Backlog health is the ability to make a credible cross-ART decision
The Solution Train backlog is a Kanban system for capabilities and enablers that enhance a large solution and its architectural runway. A long list of refined capability statements is not health. The backlog must expose outcome evidence, integration risk, NFRs, supplier conditions, ART capacity and the decisions needed before work flows.
Inspect six health dimensions
| Dimension | Healthy evidence |
|---|---|
| Relevance | Capability still advances Solution Vision and customer or mission outcome |
| Age | Time in state has an owner and decision consequence |
| Readiness | Benefit, acceptance, NFRs and major dependencies support planning |
| Flow | WIP fits analysis, ART and integration capacity |
| Architecture | Runway, interfaces and Solution Intent are sufficient |
| Validation | Integrated and operational evidence can be produced within the PI |
Age by state, not one creation date
An old funnel item may need removal; an old analyzing item may reveal missing authority; an old implementing capability may expose batch size or integration failure. Set state-specific thresholds that trigger revalidation, splitting, escalation, supplier action or deletion. Ageing is a prompt for action, not an automatic performance judgment.
Readiness should preserve known uncertainty
- Name the outcome and benefit hypothesis.
- Split capability into coherent ART features and enablers.
- Identify interfaces, suppliers and cross-ART needed-by dates.
- Attach persistent NFRs and context constraints.
- Define the earliest integrated evidence path.
- Record unresolved assumptions teams can safely learn during the PI.
Monthly backlog health review
- Remove obsolete or strategically displaced items.
- Review ageing and decisions waiting for authority.
- Compare WIP with integration and supplier capacity.
- Split capabilities that cannot produce PI evidence.
- Sequence enablers with the option they protect.
- Check released capability benefits against hypotheses.
SAFe POPM training helps connect capability intent to ART features. SAFe RTE training supports solution-level flow, dependencies and integrated planning.
A healthy backlog regularly loses items through delivery, merging, invalidation and stopping. If it only grows, the Solution Train is recording demand rather than making economic choices.
Worked backlog decision
A capability remains in analysis for ninety days because three ARTs disagree on the identity interface and a supplier estimate is missing. More refinement will not solve the decision. Solution Management narrows the first outcome, architecture defines a compatibility experiment, the supplier provides evidence by a fixed date, and the STE tracks the dependency. If the experiment fails, the capability returns to discovery or stops rather than occupying implementation WIP.
Avoid readiness theatre
A capability is not ready because every template field is complete. Ask whether ARTs can plan coherent feature slices, integration and NFR evidence can arrive within the PI, major supplier conditions are visible, and remaining uncertainty is safe to learn during execution. A red unknown is more useful than a green guess.
Metrics to pair with backlog age
- Capability arrival and finish rate.
- Analysis WIP and decision latency.
- Time to first integrated evidence.
- Percentage stopped or split before implementation.
- Blocked time by supplier, architecture, environment, or authority.
- Realized benefit after release.


