How can organisations define, rationalise, and deploy development methods in a way that practitioners actually use — and that improves continuously over time?
Methods are only valuable if they are used — yet most organisations find their development methods ignored, inconsistently applied, or actively resisted. This paper documents a real engagement at a large financial services organisation where existing methods had proliferated across wikis, SharePoint, and Office documents into a sprawling, inconsistent tangle: over 100 artefacts required on an average project, unclear rationale for most of them, poor integration between business change, project management, and development lifecycles, and a developer community pushing back against perceived administrative overhead. The investigation revealed that document-driven method definition was the root cause of most problems.
The response was a goal-driven method engineering workbench, built on the EVA platform using a comprehensive meta model that connects principles, goals, artefact types, task types, resource types, risks, quality checks, and techniques in an integrated repository. The key innovation is structuring the entire method around agreed goals: once teams align on what a project should achieve, the method content — artefacts, tasks, roles — can be automatically derived, filtered, and presented in a tailored view. A "one page portal" interface makes the result accessible to practitioners without requiring modelling expertise, delivering a project-type-specific view of the lifecycle, expected deliverables, and quality checks with drill-down navigation throughout.
The paper is candid about adoption challenges — two key sponsors departed during the engagement — and reflects honestly on the reality that method adoption is a social and organisational challenge as much as a technical one. For anyone designing or rationalising methods in a large organisation, the lessons here are as valuable as the approach itself.
Pages: 9
Originally published as an experience paper by Graham McLeod, prepared for the 24th International Conference on Evaluation and Modeling Methods for Systems Analysis and Development (EMMSAD 2019), held at CAiSE 2019, Rome, Italy, June 2019.