Digital transformation @scale impacts an enterprise’s operating model, enterprise architecture development processes, and solution delivery. Therefore, architects need to adopt new agile iterative practices and embrace more collaborative approaches in product development flow. The need for enterprise adaptability and speed in a fast-moving-and-growing digital economy is putting increasing time-constraint-related pressure on architects to deliver good quality architecture. In the past, they generally practiced BUFD (big up-front design), which can be very time consuming and often resulting in major delays for delivery teams.
Such an approach doesn’t lead most of the time to the optimum delivery of value, as successful products generally need to be developed iteratively through the effective participation of various stakeholders from an early stage of product idea until implementation. BUFD often relies on extensive and complex documentation that too often doesn’t fully facilitate the process of communicating the architects’ intent. Therefore, frequent meetings end up being needed in the implementation governance phase to assure conformant target product architecture.
Complicating this is the fact that the ultimate target architecture delivered can end up differing from what the projected/required one needed to be. Therefore, to communicate and to collaborate successfully with agile teams, architects need to deliver architecture deliverables/artifacts in sprints that represent MVA (minimum viable architecture), aligning product landscape descriptions to the appropriate information requirements needed for that stage of development. The level of uncertainty about the target product shrinks with new information gained during product development and delivery.
When we visualize complex architecture landscape descriptions (ref.1 below), illustrating the complex set of relations between the building blocks from various domains (represented here in different colors or cross domains dependencies), we can see the communication challenges involved. Therefore, architects need to be much more successful in their communications and collaborations with various stakeholders, including multi-skilled agile delivery teams, by scoping their deliverables more appropriately to the required MVA (minimum viable architecture) content. Simpler visuals and smaller artifacts representing only required perspectives at that junction is more likely to lead to better collaboration overall and eventually enhanced products.
The progression in product engineering from product idea to product delivery can be represented as a cone of uncertainty. In the context of enterprise strategy development and delivery, architects provide a higher initial contribution to the definition of the target product architecture (represented here as intentional architecture). However, their share of contribution toward the final product content (represented as MVAs deliverables) will increasingly shrink as agile teams start to deliver their MVPs (minimum viable products). This process is one moving from predominantly intentional architecture to emergent design.
Collaborative architecture work of the type just described above can lead to the faster delivery of successful products that best satisfy the end user. In addition, it can lead to enhanced gap analyses between baseline and target states of the product being engineered because agile teams are repeating these in every iterative cycle. Finally, it can also lead to better collaboration where there are often many dependencies across teams due to the size and complexity of products being engineered, also considering that some teams are delivering enabling building blocks needed by other teams.
Authored by Alex Wyka, Senior Consultant and Trainer, EA Principals