In a recent LinkedIn post, the founder of Architecture Thinking Association, Wolfgang Goebl, states that “Most IT landscapes of larger companies consist of hundreds of applications that are interconnected via poorly designed interfaces.”
Note the two issues raised i.e., “interconnected applications” and “poorly designed interfaces”.
The mess IT is in is the result, in the main, from picking solutions instead of analyzing requirements.
One remedy is to put in place a process for building a list of requirements before trying to build \ implement Enterprise Architecture for an organization.
Our group has used the following model for enterprise-wide strategy building over 30 years.
Infrastructure\Resources -> Capabilities -> Prospective Initiatives (problems or opportunities) -> Risks\Vulnerabilities -> Ranking of Initiatives -> Shortlisting (ROI/SROI, business fit) -> Funding a few Initiatives
We knew from RBV (Resource-Based View) that organizations make better decisions when staff are able to “view/see” the factors impacting decisions. To this end, we built platforms that made it easy to consolidate data for conversion to information and knowledge and we implemented one-stop-shop methods and tools at platforms that made it easy to onboard strategists and business analysts.
Here below are two screen images illustrating how not to build strategies and how to build strategies.
What we have below is a shadow ontology of some 1,000 nodes belonging to nine (9) Entities for an IED Demining Operation (Suppliers, Business Partners, Competitors, Country Profiles, Customers, Standards, Best Practices, Reference Material, Sites).
The screenshots have been blurred for confidentiality reasons . . . .
Each Entity is a shallow hierarchy with NO crosslinks between any node in one Entity that may have something in common with other nodes in other Entities and such other nodes.
The time to construct this shadow e-map is a small fraction of the time that would be needed to put in place “all” possible linkages for a multi-root hierarchy.
Full mapping for this demining application would have taken several months resulting in over 100,000 linkages, with no guarantee that analysts would succeed in putting place all possible linkages.
The secret-sauce is to let the search engine build linkages “on-the-fly. (i.e. Launch, search, reduce clutter, display)
Easy to build (the construction can be automated from ordinary delimited text “records”).
Fast to build, but not a very friendly user interface.
Here is the same content, with nodes organized under nine Entities. Some of these e-canvases have 50-100 Entities, each with 50-1000 nodes.
Let’s launch a search for ‘dog’ – we get 22 hits.
Launch a search, the search engine highlights the ‘hits’ and hides any Entities where there are no ‘hits’.
Clicking on any node results in pop-up of the underlying document, text, image or video.
Examples: (Improvised Explosive Device Detection Command and Control application)
- Build, via drag and drop, outlines comprising clones of certification standards for assignment to individual field operator for training\re-training.
- Set up Entities for various IED work sites, allocate moveable resources to the sites, track their use, ensure return to sourcing warehouses.