the decision context and the institutional or stakeholder constraints under which the analysis will be interpreted and implemented. Given these insights on method selection, this article examines the most popular MCDA brands from six perspectives—web presence, the academic literature, publication count, citations, academic use cases, and industrial use cases (including public sector)—beginning with a total ranking synthesised from the six different perspectives investigated.

3.1. Data Collection and Preprocessing

The purpose of the empirical mapping was to compare the visibility and adoption of widely known MCDA methods across several forms of dissemination and use. The study was therefore designed as a structured comparative mapping rather than as a systematic review of methodological performance. The unit of analysis was the named MCDA method or method family, not individual method variants, software implementations, or fuzzy/probabilistic extensions.

The candidate set of methods was first identified from major MCDA textbooks, surveys, bibliometric reviews, and application reviews, including [1,3,8,10,11], and related method-specific review papers. Methods were included if they met at least one of the following criteria: (i) repeated appearance in general MCDA/MCDM surveys, (ii) substantial publication or citation presence, (iii) documented use in academic application studies, or (iv) documented use in real-world organisational, industrial, or public-sector decision processes.

Searches were conducted during 2025 and updated in April 2026. The search period for academic publications and citations extended from the year of introduction of each method to that final update date. For each method, searches used both the acronym and the full method name where applicable. For methods with spelling variants, diacritics, or translations, variant forms were included, covering different spellings. For ambiguous acronyms, such as AHP, ANP, SAW, or SMART, searches were combined with full names and MCDA/MCDM-related terms to reduce false positives.

The six evaluation criteria were selected from different but complementary sources. General web presence was assessed using general-purpose web search results, tutorials, software pages, professional blogs, and visible non-academic material. Academic mentions and publication counts were assessed using bibliographic databases and the published counts in recent bibliometric reviews. Citation impact was assessed using Google Scholar and cited resources. Academic use cases were identified from application studies and review articles in which a method was actually used to analyse a decision problem, rather than merely mentioned. Industrial and public-sector use cases were identified from documented case studies, software-provider material, consultancy reports, public-sector applications, and peer-reviewed reports of real organisational use.

Several preprocessing rules were applied before ranking. First, method variants were consolidated into method families where they clearly shared the same methodological brand and intellectual origin. Thus, SMART/SMARTS/SMARTER/SAW/weighted-sum approaches were treated as a broader weighted-scoring family. Second, fuzzy, stochastic, interval, grey, or hybrid extensions were not counted as separate methods unless the extension had become a distinct named method family. Third, generic uses of weighted scoring were included only where the procedure clearly corresponded to a weighted additive model. Fourth, records were screened to exclude false positives caused by acronym ambiguity, unrelated uses of method names, duplicate records, and papers where a method was only mentioned but not used.

Because the six criteria differ in measurement scale and evidential reliability, the study did not aggregate raw counts directly. Search-engine hit counts, bibliographic counts, citation totals, and documented organisational cases are not commensurable quantities. Instead, the evidence was used to construct six ordinal rankings, one for each perspective. These rankings were then normalised and aggregated as described in Section 3.2. The resulting comprehensive ranking should therefore be interpreted as a reproducible synthesis of visibility and adoption indicators, not as a claim about theoretical validity or decision-analytic performance. See

Table 1. Data sources and search logic for the six evaluation criteria.

Criterion	Main Data Sources	Search Logic
General web presence	General web search, tutorials, software pages, professional blogs, public explanatory material.	Acronym and full method name; spelling variants; MCDA/MCDM context terms for ambiguous acronyms.
Academic mentions	Scopus, Web of Science, Google Scholar searches, and recent MCDA reviews.	Method name in title, abstract, keywords, or method discussion.
Publication count	Bibliometric reviews and database counts.	Published method-specific counts supplemented by database searches.
Citation impact	Google Scholar, Scopus/Web of Science citation data, and landmark publications.	Seminal-publication and method-family citation searches.
Academic use cases	Application papers, review articles, and domain-specific MCDA surveys.	Method name combined with terms such as application, case study, supplier selection, site selection, prioritisation, or ranking.
Industrial/public-sector use cases	Published organisational case studies, public-sector reports, software-provider case material, and consultancy examples.	Method name combined with terms such as industry, public sector, government, software, decision support, and case study.

for a summary of the data collection and

Table 2. Preprocessing and ranking outputs for the six evaluation criteria.

Criterion	Screening/Preprocessing	Output Used for Ranking
General web presence	False positives removed; method variants consolidated; generic mentions assessed cautiously.	Ordinal rank of visible non-academic presence.
Academic mentions	Duplicate and unrelated records removed; acronym ambiguity checked.	Ordinal rank of academic visibility.
Publication count	Variants grouped by method family; hybrid, fuzzy, stochastic, interval, or grey extensions folded into the parent method unless distinct.	Ordinal rank of publication volume.
Citation impact	Inflated or duplicate citation counts treated cautiously; emphasis on order of magnitude.	Ordinal rank of citation influence.
Academic use cases	Included only when the method was actually applied to a decision problem, not merely mentioned.	Ordinal rank of academic application frequency.
Industrial/public-sector use cases	Purely illustrative examples excluded; public-sector use counted as real-world use.	Ordinal rank of real-world adoption.

for the preprocessing.

Numerical and adoption evidence are reported according to the following convention. Publication counts are linked to bibliometric or systematic reviews where available, and otherwise to author database searches conducted during the original collection period and updated in April 2026. Citation estimates are linked to landmark publications and citation-database or Google Scholar records but are interpreted as conservative order-of-magnitude indicators because they vary by database, duplicate handling, and search date. Industrial and public-sector adoption claims are based on documented organisational cases, public reports, peer-reviewed case studies, software-provider material, or consultancy examples; where provider-reported material is used, this is identified as such. Across all categories, numerical values are used to support ordinal component rankings rather than as precise, directly commensurable measurements. Thus, reported numbers should be viewed as general indicators, pointing out the order-of-magnitude differences between top-ranked and lower-ranked methods.

3.2. Normalisation, Weighting, and Aggregation

The ranking workflow consisted of five steps. First, candidate MCDA methods and method families were identified from major textbooks, systematic reviews, bibliometric studies, and method-specific review papers. Second, evidence was collected for six visibility and adoption indicators: general web presence, academic mentions, publication count, citation impact, academic use cases, and industrial or public-sector use cases. Third, method variants and spelling variants were consolidated into common method families, and ambiguous acronym matches and duplicate records were screened. Fourth, each of the six indicators was converted into an ordinal top-ten ranking and normalised using a Borda-type rank score. Fifth, the six normalised scores were assigned equal weights and averaged to obtain the composite popularity score used for the comprehensive ranking.

