Not ten dashboards.
One place.
Forty surfaces produce forty trickles of data that nobody reads. The control room is where signal becomes queryable: which surfaces use a given capability, which have drifted on the primary token, which are losing ground. It's the technical precondition for any governance at scale.
Six readings, one source
Each panel answers a question that stops being answerable by eye at forty surfaces.
KPIs by surface
Conversion, churn, acquisition and retention per surface — read against intent, not in isolation. A number without the decision it belongs to is trivia.
Pains & gains over time
What's working in blue, what's hurting in red, plotted over time with product events anchored to the curve. You see the release and the movement in the same frame.
Recommendations with estimated impact
The next move, with an estimated impact attached before you commit. The layer narrows the field; the call stays yours.
Running experiments
Which hypotheses are live, on which surfaces, and what they've earned so far. A variant is an alternative value on a declared property — not a fork of the project.
Identity drift
Where surfaces have diverged from the source, and whether each divergence is a declared override or silent drift. Drift is the only thing that compounds without effort.
Compliance
Accessibility and consent status per surface, emitted alongside the surface rather than audited after the fact. WCAG 2.1 AA and GDPR are outputs, not projects.
The reason the data is readable
is that the surface was born instrumented.
A dashboard wired to independently generated surfaces can only aggregate numbers that don't speak to each other. In Ekitai a touchpoint is a declared object — manifest, sections, atoms, capabilities — so data arrives already correlated to the structure that produced it. You can trace from a metric that moved to the block that moved it, and from there to the source you need to change.