• Home
  • BVSSH
  • C4E
  • Playbooks
  • Frameworks
  • Good Reads
Search

What are you looking for?

Standard : Data confidence levels are visible and understood at decision time

Purpose and Strategic Importance

This standard ensures that confidence levels in data are clearly visible and well-understood at the point of decision-making. It enables teams to make faster, safer, and more informed choices by surfacing data reliability alongside the data itself.

Aligned to our "Data-Driven Decision-Making" policy, this standard strengthens trust in systems, reduces rework caused by poor-quality data, and supports a more transparent, accountable engineering culture.

Strategic Impact

  • Improved consistency and quality across teams
  • Reduced operational friction and delivery risks
  • Stronger ownership and autonomy in technical decision-making
  • More inclusive and sustainable engineering culture

Risks of Not Having This Standard

  • Slower time-to-value and increased rework
  • Accumulation of inconsistency and process debt
  • Reduced trust in engineering data, systems, or ownership
  • Loss of agility in the face of change or failure

CMMI Maturity Model

Level 1 – Initial

Category Description
People & Culture Data is trusted by default or assumed accurate.
No awareness of its reliability or limitations.
Process & Governance No defined approach for surfacing data confidence.
Decisions are made blindly or by gut feel.
Technology & Tools No metadata or indicators accompany data.
Teams rely on raw outputs without context.
Measurement & Metrics Confidence is not measured or reflected in any reporting.
Errors are discovered post-decision.

Level 2 – Managed

Category Description
People & Culture Some awareness of data reliability emerges.
Confidence is discussed informally.
Process & Governance Teams begin adding quality annotations manually,
but definitions vary.
Technology & Tools Spreadsheets or dashboards may show basic status (e.g., ‘trusted’, ‘incomplete’).
No standard approach.
Measurement & Metrics Quality signals (e.g., age or freshness) are monitored,
but not central to decision-making.

Level 3 – Defined

Category Description
People & Culture Teams understand how to interpret and act on
confidence levels presented with data.
Process & Governance Confidence scoring is standardised across sources
and reviewed periodically.
Technology & Tools Dashboards display quality indicators with each metric
(e.g., confidence bands, source trust).
Measurement & Metrics Data sets include quality dimensions like accuracy,
completeness, and latency.

Level 4 – Quantitatively Managed

Category Description
People & Culture Teams actively use confidence scores in discussions
and trade-off decisions.
Process & Governance Confidence thresholds are integrated into governance.
Low-trust data triggers review.
Technology & Tools Decision-support tools automatically factor in data
reliability at time of use.
Measurement & Metrics Metrics include confidence scores derived from
validated criteria and SLAs.

Level 5 – Optimising

Category Description
People & Culture Teams treat data reliability as a first-class concern,
continually improving literacy and trust.
Process & Governance Feedback loops refine how confidence is scored,
prioritised, and governed.
Technology & Tools Data confidence insights drive improvements in
architecture and pipeline reliability.
Measurement & Metrics Trends in trust signals are tracked to identify weak spots,
enabling proactive improvements.

Key Measures

  • Percentage of decisions made using high-confidence data
  • Coverage of confidence scores across key dashboards and tools
  • Frequency of decisions challenged or revised due to confidence issues
  • Improvement in decision lead time when confidence scores are surfaced
  • Team understanding of data quality signals through feedback or surveys
Associated Policies
  • Data-Driven Decision-Making
  • Decentralised Decision-Making
Associated Practices
  • Chaos Engineering
  • Exploratory Testing
  • Software Composition Analysis (SCA)
  • Hypothesis-Driven Development

Technical debt is like junk food - easy now, painful later.

Awesome Blogs
  • LinkedIn Engineering
  • Github Engineering
  • Uber Engineering
  • Code as Craft
  • Medium.engineering