The comprehensive ranking was constructed from the six perspectives reported in Section 4: general web presence, academic mentions, publication count, citation impact, academic use cases, and industrial or public-sector use cases. These six perspectives were treated as complementary indicators of method visibility and adoption, not as indicators of theoretical quality or normative validity.

Because the underlying evidence differs in scale and reliability across the six perspectives, direct aggregation of raw counts was not used. Web search counts, bibliometric publication counts, citation totals, and documented use cases are not commensurable quantities and are affected by different sources of measurement error. Instead, each perspective was first converted into an ordinal rank list. For each perspective c, a method or method family i appearing in position r_ic received a normalised rank score.

$${\mathit{s}}_{\mathit{i}\mathit{c}}={\displaystyle \frac{\mathit{K}-{\mathit{r}}_{\mathit{i}\mathit{c}}}{\mathit{K}-1}}$$

where K = 10 for the six perspective-specific lists in Section 4. Thus, the first-ranked method received a score of 1, the tenth-ranked method received a score of 0, and intermediate ranks were linearly interpolated. Methods not appearing in a given top-ten list were assigned a zero score.

The six perspectives were assigned equal weights, reflecting the purpose of the article: to capture broad method visibility across several forms of dissemination and use, rather than privileging one particular kind of evidence. The composite popularity score for method i was therefore computed as

$${\mathit{S}}_{\mathit{i}}={\displaystyle \frac{1}{6}}{\displaystyle \sum _{\mathit{c}=1}^6}{\mathit{s}}_{\mathit{i}\mathit{c}}$$

Equal weighting was chosen because there is no non-arbitrary basis for claiming, for example, that citation impact should count more than industrial uptake, or that web visibility should count less than academic applications. The resulting ranking should therefore be interpreted as a transparent synthesis of six visibility indicators, not as an optimisation result or a measure of methodological superiority. Where composite scores were close, the final ordering was checked against the qualitative evidence in Section 4, especially whether the method showed balanced visibility across several perspectives or was concentrated in only one domain.

Table 3. Ranking results and composite score.

Overall Rank	Method Family	Web Presence	Academic Mentions	Publication Count	Citation Impact	Academic Use Cases	Industrial Public Use	Composite Score
1	AHP	1	1	1	1	1	1	1.000
2	TOPSIS	2	2	2	2	2	5	0.833
3	PROMÉTHÉE	4	4	4	3	3	4	0.704
4	VIKOR	6	3	3	5	4	–	0.537
5	ÉLECTRE	5	5	6	4	5	7	0.519
6	SMART/SAW	3	7	7	7	–	2	0.444
7	ANP	7	6	5	6	6	8	0.407
8	PAPRIKA	8	–	–	–	9	3	0.185
9	BWM	9	8	–	8	7	9	0.167
10	MACBETH	10	9	–	9	8	6	0.148
11	TODIM	–	–	8	–	–	–	0.037
12	COPRAS	–	–	9	–	–	–	0.019

reports the 12 method families that appeared in at least one of the six top ten component rankings in Section 4. Additional methods discussed below, including WASPAS, MULTIMOORA, SWARA, DRSA, UTA/UTADIS, FITradeoff and THOR, are mentioned as relevant specialised methods but are not assigned separate composite ranks.

As a sensitivity check, the ranking was recalculated under two simple perturbation schemes; see

Table 4. Sensitivity of composite ranking under alternative weighting checks.

Method	Equal-Weight Rank	Leave-One-Out Rank Range	One-Criterion-Doubled Rank Range
AHP	1	1–1	1–1
TOPSIS	2	2–2	2–2
PROMÉTHÉE	3	3–3	3–3
VIKOR	4	4–5	4–6
ÉLECTRE	5	4–6	4–5
SMART/SAW	6	4–7	4–7
ANP	7	6–7	6–7
PAPRIKA	8	8–10	8–9
BWM	9	8–9	8–10
MACBETH	10	8–10	9–10
TODIM	11	11–11	11–11
COPRAS	12	11–12	12–12

. First, each of the six criteria was excluded in turn and the composite score was recomputed from the remaining five criteria. Second, each criterion was doubled in weight in turn, while the remaining five criteria were assigned equal shares of the remaining weight. These checks test the stability of the aggregation results reported in Table 3.

The results show that the top three methods are highly stable. AHP remains first, TOPSIS second, and PROMÉTHÉE third in all tested scenarios. The middle-ranked methods are more sensitive to weighting assumptions: VIKOR varies between ranks 4 and 6, ÉLECTRE between ranks 4 and 5, SMART/SAW between ranks 4 and 7, and ANP between ranks 6 and 7. The lower-ranked methods also show some movement, especially PAPRIKA, BWM, and MACBETH, reflecting their more uneven visibility across academic and practical indicators. Overall, the sensitivity analysis supports the conclusion that the ranking is robust at the top but less stable in the middle range, where different defensible weighting choices can shift the ordering.

These checks provide a quantitative basis for the interpretation. Methods such as AHP and TOPSIS remain highly ranked across nearly all criteria, indicating broad visibility and cumulative adoption.

3.3. The Total Ranking

The total ranking list is ordered by the composite popularity score obtained from the six normalised perspective-specific rank scores described above. The rank positions reported in Section 4 constitute the input to the aggregation procedure. The resulting composite score is reproducible from the six component rankings, while the underlying evidence for each component ranking is discussed qualitatively in the corresponding subsection. The total list and the separate topic lists have the same overall format. They present ten entries (the overall list has 12) in ranking order, with the highest ranked according to the total (this section) or specific metric (Section 4).

1. Analytic Hierarchy Process (AHP). This is by far the most widely used MCDA method across the board. Developed by Saaty [7], it structures decisions into hierarchies and uses pairwise comparisons to derive priority weights. Web presence: AHP has a massive general web footprint, with its Wikipedia article and innumerable tutorials, tools, and case studies online. Academic prevalence: It dominates academic literature with the highest presence of any MCDA method. It is the top method in publication count globally. The seminal AHP book by Saaty [12] has accumulated tens of thousands of citations, reflecting its exceptional influence; the exact count depends on the citation database and search date. Use cases: AHP appears in virtually every application domain (business, government, engineering, healthcare, education, etc.). In industry, AHP is frequently used for complex decision-making and is supported by commercial software (e.g., Expert Choice). It is employed by organisations for project prioritisation, vendor selection, policy planning, and much more.

