Independent educational resource. Not affiliated with CAST, SonarSource, or any code-analysis vendor. Data sourced from CISQ, Stripe, McKinsey, and DORA reports.

Section IV, Assessment

Score your codebase across eight dimensions.

A weighted scorecard that turns intuition into a defensible baseline. Score each dimension 1 to 5 against the rubric, see your weighted grade, identify the dimension dragging the score down, and export results for tickets or presentations.

Score each dimension 1 to 5 against the rubric. Weighted average produces the grade.
01 / 08

Architecture

Weight 20%

Clear bounded contexts, some seam coupling between modules

3
02 / 08

Code Quality

Weight 15%

Reasonable structure, code reviews catch most issues

3
03 / 08

Test Coverage

Weight 15%

30 to 60% coverage, critical paths tested, some flakiness

3
04 / 08

Dependency Health

Weight 10%

Most dependencies current, monthly review cadence

3
05 / 08

Documentation

Weight 10%

Architecture and onboarding docs, runbooks for top incidents

3
06 / 08

Infrastructure & CI/CD

Weight 10%

Pipeline based deploy, environments mostly aligned

3
07 / 08

Security Posture

Weight 10%

SAST in CI, secret scanning enabled, auth patterns documented

3
08 / 08

Observability

Weight 10%

Metrics on golden signals, structured logs, some tracing

3
3.00
of 5.00
C
Manageable
Architecture
Score 3 / 5, prioritise first

All scoring runs in your browser. No data leaves the page.

Section II

What each dimension covers.

The rubric in the scorecard above gives you the level descriptions. This table shows how the eight dimensions map to your day to day operational categories.

DimensionWeightFocus
Architecture20%Module boundaries, coupling, evolvability
Code Quality15%Duplication, complexity, naming, function length
Test Coverage15%Unit, integration, e2e, critical path coverage
Dependency Health10%Currency, vulnerability count, upgrade readiness
Documentation10%API docs, ADRs, runbooks, onboarding speed
Infrastructure & CI/CD10%Deploy automation, environment parity, rollback
Security Posture10%Scanning, secrets, auth, threat modelling
Observability10%Metrics, tracing, alerting, SLOs
Section III

Reading the grade.

The weighted score maps to a five band grade. Each grade has a recommended action.

A
4.5 to 5.0
Best in class

Maintain. Allocate minimal debt time.

B
3.5 to 4.4
Healthy

Boy Scout rule. Quarterly retrospective.

C
2.5 to 3.4
Manageable

Allocate 20% sprint capacity to debt.

D
1.5 to 2.4
Concerning

Dedicated debt sprints. Set ceiling.

F
Below 1.5
Critical

Pause new features. Triage by risk.

Section IV

Common questions.

01Why eight dimensions instead of one number?+

A single TDR or SQALE rating tells you the magnitude of the problem but not where it lives. The eight dimensions on this page map to the operational categories that engineering leaders actually budget against. Knowing your codebase scores 2.3 overall is less useful than knowing your test coverage is a 1 and your architecture is a 4. The weighted total gives you the grade. The dimension breakdown tells you what to fix first.

02How are the weights set?+

Weights reflect typical business impact. Architecture is highest (20%) because architectural debt has the largest blast radius and is the hardest to remediate incrementally. Code quality and test coverage are 15% each because they affect every change. The remaining five dimensions sit at 10% each. Override the weights for your context if security or observability matters more than average for your business.

03How does this compare to SQALE or TDR?+

SQALE and TDR measure code-level debt, scoped to what static analysis can see. This assessment is broader: it captures architectural, process, and operational debt that static analysis misses. Best practice is to use both. SQALE or TDR for the precise code-quality figure. This assessment for the full picture you present to leadership.

04Should I score honestly or aspirationally?+

Honestly. The point of the assessment is to identify your weakest dimension so you can prioritise. Inflating scores to look good defeats the purpose. If you are unsure between two scores, pick the lower one. The conservative version gives you a more defensible baseline to track improvement against.

05Can I run this for multiple systems?+

Yes, and at portfolio scale this becomes the input to enterprise prioritisation. Run the assessment per system, capture the dimension scores, and feed them into the portfolio matrix on the enterprise page. The combination produces a system level prioritisation list with action recommendations.