Analysis Artifact discovery intermediate

Technology Radar

What the organisation has decided to adopt, trial, assess or retire — a decision record about technology choice, not a fashion report.

Vendor Selection Greenfield Build Cost Optimisation Cloud-Agnostic Cross-Industry Architecture Review Board Engineering CxO

quadrantChart
  title Platform technologies
  x-axis "Retire" --> "Adopt"
  y-axis "Low usage" --> "High usage"
  quadrant-1 "Adopt — default choice"
  quadrant-2 "Hold — in use, do not extend"
  quadrant-3 "Retire — plan removal"
  quadrant-4 "Trial / Assess — bounded use"
  "Kubernetes (managed)": [0.92, 0.88]
  "PostgreSQL": [0.95, 0.93]
  "Kafka (managed)": [0.86, 0.72]
  "Terraform": [0.90, 0.80]
  "OpenTelemetry": [0.78, 0.55]
  "Service mesh": [0.55, 0.30]
  "Self-managed Elasticsearch": [0.18, 0.62]
  "Oracle DB": [0.12, 0.70]
  "Jenkins": [0.20, 0.45]
  "Nomad": [0.60, 0.08]
  "DuckDB (analytics)": [0.62, 0.12]

What it is

Technologies placed in rings by how confidently the organisation has decided they may be used: adopt is the default choice, trial is bounded and observed, assess is investigation only, hold means in use but do not extend it, and retire has a removal plan.

It is a decision artifact, not a trend report. The distinguishing test is whether anything ever moves outward — a radar with no retire ring is a wish list.

When you produce it

Establish it when the estate has enough duplication that new projects are picking technology by preference, and refresh it quarterly. Each move between rings should have a short rationale, and the movement over time is the useful history.

Who reads it

Engineers starting something new, who should be able to answer "can I use X?" without a meeting. Architects, who use it to make review scale — on-radar choices need no review, off-radar ones do. Procurement and finance, who see the consolidation intent.

What good looks like

Adopt is short. If everything is adopted, nothing is a default.
The hold and retire rings are populated, with owners and dates for retirement. That is where the cost saving is.
Each entry has a sentence of rationale and a date of last movement.
It is tied to governance: on-radar is self-service, off-radar needs an ADR and a review. Otherwise it has no force.
Organisation-specific. A published industry radar is input, not output — it knows nothing about your skills, contracts or estate.

Common mistakes

Copying a public radar, which ignores every constraint that makes yours different.
Adopt-only, so it never drives consolidation.
No consequence attached to the rings, making it advisory and therefore ignored.
Annual refresh, which is slower than the decisions it is meant to inform.