2. TOPSIS (Technique for Order Preference by Similarity to Ideal Solution). This is a popular distance-based ranking technique [13]. Web presence: It is well-represented online through examples in engineering and data science communities (e.g., many “how-to” guides on decision matrices). Academic prevalence: It consistently ranks just behind AHP. Basílio et al. [10] report TOPSIS as one of the most frequent MCDA methods in publication counts, with several thousand indexed publications. Bibliometric analyses place it as the second most-used MCDA method globally. Use cases: It has broad appeal because of its intuitive concept of an “ideal” solution that has somehow caught on as a USP (unique selling point). It is applied in numerous domains. Industrial uptake exists where a simple computational method is needed for ranking alternatives; TOPSIS is often used in decision-support tools and by analysts (not least in engineering companies) for tasks like supplier selection or project ranking.

3. PROMÉTHÉE (Preference Ranking Organisation Method for Enrichment Evaluations). This is a family of outranking methods [6]. Web presence: It has a moderate web presence—known in OR/MS circles with dedicated software (Visual PROMÉTHÉE, D-Sight, etc.) and a well-visited Wikipedia entry. Academic prevalence: PROMÉTHÉE is the most popular outranking MCDA approach. Approximately 1400 academic publications involve PROMÉTHÉE. In global rankings, it is always one of the top five methods used. Use cases: It has been applied in diverse fields including business, finance, hydrology, environmental management, logistics and more. Its strength in handling preference flows makes it popular for environmental and sustainability decision problems. While not as ubiquitous in industry as AHP, it has seen use in practical contexts (e.g., energy planning, supply chain evaluation) via consulting projects and decision support systems.

4. VIKOR (VlseKriterijumska Optimizacija I Kompromisno Resenje). VIKOR, originally IKOR [14], is a compromise-ranking method emphasising closeness to an ideal solution [15]. Web presence: Primarily discussed in academic and technical sources; modest general web mentions. Academic prevalence: Its academic footprint grew rapidly in the 2000s after its brand change. It appears in around 1500 publications, making it one of the top three methods in recent decades. For example, in international publications, it is ranked only after AHP and TOPSIS in frequency of use. Use cases: It is popular in scholarly research for engineering, manufacturing, and supply chain problems that seek a “compromise” solution. It often appears alongside TOPSIS for comparative studies. Industrial use is less common; VIKOR tends to be used by researchers and in analytical evaluations rather than as a standard industry tool. Still, it has been applied to practical problems (e.g., selecting project sites, evaluating policies) in collaboration with domain experts, especially in Asia and Eastern Europe, where it originated.

5. ÉLECTRE (ELimination Et Choix Traduisant la REalité). It is a family of outranking methods focusing on “choosing” and “ranking” via pairwise dominance and concordance/discordance tests [5]. Web presence: Moderate but less than AHP/TOPSIS. It is very well-known in specialised MCDA contexts, especially in Europe. Academic prevalence: Across all ÉLECTRE variants (I, II, III, IV, etc.), it appears in roughly 1000 academic publications. ÉLECTRE is a classical outranking approach, though its usage in recent years has been overtaken by PROMÉTHÉE in the outranking category. Use cases: Historically, ÉLECTRE was employed in European government and infrastructure decision projects (e.g., energy, transportation planning) and is well-known in theoretical MCDA literature. Many academic case studies (urban planning, public resource management, etc.) have used ÉLECTRE. In industry, its adoption has been limited; it is more common in public sector or policy analysis settings than in private companies. Nonetheless, it remains an important method taught in decision analysis, and some decision support systems (especially in France) have implemented ÉLECTRE for multi-criteria evaluations.

6. Weighted Sum/Rating Techniques: SMART and variants (SMARTS/ER, SAW, WSM). These refer to the family of methods where criteria are weighted (via direct rating or simple procedures) and alternative scores are summed—including SMART (Simple Multi-Attribute Rating Technique) by Edwards [16], its variants like SMARTS or SMARTER, and the closely related SAW (Simple Additive Weighting) or WSM (Weighted Sum Model) families. Web presence: High in concept, but low under specific names. The approach of weighted scoring is ubiquitous on the web (e.g., countless “decision matrix” or “pro–con weighting” examples. However, terms like “SMART” or “SAW” are less often explicitly cited online compared to AHP or TOPSIS. Academic prevalence: Few papers focus on these as novel methods. Nonetheless, weighted scoring is the de facto baseline in many studies (often used but not always named). For instance, SAW is noted as a “simple, weight-based sum method closest to daily life use”, underscoring its practical importance. Use cases: This category is extremely common in practice, arguably the most widely applied approach in industry for multi-criteria decisions, due to its simplicity. Many companies and government bodies use weighted scorecards or utility scoring (which is essentially SMART/SAW) for decision-making without needing specialised software. Academically, SMART/SAW often appears as a benchmark to compare more complex MCDA methods. In industrial settings, when decisions need to be made quickly and transparently, a weighted criteria scoring (sometimes under the guise of “ranked voting” or “point systems”) is frequently used.

7. ANP (Analytic Network Process). ANP [17] generalises AHP to allow network relationships (interdependencies) instead of a strict hierarchy. Web presence: Moderate—often discussed in the context of AHP (some websites and forums cover it as an advanced Saaty method). Academic prevalence: ANP is frequently studied, with over 1300 publications. It often appears just behind the top four methods in bibliometric rankings. Its foundational literature (Saaty’s work) is highly cited, though its overall usage is much lower than AHP due to its complexity. Use cases: Academically, ANP is applied where criteria and alternatives have feedback influences (e.g., policy analysis with interrelated factors). It is one of the cornerstone methods in the American school of MCDA. Industrial adoption is niche: some organisations use ANP for strategic decisions (often via software like SuperDecisions), but the need to perform many pairwise comparisons in a network deters widespread commercial use. It remains primarily an expert tool for complex decision problems.

For the lower-ranked methods, exact publication counts are less central to the argument; the discussion therefore emphasises relative visibility and order-of-magnitude differences rather than precise counts.

8. PAPRIKA (Potentially All Pairwise Rankings of all possible Alternatives). This was introduced by Hansen and Ombler [18] as a method for scoring additive multi-attribute value models using pairwise rankings of alternatives. PAPRIKA is a more recent method that involves pairwise comparisons of hypothetical alternatives to determine criteria weights and scores. It is the engine behind the 1000minds decision software. Web presence: Moderate—largely via 1000minds (which extensively markets the method) and the health economics literature. Academic prevalence: PAPRIKA does not often appear, but the method is used implicitly in many studies via the 1000minds tool. According to the developers, 370 peer-reviewed articles/abstracts using PAPRIKA have been published since 2006. Use cases: PAPRIKA has strong real-world adoption relative to its academic footprint. It has extensive industrial and organisational use through the 1000minds platform. Academically, it is popular in health and resource allocation studies where stakeholders’ preferences are elicited via pairwise rankings. In summary, while traditional publication counts rank PAPRIKA low, its practical impact (especially in multi-criteria decision support and preference learning applications) is significant.

