There is the portfolio IT designed, and the portfolio people actually consume. Most decisions are anchored to the first one.
The one we designed
Every IT organization has a service portfolio it can show you. It was designed in workshops, structured by taxonomy, and it lives in the catalog. It says what services exist, how they’re organized, and who owns them. Real work went into it — service owners argued its categories, architects drew its boundaries, and someone maintains it still.
It is also, necessarily, a simplification. A catalog is a hierarchy, and a hierarchy holds one dimension at a time. Enterprise reality has many. The catalog keeps one true dimension of every service and sets the rest aside — not through carelessness, but because that is what the structure can hold.
The one that’s consumed
There’s a second portfolio, and it doesn’t live in any document. It’s what the people IT supports actually consume, for what purposes.
The designed portfolio says the collaboration platform sits under “communications.” The consumed portfolio knows it’s also the phone system, the meeting room, the file store, and half the development workflow. The designed version has a box for everything. The consumed version crosses every box.
The consumed portfolio is the more accurate of the two, and it has always been the harder one to see. It exists as thousands of daily choices — which tool for which task, which workaround for which gap. No one person holds it. Everyone holds a piece.
Where the two diverge
The divergence isn’t a maintenance gap that better catalog hygiene would close. It’s structural. The two portfolios are different shapes. The designed one is organized around ownership — who runs what, who funds what — because the people who maintain the catalog are the people accountable for the services. The consumed one is organized around capability — what a service does for the person using it. Both organizing logics are correct. A tree can encode one of them.
So the portfolios drift apart along a predictable line: everything cross-cutting, multi-purpose, or informally adopted lives fully in the consumed portfolio and only partially — or not at all — in the designed one. The services people lean on hardest are often precisely the ones that fit the catalog least, because heavy real-world use is what makes a service outgrow its box.
What anchors to which
Every significant IT decision is anchored to one of the two portfolios — and it’s almost always the designed one, because the designed one is the only one anyone can see.
Rationalization consolidates the services the catalog lists, in the categories the catalog draws. Investment flows along the catalog’s structure. Vendor scope is negotiated against the catalog’s definitions. Chargeback allocates cost by the catalog’s boundaries. Each of these decisions is made carefully, by capable people, on the most complete picture available. The picture is one dimension of a many-dimensional reality.
Decisions anchored to the consumed portfolio would sometimes land differently — consolidating along how services are actually used rather than how they’re classified, funding the capabilities people depend on rather than the categories that contain them. Not because the designed-anchored decisions were made badly, but because the anchor was partial.
Seeing the consumed portfolio
For most of enterprise IT’s history, this was simply the condition of the work. The consumed portfolio couldn’t be gathered — it was distributed across too many people, too many dimensions, too much unstructured evidence. So decisions anchored to the portfolio that could be written down, and the gap between the two was carried as an unavoidable cost.
That condition is lifting. The consumed portfolio leaves evidence — in what people say about their tools, in what they raise and what they never raise, in patterns of use that cross every catalog boundary. Evidence of that kind was beyond the reach of hierarchical methods. It is exactly what inference-based analysis was built for.
This is the deeper role of ground truth. It doesn’t just measure how services perform — it reveals what the portfolio actually is. And a portfolio assembled from evidence is only as good as the evidence’s independence: where it comes from, and whether the instrument that assembles it shows its own reliability. That standard has a name — structural independence — and the consumed portfolio, anchored in ground truth, is the first thing worth holding to it.
The decisions were never the problem. The anchor was. And the anchor can now be seen.
