Design Health
Score
The single org-level architectural health number. Board-reportable.
DCS × 0.40
+ (1 − ADI) × 0.25
+ IFR × 0.20
+ (1 − UDR) × 0.15
The formula above is the target design — it requires IFR and UDR, neither of which is built yet, and DCS/ADI require a completed CDA audit history. Until all four component metrics are live, DHS is computed from a different, fully-available set of inputs so the score isn't simply blank:
where DecisionScore = (100−DCR)×0.4 + AAR×0.4 + (100−ADTR)×0.2
and RiskScore = 100 − RES
This interim formula will be replaced by the target formula above as IFR and UDR ship and CDA audit history accumulates — the DHS number you see today is not final.
What it signals
DHS is the single org-level architectural health number. It exists to answer one question at the board level: is this system getting better or worse? It is designed to survive a board-level presentation without requiring the presenter to explain what a Lens evaluation is or how drift is calculated.
Trend matters more than point-in-time. A DHS of 72 that has risen for four consecutive sprints reflects a healthier governance posture than an 84 that has dropped 15 points in six weeks. When reading DHS, always load the time-series view — a single snapshot is context without signal.
DHS is a composite. It aggregates four component metrics — DCS, ADI, IFR, and UDR — each representing a distinct dimension of architectural health. A high composite can mask a failing component. Always review the component breakdown alongside the headline number.
How rkito produces it
The four component metrics are computed independently by the CDA pipeline, then combined with stated weights. ADI and UDR are inverted so that higher is always better across all inputs. The weighted sum is normalised to 0–100. DHS is computed per sprint, stored as a time series, and surfaced on the org dashboard.
Who this metric is for
DHS is the single number for architecture health without needing to understand the underlying metrics. It answers the board-level question — is our architecture getting better? — without requiring translation.
DHS is the output; architects own the inputs. DCS improves through better Lenses. IFR improves through better CHI process adoption. DHS is where accountability is visible.
How this metric gets misused
...report DHS without Governance Depth. A DHS of 94 with a Governance Depth (GD) of 2 means only two rules are being checked — almost nothing in the architecture is being evaluated. That score is not evidence of a healthy system; it is evidence of an under-governed one. Always present DHS alongside GD. “Our DHS is 94” is a misleading statement if GD is low.
GD threshold to trust DHS: At least 10 active Lenses. Below that, DHS is measuring coverage of a tiny fraction of the codebase's design surface.
...use a momentarily high DHS to defer investigation of leading indicators. DHS is a lagging indicator. It reflects what happened in the trailing window, not what is about to happen. PDT (Pillar Drift Trend) and ADRIFT are leading indicators. A DHS of 88 paired with three consecutive sprints of declining PDT and rising ADRIFT signals a system that is about to deteriorate — the high DHS is yesterday's result.
What to do at each threshold
Step-by-step response playbook for each signal state.
Architecture is in good order. The focus shifts from remediation to validation that the score is meaningful.
Confirm GD is meaningful — at least 10 active Lenses. A DHS of 87 with GD=3 is not a trustworthy score. Navigate to the org dashboard to verify Governance Depth alongside DHS.
Check that components are individually healthy. A DHS of 87 driven by very high DCS but very low IFR may mask a process failure. Verify the component breakdown — each dimension should be above its own floor.
Drift is present in at least one architectural dimension. Identify and address the weakest component.
Navigate to /changecontrols/dashboard. Identify which of the four components — DCS, ADI, IFR, UDR — is pulling DHS down. The weakest component is the remediation priority.
Each component has its own recovery path. Do not attempt to improve DHS directly — DHS is the output. Address the failing input: Lenses for DCS, CHI process for IFR, resolution velocity for UDR.
The four components collectively indicate systemic governance failure. Treat as an architectural health incident.
Declare an architectural health incident. Convene engineering leadership with the component breakdown — DCS, ADI, IFR, and UDR individually — not just the composite.
Do not attempt to improve DHS directly. Improve the underlying components. A DHS this low means governance is failing at multiple dimensions simultaneously — each requires its own remediation track.
The highest-weighted DHS input. % of PRs that passed all Lens checks — the headline conformance signal.
Number of active Lenses covering the codebase. DHS is only trustworthy when GD is meaningful.
Proportion of flagged violations that remain open. Inverted and weighted at 0.15 in the DHS composite.
What is your org's DHS?
rkito computes your Design Health Score automatically. Connect your repo and see the composite — and every component — within your first sprint.