9. BWM (Best–Worst Method). This is a relatively new MCDA method [19] for deriving criteria weights from a limited set of pairwise comparisons (best vs. others, worst vs. others). It is among the fastest growing methods and its ranking is expected to increase. Web presence: Emerging—discussed in research blogs and some online tutorials as a promising alternative to AHP for weighting. Academic prevalence: In just a decade, BWM has gained significant traction. The foundational BWM paper [19] has accumulated several thousand citations, indicating rapid academic uptake for a comparatively recent method. Dozens of BWM applications have been published (especially in fields like supply chain, sustainability, etc.). BWM would likely rank among the top five-to-seven methods by academic mentions in recent years (eclipsing some older methods). Use cases: BWM is known for requiring slightly fewer comparisons than AHP while maintaining consistency, so academics have applied it to criteria weighting problems in, e.g., logistics, supplier selection, risk assessment and more. Industrial use is nascent but growing. Its relative simplicity (compared to AHP which is its main market target) makes it attractive for practitioners to derive weights without the full overhead of AHP. There are very few industry case studies (e.g., in project portfolio selection) using BWM, often guided by academics.

10. MACBETH (Measuring Attractiveness by a Categorical Based Evaluation Technique). This is a pairwise-comparison method that uses qualitative judgments (“difference in attractiveness” categories) to derive numerical scores [20]. Web presence: Niche—primarily known through the MACBETH software (M-MACBETH) and academic references. Academic prevalence: a moderate academic following. It is not as widely cited as the top methods, but it has a dedicated community of practice in decision analysis. Use cases: MACBETH is used in academia and some public-sector decisions to derive value functions when decision makers find quantitative pairwise ratios difficult. It has been applied in project prioritisation, resource allocation, and policy decisions (e.g., urban planning, infrastructure), especially in Europe and South America. Industrial use is limited but not absent—a few consulting cases and government agency decisions have employed MACBETH (often facilitated by its creators or trained analysts). Its strength is in group settings where qualitative consensus is sought before quantification.

11. TODIM, from the Portuguese Tomada de Decisão Interativa Multicritério, is an outranking method based on prospect-theory value functions [21]. Academic prevalence: mostly in Latin America. Use cases: Applied in academic case studies (e.g., energy, mining decisions). Limited industrial uptake globally, though better known in Brazil than in many other regions.

12. COPRAS (Complex Proportional Assessment). COPRAS was introduced by Zavadskas, Kaklauskas and Sarka [22] for complex proportional assessment of projects. It is a method that ranks alternatives by proportional evaluation of criteria significance and utility degree. Web presence: Low general awareness outside academic/engineering communities. Academic prevalence: COPRAS appears mainly in engineering, construction, and sustainability journals. It has a niche following, particularly among Eastern European researchers. Use cases: Academically, COPRAS has been applied to construction project selection, supplier evaluation, etc., often in studies from the Eastern Bloc (where it originated). Industrial use beyond those contexts is limited; it is mostly confined to academic case studies or where researchers collaborate with industry to apply it.

The section ends with some unranked methods that were part of the investigation but failed to make it into the perspective ranking lists.

WASPAS (Weighted Aggregated Sum Product ASsessment). WASPAS was introduced by Zavadskas, Turskis, Antucheviciene and Zakarevičius [23] as a weighted aggregated sum-product assessment method. Academic prevalence: Moderate use in the last decade. Use cases: Mostly academic; used in similar domains to COPRAS (engineering decisions, etc.) with few documented industry adoptions.

MULTIMOORA (Multi-Objective Optimisation by Ratio Analysis, including Reference Point). MOORA and its extension MULTIMOORA are ranking methods from the late 2000s [24]. Academic prevalence: Moderate. Use cases: Mainly academic exercises in engineering optimisation problems; minimal direct industry usage.

SWARA was introduced by Keršulienė, Zavadskas and Turskis [25]. SWARA (Step-wise Weight Assessment Ratio Analysis) is a weight elicitation technique occasionally used in recent studies, often alongside other methods. DRSA was introduced by Greco, Matarazzo and Słowiński [26]. DRSA (Dominance-based Rough Set Approach) is a qualitative MCDA approach for sorting problems. UTA was introduced by Jacquet-Lagrèze and Siskos [27]. UTA/UTADIS (Utility Additive methods for multi-attribute preference modelling) are used for rating and sorting, mainly in research. These and similar techniques, e.g., FITradeoff [28] and THOR [29], have not achieved widespread adoption yet. They might be important in specific subdomains but rank lower in general popularity.

4. Separate Perspectives on MCDA Methods

This section provides the perspective-specific evidence behind the aggregated ranking in Section 3. This section lists the top ten methods found in each of the six perspectives: web presence, academic literature, publication count, citations, academic use cases, and industrial use cases (including public sector usage). The lists were collected separately. Thus, it is not the same ten methods appearing on every list, reflecting different strengths and weaknesses among the methods. However, the number one method is the same across all six lists, making it an undisputed top performer in the aggregated list in Section 3. Some lists have “others” as the last, tenth entry, indicating that it was close between several methods at that end of those lists.

The descriptions in this section are organised by evidence dimension rather than by method. Their purpose is to justify the component rankings used in the aggregation procedure. The industrial and public-sector ranking should be interpreted with particular caution, since real-world adoption is less uniformly documented than academic publication and citation evidence. It is therefore treated here as an ordinal indicator of documented real-world visibility, not as a precise measure of total adoption volume.

4.1. General Web Mentions (Popularity on the Web)

1. AHP: This has the strongest web presence. It is widely referenced in general articles, forums, blogs, and decision-making tutorials. Its use in diverse fields and the availability of simple explanations have led to numerous Google search results and a broad online footprint.

2. TOPSIS: This also enjoys high web visibility. Many engineering and data science websites discuss it for practical decision problems (e.g., as a technique in Excel or Python libraries). Its straightforward concept (ideal and nadir solutions) makes it a common example on the web.

3. Weighted Scoring (SMART/SAW): The generic approach of weighted scoring is extremely common in online content (often presented as “priority matrices” or “score cards”). Although the specific term “SMART” might not always be used, the method itself is essentially everywhere in decision-making blogs. This gives it a high implicit web presence.

