Using a Capability Map for Application Portfolio Rationalization

Application portfolio rationalization decides which applications to keep, invest in, replace or retire. Done application by application, it turns into a debate about each system's merits. Done against a business capability map, it becomes a question the map can answer: for each thing the business does, how many applications support it, how well, and how much does that capability matter?

The capability map gives the review a frame that holds still. Applications come and go and the organization reorganizes, but the capabilities stay the same, so overlaps, gaps and dead weight show up as patterns on one page.

The Method, in Six Steps

  1. Fix the capability frame. Agree the capability map and the level you will map at. L2 suits most portfolios; use L3 where applications are narrow and several of them sit inside one L2 capability.
  2. Inventory the applications. List the applications and systems in scope, with an owner for each. Include spreadsheets and other tools the business runs on, because they are often where the gaps are being filled.
  3. Map each application's footprint. For every application, score its footprint on each capability: high, medium or low. A high footprint means the capability depends on it; a low one means it touches the capability at the edges.
  4. Read the matrix. Capabilities with several high-footprint applications are overlap candidates. Capabilities with no application are gaps, or work done by hand. Applications that map to no capability need a second look: either they are dead weight, or the map is missing a capability.
  5. Add value and condition. Overlay how much each capability matters to the strategy, and how well each application serves it today: its fitness for the work, its cost and its technical health. Overlap in a low-value capability is a consolidation candidate; weak support for a high-value one is an investment case.
  6. Decide and sequence. Assign each application one of four dispositions: keep, invest, replace or retire. Before retiring one, check which data it creates and updates, and which capabilities depend on that data, so nothing is orphaned. Sequence the changes so each capability keeps working throughout.

Where AI Changes the Review

If AI agents will take on work in some capabilities, rationalize with that in mind. A capability whose target AI mode is automate or agentic needs clean, accessible data and fewer systems in the way, so consolidating applications there pays twice. A capability that stays human-led needs no change on AI's account.

Each Capstera AI-ready model carries a target AI mode and a data readiness rating for every capability at the profile level, so the two reviews can use one frame.

What the Packages Provide

Every Capstera AI-ready Capability Model and Blueprint includes a Capabilities Relationship Mapping Templates workbook, pre-filled with the package's L2 capabilities, with blank rows for your own. Its sheets include:

  • Capabilities to applications and systems: name each application in the header row and score its footprint on each capability as high, medium or low.
  • Capabilities to data entities: which capabilities create, update or read each data entity, the check to run before retiring an application.
  • Capabilities to strategy, functions, roles, locations and vendors, for the value overlay and the follow-on decisions.

Doing It in the Capstera Platform

The same mapping can live in the Capstera Platform, which links capability maps, value streams and business data models and includes cross-relationship mapping and heatmaps. Each package's import files load its capability model into the platform.

Questions

Why rationalize applications against capabilities rather than processes?

Capabilities stay stable when processes, systems and the organization chart change, so the review doesn't need redoing after the next reorganization. They also put every application on one shared frame, which is what makes overlaps and gaps visible.

What level of the capability map should applications be mapped to?

L2 for most portfolios. Map to L3 where applications are narrow, so several of them sit inside one L2 capability and the L2 view would hide which does what.

What does an application that maps to no capability mean?

Either the application does nothing the business needs, which makes it a retirement candidate, or it supports work the capability map is missing, which means the map needs a capability added. Check which before deciding.

Can the capability model be used with an enterprise architecture tool?

Yes. The workbook keeps the capabilities with stable IDs on their own sheets, ready to import into an enterprise architecture tool, and the packages include import files for the Capstera Platform.