Why there is no portfolio score¶
Roll a portfolio report in front of anyone senior and the first question is always the same: so what's our number?
There isn't one. Not because it hasn't been built yet — because a single number would destroy the only thing the portfolio view adds.
The average erases the finding¶
The reason to look across repositories at all is that one team's cognitive read, measured against several different habitats, produces several different gaps. The spread of those gaps is the finding.
Consider two subjects:
| Subject | Habitat maturity | Cognitive read | Gap | Regime |
|---|---|---|---|---|
greenfield-api |
3.9 | 4 | +0.1 | Coherent |
legacy-batch |
1.6 | 4 | +2.4 | Ambition outpaces enablement |
Average those gaps and you get +1.25 — "mildly ambitious". That describes neither subject. It hides a team working comfortably in one place and fighting its environment in another, and it points at no action, because the two situations need opposite responses: hold the first, invest heavily in habitat for the second.
The average is not a summary of the two facts. It is the destruction of both.
A ceiling is not a mean¶
The single-repo instrument already refuses to average — the weakest discipline is the ceiling, because a specification you cannot verify is aspiration, not control. The portfolio view applies the same rule one level up, and splits the result:
- Common weak — dimensions weak across most of the estate. This is the enablement backlog. It belongs to whoever provides the habitat.
- Locally weak — dimensions weak in one or two subjects. This is that team's own backlog.
That split is doing real work. Without it, a matrix across teams becomes a league table, and the fastest way to make an honest assessment useless is to give people a reason to game it. Separating what enablement owes the estate from what a team owes itself keeps the report a diagnostic rather than a ranking.
What to say when you are asked anyway¶
Show the spread and the split ceiling. They answer the question the number was standing in for — how are we doing, and what do we do next? — and they answer it in a form somebody can act on tomorrow morning.
A grade answers neither. It just ends the conversation.