4. PROMÉTHÉE: Moderate web presence. Outside academic or professional circles, it is mentioned less, but it has dedicated sites (e.g., those for PROMÉTHÉE software and tutorials) and appears in some specialised discussions (especially in environmental decision forums).

5. ÉLECTRE: Similar to PROMÉTHÉE, ÉLECTRE has a niche web presence. It is known among practitioners in Europe; some websites and course notes discuss ÉLECTRE, but it is not as ubiquitous in general decision-making discussions.

6. VIKOR: Fewer general web mentions—mostly found on academic or technical sites. It is less likely to be encountered by non-specialists browsing the web compared to the above methods.

7. ANP: Low web profile. It is typically discussed in conjunction with AHP. It appears mainly on academic pages or forums, such as ResearchGate.

8. PAPRIKA: Its web presence comes through 1000minds’ own website and some news articles about its award-winning software. It is rather visible due to the 1000minds marketing (e.g., case studies, client testimonials), but it is not a known name outside those contexts.

9. BWM: Emerging presence. It is beginning to appear on webpages aimed at researchers, but it is still largely confined to academic discussions online. Its presence is on the rise.

10. MACBETH: This has minimal general web presence. It is usually found only by targeted search or on academic sites.

4.2. Academic Mentions (Presence in Academic Literature)

1. AHP: Absolutely dominant in academic mentions. It appears in titles/abstracts/keywords of thousands of publications—more than any other MCDA method. Essentially, every literature review of MCDA notes AHP’s prevalence.

2. TOPSIS: Second-most mentioned in academic contexts. It is frequently included in comparative studies, used as a baseline, or mentioned in methodology sections across diverse research fields.

3. VIKOR: High academic mention frequency, especially in the last two decades. Many papers, particularly from Asia and Eastern Europe, include VIKOR either as a primary method or in comparisons, giving it a strong presence in academic databases.

4. PROMÉTHÉE: Also frequently mentioned, especially in European journals and the OR literature. It often appears in multi-method comparisons and has a long history of being cited in MCDA research.

5. ÉLECTRE: Widely cited in older and theoretical MCDA literature. Almost any discussion of outranking methods in academia will mention ÉLECTRE, since it founded the outranking or French school of MCDA, so it continues to show up a lot in the literature, though new applications are fewer than for the above methods.

6. ANP: Frequently mentioned in academic work (often alongside AHP). Researchers discuss ANP when interdependence is an issue, so it has a solid mention count in fields such as supply chain, project selection, etc.

7. SMART/SAW: While not always labelled SMART, the concept is nearly always mentioned as a baseline or in discussions of MAUT-based MCDA. Academic papers often reference the idea of a simple additive weighting method (SAW) as a comparison point. This means the family has a consistent, if somewhat background, presence.

8. BWM: Rapidly growing academic mentions. Since 2015, BWM has been mentioned in an expanding array of papers. It is not uncommon to read that BWM was used to determine weights in recent decision-making studies, increasing its presence.

9. MACBETH: It is mentioned less frequently. It might be cited in discussions of value functions.

10. Others (COPRAS, etc.): These are even less frequently mentioned. COPRAS/WASPAS appear in niche engineering decision papers. PAPRIKA is seldom mentioned by name in journal articles.

4.3. Academic Papers Published (Publication Count)

The ranking list by publication count aligns closely with the “academic mentions” ranking above. Notably, AHP, TOPSIS, and VIKOR form the top triad in publication volume over the past decades. Section 4.3 is based on the recent study by Basílio et al. [10].

1. AHP: Stands at the top with the largest number of dedicated papers. A recent count found 6835 papers in scholarly databases involving AHP, the highest of any single MCDA method. This includes methodological papers and a vast array of applications over 40+ years.

2. TOPSIS: Over 4900 papers in WoS/Scopus involve TOPSIS. This is the second-highest count, confirming TOPSIS as a favourite method in research publications.

3. VIKOR: Around 1500 papers. Notably, VIKOR’s count, despite being introduced later, surpasses many older methods, reflecting its popularity especially from 2000 onward.

4. PROMÉTHÉE: Approximately 1400 papers. PROMÉTHÉE’s publication count is close to VIKOR’s, indicating a substantial body of research using or studying the method.

5. ANP: About 1300 papers. ANP has a substantial publication count, underscoring that many studies required network-based decision modelling.

6. ÉLECTRE: Over 1000 papers. A significant count, given ÉLECTRE’s many variants; it is still a foundational outranking method in many studies.

7. SMART/SAW: A total of 400 papers explicitly mention using SAW. This number is modest because many use it without calling it by name, making a better estimate hard to obtain.

8. TODIM: A total of 300+ papers.

9. COPRAS: Slightly fewer than 300 papers.

10. Others: Includes BWM and WASPAS, among others, with smaller counts.

4.4. Total Academic Citations (Influence via Citations)

1. AHP: This is the most-cited MCDA methodology. Saaty’s original works have accrued enormous citations. The 1980 book alone has almost 60,000 citations on Google Scholar. Summing across all AHP-related papers, the citation count runs into several hundred thousand. This reflects both AHP’s first-mover status and extensive use.

2. TOPSIS: This also has a high citation impact. TOPSIS has on the order of 150,000 citations in total across over 7000 publications, which underscores a large influence. The initial book on TOPSIS by Hwang and Yoon [30] has over 3000 citations.

3. PROMÉTHÉE: Its key publications have several thousand citations combined. While an exact total is hard to estimate because citations are spread over multiple variant papers, PROMÉTHÉE is the highest cited in the French outranking category.

4. ÉLECTRE: Although less cited than PROMÉTHÉE, it still has a high count due to its long-standing position in the field and its legacy as the grandparent of all outranking methods, as well as an intellectual parent of both TOPSIS and VIKOR.

5. VIKOR: The main paper by Opricovic and Tzeng [15] is well-cited. Overall, VIKOR’s citation body has grown rapidly since its introduction, but its later start means that its cumulative citation total remains below that of older methods such as AHP, TOPSIS, PROMÉTHÉE and ÉLECTRE.

6. ANP: Saaty’s ANP (1996) book and papers are highly cited (thousands of citations), given that it is often referenced alongside AHP. The influence of ANP in citations is strong. Although far fewer people apply it compared to AHP, many still cite it in methodology discussions.

7. SMART/SAW: The citations here are diffuse. The original SMART paper by Edwards [16] and related works have been cited frequently in the decision analysis literature, but far from the level of AHP or TOPSIS. Still, because virtually every MCDA textbook or review mentions additive weighting in some form, the cumulative citation footprint is significant but spread across many sources.

