Context-Aware BPM: Method Assessment and Selection (CAMAS)

Summary

Vom Brocke et al. (2021) address a gap in context-aware BPM: while prior work has catalogued context factors (Rosemann & vom Brocke 2015; vom Brocke et al. 2016), there is no systematic guidance for assessing existing/new BPM methods or selecting an appropriate method for a given context. Following Design Science Research (Gregor & Hevner 2013; Peffers et al. 2008) and Situational Method Engineering (Brinkkemper 1996; Henderson-Sellers & Ralyté 2010), the authors develop the CAMAS Method as the central artefact and instantiate it as an Excel prototype with a Method Base of 103 BPM methods retrieved via SLR.

CAMAS comprises three components:

Classification Framework — a 3-D meta-model:
- Lifecycle dimension: design · implementation · monitoring · improvement & innovation · project & program management.
- Goal dimension: exploitation vs. exploration (Benner & Tushman 2003).
- Context dimension: three sub-dimensions with binary characteristics (Christenfeld 1995):
  - Process — value contribution, repetitiveness, knowledge-intensity, creativity, interdependence, variability.
  - Organization — scope, industry, size, culture, resources.
  - Environment — competitiveness, uncertainty.
Assessment Process (4 activities) — performed by a BPM method engineer for a new method, or by a method user for an unclassified method:
- A1 Identify method (literature review, definition check).
- A2 Classify lifecycle dimension (multi-stage allowed).
- A3 Classify goal dimension.
- A4 Classify context dimension using {applicable (a), not applicable (na), cannot assess (–)} per characteristic.
- Reliability via Cohen's Kappa with ≥2 independent judges.
- Output: DCS (Degree of Context Specificity) indicator distinguishing special-purpose from general-purpose methods (Eq. 1).
Selection Process (4–5 steps) — performed by a BPM method user (e.g. process manager):
- S1 Define target lifecycle stage(s).
- S2 Define target goal characteristic(s).
- S3 Define target context characteristics with weights via AHP (Saaty 1990).
- S4 Select method(s) for a single context using DA (Degree of Applicability) (Eq. 2), with risk-averse vs. risk-taking calculation modes for unassessed (–) values.
- S5 (optional) Iterate S1–S3 and select for multiple contexts.

Method components (Table 1)

A BPM method has attributes (goal orientation, systematic approach, principles orientation, repeatability) and elements (meta-model, activity, technique, tool, role, defined output).

Evaluation

Two evaluation studies (Sonnenberg & vom Brocke 2012):

Use Case 1 — Assessment: 2 co-authors classified 103 methods; 12 BPM method engineers re-assessed a sample of 20. Hit ratio ≈ 97%; Cohen's κ ≈ 0.66 (moderate-substantial agreement, Landis & Koch 1977).
Use Case 2 — Selection: 12 method engineers applied the Selection Process across 6 real-world processes (2 contexts each).

Headline findings

Most cataloged BPM methods are general-purpose; explicit context-specificity is rare.
The literature is biased toward exploitative BPM (incremental improvement); explorative BPM (radical/innovation) is under-served by available methods.
Calls for stronger context-awareness in future BPM method design.

Connections

Builds on the BPM lifecycle (concepts/bpm-lifecycle) of Rosemann & vom Brocke 2015 — see entities/michael-rosemann.
Builds on the BPM context framework (vom Brocke et al. 2016) — see frameworks/bpm-context-framework (referenced-not-ingested for the 2016 paper itself).
Introduces a new concept page: concepts/context-aware-bpm.
New entity: entities/jan-vom-brocke.
Complements methods/process-discovery-methods — gives a formal way to choose between Dumas's evidence/interview/workshop discovery methods given a context (e.g. start-up + low-knowledge-intensity vs. large-org + high-creativity). See also syntheses/interview-structuring-for-process-models.
DSR / SME methodological lineage (referenced-not-ingested): Gregor & Hevner 2013; Peffers et al. 2008; Brinkkemper 1996.