An overview of process model quality literature - The Comprehensive Process Model Quality Framework

The rising interest in the construction and the quality of (business) process models resulted in an abundancy of emerged research studies and different findings about process model quality. The lack of overview and the lack of consensus hinder the development of the research field. The research objective is to collect, analyse, structure, and integrate the existing knowledge in a comprehensive framework that strives to find a balance between completeness and relevance without hindering the overview. The Systematic Literature Review methodology was applied to collect the relevant studies. Because several studies exist that each partially addresses this research objective, the review was performed at a tertiary level. Based on a critical analysis of the collected papers, a comprehensive, but structured overview of the state of the art in the field was composed. The existing academic knowledge about process model quality was carefully integrated and structured into the Comprehensive Process Model Quality Framework (CPMQF). The framework summarizes 39 quality dimensions, 21 quality metrics, 28 quality (sub)drivers, 44 (sub)driver metrics, 64 realization initiatives and 15 concrete process model purposes related to 4 types of organizational benefits, as well as the relations between all of these. This overview is thus considered to form a valuable instrument for both researchers and practitioners that are concerned about process model quality. The framework is the first to address the concept of process model quality in such a comprehensive way.

💡 Research Summary

The paper addresses the growing interest in business process modeling and the concomitant need to ensure high‑quality process models, while noting that the research field suffers from fragmented findings, inconsistent terminology, and a lack of a comprehensive overview. To remedy this, the authors conduct a tertiary systematic literature review (SLR) that builds upon existing secondary and tertiary studies rather than re‑examining primary research directly.

Methodology
Six major digital libraries (ACM Digital Library, Google Scholar, IEEE Xplore, ScienceDirect, Scopus, SpringerLink) were queried with a three‑part Boolean string: “business process model” AND “quality” AND a set of synonyms for “literature review”. Inclusion criteria required that the primary focus be the quality of business process models, that the study present conclusions or actions related to model quality, and that it be a literature review (or contain a representative review). Exclusion criteria filtered out domain‑specific studies (e.g., healthcare, project management), works focusing solely on process mining, simulation, or language issues, and studies that only examined modeling efficiency without addressing model quality. Practical limits (English language, full‑text accessibility) were also applied.

The initial search yielded 121 records; after title/abstract screening, full‑text assessment, and snowballing of references, a final corpus of 42 secondary studies was assembled. Each study was evaluated using the DARE quality checklist (C1‑C4), which assesses the explicitness of inclusion/exclusion criteria, comprehensiveness of the search, assessment of primary study quality, and adequacy of source description. Only five papers satisfied all four criteria, highlighting the overall methodological weakness of many prior reviews.

Findings from the Literature
The authors identify three dominant research streams: (1) theoretical frameworks that define and classify quality dimensions (e.g., Krogstie et al., 2006); (2) metric development that quantifies specific properties such as element count, connectivity, or label clarity (e.g., Vanderfeesten et al., 2007); and (3) practical guidelines or tool‑supported approaches aimed at improving model quality in practice (e.g., Mendling et al., 2010). However, these streams use divergent vocabularies (e.g., “understandability” vs. “comprehensibility”) and lack a unified structure, which hampers cumulative knowledge building.

The Comprehensive Process Model Quality Framework (CPMQF)
To synthesize the dispersed knowledge, the authors propose CPMQF, a multi‑layered model that interlinks quality dimensions, drivers, metrics, realization initiatives, model purposes, and organizational benefits. The framework comprises:

• 39 quality dimensions grouped into four high‑level categories (structural, visual, content‑based, user‑centric). Each dimension is further broken down into sub‑dimensions such as “element count”, “connectivity”, “label clarity”, “visual layout”, “semantic completeness”, and “stakeholder understandability”.

• 21 quality metrics that operationalize the dimensions. Metrics range from automatically extractable counts (e.g., number of nodes, cyclomatic complexity) to survey‑based assessments (e.g., perceived understandability).

• 28 quality (sub)drivers that influence dimensions positively or negatively. Examples include “use of standardized notation”, “stakeholder involvement”, “modeling training”, and “tool support”.

• 44 driver metrics that measure the presence or intensity of each driver (e.g., training completion rate, frequency of tool‑based validation runs).

• 64 realization initiatives – concrete actions that organizations can undertake to improve quality, such as “conducting modeling workshops”, “integrating automated consistency checkers”, “establishing governance policies”, and “maintaining a model repository”.

• 15 model purposes (communication, analysis, automation, compliance, etc.) that clarify why a model is created and guide the selection of relevant dimensions and drivers.

• 4 organizational benefit types (operational efficiency, cost reduction, risk management, strategic decision support) that link improved model quality to business value.

The framework is visualized as an interconnected matrix, enabling researchers to select specific dimensions, drivers, and initiatives for hypothesis testing, and allowing practitioners to map their current state, prioritize improvements, and trace expected benefits.

Limitations and Future Work
The authors acknowledge several constraints: (1) reliance on secondary literature may omit the latest primary studies (e.g., AI‑driven model validation); (2) the framework does not assign quantitative weights to the relationships among dimensions, drivers, and initiatives, which could complicate practical application; (3) some driver metrics lack tool support and would require manual data collection. Future research directions include empirical validation of metric weights, domain‑specific extensions (e.g., healthcare, manufacturing), integration with automated modeling environments, and the development of training and certification programs based on CPMQF.

Conclusion
The Comprehensive Process Model Quality Framework represents the first attempt to aggregate, structure, and interrelate the extensive but scattered body of knowledge on process model quality. By providing a unified taxonomy, a set of measurable metrics, and a catalog of actionable improvement initiatives, CPMQF bridges the gap between academic research and practical process‑modeling efforts. It equips scholars with a solid foundation for further theoretical development and offers practitioners a concrete roadmap to enhance model quality and, consequently, to realize tangible organizational benefits.