8. BWM: Remarkably high impact for a new method—the 2015 BWM main paper has over 4000 citations [19]. This indicates that in just a decade, BWM’s citation count has exceeded those of many older methods’ seminal papers. BWM’s overall citations (including follow-up papers) are climbing quickly, making it one of the most influential newer MCDA brands.

9. MACBETH: Has a moderate citation count. Key papers by Bana e Costa in the mid-1990s and early 2000s have several hundred citations each. Overall, MACBETH’s influence is respectable but not in the very top tier.

10. Others: The PAPRIKA method paper [18] has on the order of a hundred citations. However, many studies using PAPRIKA do cite the 1000minds tool. If those are counted, the number would reach several hundred citations in total, which is still modest compared to methods higher on the list. Other remaining methods including COPRAS, WASPAS, and TODIM have lower citation totals, generally reflecting their niche usage.

4.5. Academic Use Cases (Application Studies in Literature)

1. AHP: It has been applied in an unparalleled number of academic case studies. From site selection problems to medical decision-making, thousands of academic papers report using AHP as the main technique. Its versatility means it appears in case studies across all disciplines. AHP has, by far, the most applications in the literature and has been around for almost the longest (48 years in 2025).

2. TOPSIS: It is the second-most common choice for applied studies. Hundreds of academic case studies (in engineering, management, environmental science, etc.) utilise TOPSIS to rank alternatives. It is especially prevalent in applied engineering journals due to its computational ease. For some reason, TOPSIS is the method most used in computer science-related MCDA applications.

3. PROMÉTHÉE: This method has a very large number of academic use cases. Reviews of usage show a large number of papers related to environmental and sustainability topics [31]. PROMÉTHÉE’s many variants allow it to be adapted to different contexts, and numerous studies report its use for things like watershed management, supplier ranking, etc.

4. VIKOR: It has become a go-to method in many applied papers since the 2000s. Academic use cases include disaster management, supplier selection, and policy evaluation, among others. Its share of use in academic case studies in one domain-specific review (COVID-19 decisions) was third after AHP and TOPSIS [30], constituting an example of VIKOR quite frequently featuring in recent case-study papers.

5. ÉLECTRE: Over the decades, many academic case studies, especially in the 1970s–1990s, employed ÉLECTRE for public and urban planning decisions. Although its frequency in recent case studies is lower, it still appears in some modern applications like water resource planning or transportation project ranking.

6. ANP: While significantly fewer in number than AHP case studies, ANP use cases are well-documented in the academic literature, often in high-impact areas like supply chain risk where criteria influence each other.

7. BWM: Since 2015, a growing number of academic case studies report using BWM for criteria weighting in real problems (e.g., evaluating sustainable suppliers, selecting logistics strategies). The number of distinct use-case papers is rapidly increasing each year, reflecting researchers’ enthusiasm for applying BWM. It is becoming common in journals related to operations and sustainability to see BWM in the methodology.

8. MACBETH: There are several notable academic case studies using MACBETH (e.g., evaluating infrastructure projects, defining strategic priorities for organisations). The volume is not high compared to AHP or TOPSIS, but the use cases are well-documented and often involve decision-makers in the process, owing to MACBETH’s interactive nature.

9. PAPRIKA: Many academic application papers (especially in health economics and prioritisation research) use the PAPRIKA method via 1000minds. For example, studies have used it to prioritise healthcare interventions, rank policy options, or conduct conjoint analysis surveys. The company claims that 370+ research publications are using it, which, while being a mere fraction of AHP’s total, is still substantial. These use cases are often in specialised journals or conference proceedings.

10. Others (COPRAS, WASPAS, TODIM, etc.): These methods collectively account for a fair number of academic case studies, but each individually has a limited set. COPRAS and WASPAS, for instance, appear in some construction engineering case studies. TODIM is used in a handful of Brazilian case examples (like selecting investments with a risk attitude element). SWARA is sometimes reported as the weighting method in case studies, alongside another ranking method.

4.6. Industrial (Real-World) Use Cases

1. AHP: This has a large number of documented real-world applications across business, government, engineering, healthcare, and education, supported both by peer-reviewed case studies and by commercial decision-support software. Real industrial use cases abound: companies use AHP for vendor selection, resource allocation, strategic planning, etc., and public agencies use it for policy and infrastructure decisions. Its popularity is due to its seemingly logical structure and the availability of user-friendly software. Ubiquitously, AHP has been used by everything from governments for project prioritisation to tech companies for product feature decisions, and by military and healthcare institutions for complex choices. It is often stated that AHP is used “in many fields including business, government, engineering, healthcare and education,” which is demonstrated by practical examples.

2. Weighted Scoring (SMART/SAW): In practice, simple weighted scoring (sometimes without a formal name) is extremely common. Many “MCDA” analyses in industry are done via spreadsheets where criteria are weighted and scored—essentially the SAW approach. This method is inherent in numerous commercial decision processes (from hiring matrices to product feature prioritisation in tech companies). Because it requires no specialised software or training, it is the default in many organisations. Thus, while one might not hear the acronym SMART in boardrooms, the method itself is arguably the second-most practiced in industry (after or on par with AHP) in various forms.

3. PAPRIKA: This has a stronger practical footprint than its academic publication count alone suggests, largely through the 1000minds software platform. However, because part of this evidence comes from software-provider documentation, these claims should be interpreted as provider-reported adoption evidence rather than independent bibliometric evidence.

4. PROMÉTHÉE: It has seen real-world use, though more limited to certain sectors. There are software implementations (Visual PROMÉTHÉE) that have been applied in consulting projects for logistics, environmental management, and finance. Some European utility companies and government bodies have used PROMÉTHÉE for project ranking and policy analysis. Its use is notable in cases where decision or operational research specialists are involved. Still, compared to AHP, its industry penetration is lower and often regional (French-speaking countries).

5. TOPSIS: Many industries use techniques akin to TOPSIS without explicitly labelling them. However, there are documented instances of it in practice—e.g., manufacturing firms using it to rank supplier performance or telecom companies applying it for site selection. Also, some decision support systems include it as an option. Its straightforward nature means some practitioners implement it ad hoc for multi-criteria rankings (especially in countries where engineers are familiar with it from university). Overall, its direct industrial use is moderate but present.

6. MACBETH: This has been used in some organisational decision processes, particularly facilitated by consultants/academics trained in the method. For example, it has been applied in municipalities for budget decisions and by defence agencies for evaluating policy options (the method’s developers have themselves reported such cases in the literature). While not mainstream, it has a track record of real decision support, especially in group settings requiring consensus.

7. ÉLECTRE: Historically, ÉLECTRE was used in governmental decision processes in France and the EU (such as water management and transportation planning in the 1970s–80s). In modern times, it is less commonly employed in the real world. Those who need outranking often opt for PROMÉTHÉE or other more consistent methods instead. Still, some consulting cases in environmental management have used ÉLECTRE, and some defence and aerospace evaluations mention using it for ranking alternatives when a purely outranking approach is desired. Its current industrial use is limited and mostly within expert and specialist circles.

8. ANP: A few specialised cases of ANP in industry exist. For instance, Boeing reportedly used ANP for prioritising engineering design requirements, and some oil and gas companies have used it for project selection. These cases are driven by collaboration with MCDA experts. ANP’s complexity has prevented widespread casual use, but in high-stakes, interrelated decisions, some organisations have invested in ANP analyses.

9. BWM: Being new, pure BWM use in industry is still emerging. However, early signs show that some consulting firms and forward-thinking companies are experimenting with it because of its efficiency in eliciting weights. As more user-friendly implementations appear, BWM is expected to gain traction. Currently, most real use cases are in pilot projects or decisions guided by researchers (e.g., a company working with a university to apply it to a purchasing decision).

10. Others: Most other methods (COPRAS, WASPAS, TODIM, etc.) have smaller footprints in everyday societal use. They are largely confined to academia or a small cluster of organisations where the method was developed or explicitly introduced.

5. Discussion: Branding, Visibility, and Adoption Mechanisms

5.1. Branding and Adoption Mechanisms

The branding interpretation should be understood as exploratory. The present study does not directly measure cognitive bias among researchers or practitioners; rather, it identifies visibility and adoption patterns that are consistent with mechanisms such as availability, familiarity, software-supported diffusion, cumulative citation effects, innovation diffusion and academic citation dynamics. The six visibility and adoption indicators therefore point to complementary mechanisms that may influence how methods become known, cited, taught, implemented and used in practice. Throughout this discussion, the ranking is interpreted as evidence of method visibility and adoption, not as evidence of methodological performance. A highly ranked method is therefore not necessarily theoretically superior or more appropriate for a particular decision context; it is more visible, more frequently cited, more widely reported, or more strongly represented across the indicators used in this study.

First, several highly ranked methods benefit from strong name recognition. Acronyms such as AHP, TOPSIS, VIKOR, and PROMÉTHÉE are memorable and easily transferable across articles, teaching materials, software packages, and applied case studies. This does not make them methodologically superior, but it may increase their availability in the minds of researchers and practitioners selecting a method. Second, first-mover and cumulative citation effects appear to matter. AHP, ÉLECTRE, and PROMÉTHÉE entered the field early and became embedded in textbooks, course syllabi, review papers, and software ecosystems. Such early visibility may reinforce later adoption independently of any fresh comparative evaluation. Third, software support appears to translate methodological visibility into practical uptake. AHP, PROMÉTHÉE, MACBETH, and PAPRIKA/1000minds illustrate different ways in which software can stabilise a method as a practical decision-support option.

Fourth, the comparison between academic and industrial visibility shows that publication volume and practical use are not identical forms of popularity. Weighted scoring and SMART/SAW-like approaches are less visible as named academic brands but are highly visible in practice because they are easily implemented in spreadsheets and organisational scorecards. Conversely, methods such as ANP, MACBETH, COPRAS, and TODIM may have identifiable academic communities but more limited general organisational reach. This divergence supports the central distinction made in this article between methodological performance, academic visibility, and practical adoption.

The branding interpretation should therefore not be read as a claim that researchers and practitioners choose MCDA methods irrationally. Rather, it suggests that method selection is partly shaped by ordinary dissemination mechanisms: memorable labels, pedagogical simplicity, software availability, citation visibility, and institutional familiarity. These mechanisms are not illegitimate; they help methods spread and become usable. The concern arises only when visibility is mistaken for methodological adequacy, or when a method is selected because it is familiar rather than because it fits the decision context.

High citation rates should not be interpreted as direct evidence of theoretical superiority. Citations may reflect methodological usefulness, but they may also reflect age, textbook adoption, recognisable terminology, software availability, ease of teaching, and repeated use as a benchmark in comparative studies. For example, older methods have had more time to accumulate citations, and methods with distinctive acronyms are easier to search for, teach, and reuse as named methodological brands. Conversely, generic weighted-additive approaches may be widely used in practice without accumulating comparable method-specific citation counts, because they are often implemented as ordinary scorecards or spreadsheet models rather than cited as a named MCDA brand. The present study therefore treats citation impact as one indicator of visibility and influence, not as a proxy for theoretical validity.

A further implication of the ranking framework concerns newer MCDA methods whose visibility is increasing rapidly. A method may improve its position in popularity rankings because it is theoretically useful, but also because it benefits from novelty effects, active dissemination, software support, ease of use in application papers, or frequent inclusion in hybrid MCDA models. This creates a risk that rising publication counts or citation rates are interpreted as evidence of methodological superiority before the method has been sufficiently examined with respect to axiomatic properties, robustness, transparency, elicitation burden, and decision-context fit. Newer methods such as BWM and software-supported methods such as PAPRIKA illustrate how rapid academic or practical visibility can emerge through different channels. This is not a criticism of these methods; rather, it highlights the need to distinguish between methodological innovation, dissemination success, and demonstrated decision-analytic adequacy.

5.2. Limitations of the Ranking Framework

Several limitations of the proposed ranking framework should be noted. First, the ranking is necessarily time-dependent. Publication counts, citation counts, web visibility, software uptake, and documented use cases change continuously. Older methods such as AHP, ÉLECTRE, PROMÉTHÉE, and TOPSIS benefit from longer citation histories and greater textbook presence, whereas newer methods such as BWM may show rapid recent growth without yet having comparable cumulative totals. Conversely, some older methods may retain high cumulative visibility even if their relative current use is declining.

Second, the six indicators are affected by different data-source biases. Bibliographic databases such as Scopus, Web of Science, and Google Scholar differ in journal coverage, conference coverage, language coverage, duplicate handling, and citation indexing. Search engines introduce additional instability because results depend on indexing practices, localisation, query formulation, spelling variants, and temporal fluctuations. For this reason, web visibility was treated as an ordinal indicator rather than as a precise numerical count.

Third, several MCDA method names and acronyms are ambiguous. Terms such as AHP, ANP, SMART, SAW, and weighted scoring can generate false positives outside MCDA unless combined with full method names or decision-analysis context terms. Conversely, some methods are used without being named explicitly, especially simple weighted-sum or scorecard approaches. This may underestimate the practical prevalence of generic additive methods while over-representing methods with distinctive acronyms and searchable brand names.

Fourth, the evidence base differs between academic and real-world use. Academic publication and citation data are relatively visible, although still imperfect. Industrial and public-sector use is less systematically documented. Some real-world applications are reported in peer-reviewed case studies or public reports, whereas others are described only in software-provider materials, consultancy materials, or organisational examples. Such evidence was included cautiously, but it should not be interpreted as having the same evidential status as bibliometric data.

Fifth, the consolidation of method variants into broader method families involves judgement. Grouping variants under AHP, ÉLECTRE, PROMÉTHÉE, SMART/SAW, or related families improves readability and avoids fragmenting the ranking, but it may also hide differences between individual variants, extensions, and software implementations. Fuzzy, stochastic, grey, interval, and hybrid extensions further complicate attribution because they may spread a method name while also changing the actual computational procedure being used.

Finally, the ranking captures visibility and adoption rather than methodological quality. A high rank does not imply that a method is theoretically superior, normatively more defensible, or more suitable for a given decision problem. Likewise, a lower rank does not imply inferior methodological value. The ranking should therefore be read as a structured map of prominence, dissemination, and documented use. Its purpose is to support reflection on method choice, especially where non-expert users may otherwise rely on familiarity, availability, or software presence as proxies for decision-context fit.

The final ranking is sensitive to the choice and weighting of evaluation criteria. If publication counts and citation impact are emphasised, older and heavily cited academic methods, such as AHP, TOPSIS, PROMÉTHÉE, ÉLECTRE, and ANP, are favoured. If industrial or public-sector use is emphasised, simpler and software-supported approaches such as weighted scoring/SMART/SAW, AHP, and PAPRIKA gain relative importance. If recent growth is emphasised, newer methods such as BWM would rank higher than their cumulative publication or citation totals alone suggest. Thus, the composite ranking should not be interpreted as an invariant ordering. It reflects the balanced six-criterion framework used in this study. Different but defensible weighting schemes would shift some mid-ranked methods, although AHP remains dominant across all six perspectives.

The industrial and public-sector use indicator is the least homogeneous of the six evidence categories. Unlike publication counts or citation records, real-world adoption is unevenly documented across methods. Some applications are reported in peer-reviewed literature or public-sector documents, while others are available mainly through software-provider or consultancy materials. Consequently, the industrial ranking should be interpreted as an ordinal map of documented real-world visibility rather than as a precise measure of total adoption.

5.3. Implications for Teaching, Software Development, and Practitioner Training

The findings have practical implications for how MCDA methods are taught, implemented, and communicated to non-expert users. In MCDA teaching, popular methods should not be presented merely as a list of competing techniques. Instead, teaching should make explicit the distinction between method visibility, methodological assumptions, theoretical properties, elicitation requirements, and decision-context fit. Students should be encouraged to ask why a method is commonly used, what assumptions it imposes, and whether those assumptions are appropriate for the decision problem at hand.

For software development, the results suggest that decision-support tools should avoid presenting method availability as an implicit recommendation. Software interfaces can encourage better method choice by explaining the assumptions behind each method, warning users about common misuse, supporting sensitivity analysis, and documenting how changes in weights, scales, normalisation, or thresholds affect the results. This is particularly important for non-expert users, who may otherwise interpret a branded method or software-supported procedure as methodologically authoritative.

For practitioner training, the central recommendation is to treat MCDA method choice as a modelling decision rather than as a default selection from a familiar toolbox. Training should therefore include simple diagnostic questions: Are trade-offs meaningful in this decision context? Are criteria compensatory or non-compensatory? Are weights elicited, assumed, or imported from elsewhere? Are rankings stable under reasonable perturbations? Is the method being chosen because it fits the problem, or because it is familiar, well marketed, or easy to implement? Such questions can help practitioners distinguish appropriate use of popular methods from popularity-driven method selection.

6. Conclusions

This article examines the popularity and branding dynamics of widely known Multi-Criteria Decision Analysis (MCDA) methods across academic and industrial domains. Drawing on six evaluative criteria—web presence, academic mentions, publication volume, citation impact, academic applications, and real-world use—we have developed a composite ranking of these methods. These findings should be interpreted with the limitations of the evidence base in mind: the ranking is sensitive to database coverage, search strategy, temporal effects, and the uneven visibility of academic and real-world applications.

The aim of this analysis is not to criticise popular MCDA methods for being widely used. They are valuable, well-established, and practically useful. Rather, the concern is that non-expert or semi-expert users may sometimes treat visibility, familiarity, software availability, and memorable branding as proxies for methodological suitability. In such contexts, popularity can exert a disproportionate influence on model choice. The purpose of the present ranking is therefore to make this influence visible and to encourage a more reflective distinction between method prominence, methodological quality, and decision-context fit.

The results underscore that AHP remains the dominant MCDA method across every dimension assessed. Its simplicity and strong first-mover advantage have made it the dominant standard across both academia and industry. TOPSIS and PROMÉTHÉE follow thereafter, benefitting from conceptual clarity and effective positioning within the engineering and operations research communities. VIKOR has carved out a niche with its compromise programming logic, while ÉLECTRE, ANP and weighted scoring variants maintain steady, if more specialised, usage. Emerging methods like BWM and tools like 1000minds (PAPRIKA) show how targeted innovation and software-driven outreach can elevate newer approaches into practical relevance. These are “newer” in the sense that all top-five methods are at least 40 years old or more.

However, the article has also highlighted a possible tension in MCDA’s historical development: the risk that branding and visibility may be conflated with methodological rigour. Several methods benefit from their acronyms or alignment with influential academic institutions, which is consistent with possible brand-based lock-in and network effects. This presents a paradox within decision science: while MCDA methods are designed to help mitigate cognitive biases in decision-making, their own adoption by researchers and practitioners may partly be associated with availability, familiarity, and affective association mechanisms: some of the very biases MCDA use seeks to overcome.

Another central insight is the multi-layered nature of method adoption. Academic prevalence does not always correlate with real-world use, and vice versa. For example, weighted scoring methods are rarely the focus of academic publications, yet they are common in real-world use. Conversely, methods like MACBETH or ANP have strong academic appeal but limited industrial uptake, perhaps due to low industrial market reach or lack of accessible implementations. This distinction has practical implications: MCDA education and dissemination should consider not only what is theoretically elegant or academically prolific but also what is suitable and easy to use for varied decision contexts.

Finally, a practical implication is that MCDA education, software interfaces, and practitioner training should not treat widely used methods as default choices. They should instead help users distinguish popularity from suitability by making method assumptions, elicitation demands, sensitivity, and decision-context fit explicit.