Example library

Preserve the original applications. Improve the outcome.

The original backlog, discovery, journey, sprint, testing, bug, and release examples are retained and constructively refined. Cross-industry examples extend the same CLEAR thinking into Product, IT, OT, public services, capability, sustainability, cybersecurity, and operations. New cases cover customer discovery, product strategy, pricing, SaaS, cloud migration, DevOps, service transformation, applied capability, industrial water use, circularity, drilling operations, and safety-critical assets.

Illustrative examples | Numbers are not industry benchmarks
Collection
Context
ProductOriginal applications revisited

Backlog management

A product group whose backlog has grown faster than its ability to make evidence-based priority decisions.

Starting statement

Ensure that 90% of backlog items are aligned with the product roadmap and strategic goals.

CLEAR outcome

Within two quarters, at least 90% of product-development capacity is allocated to work with a documented link to a current product objective and supporting evidence, compared with 58% at baseline.

Inspect the outcome record

Why the first statement fails

  • Counts aligned items rather than consequential allocation decisions
  • A large low-value backlog could still satisfy the measure
  • Does not explain how legitimate urgent work is handled

Evidence

  • Capacity allocation by objective, not only item count
  • Decision records and evidence supporting prioritisation
  • Age and disposition of unlinked backlog demand

Guardrails

  • Safety, regulatory, security, and urgent customer work can bypass normal sequencing
  • Teams do not relabel work merely to appear aligned
  • Product discovery remains able to challenge the roadmap

Candidate interventions

Backlog retirement · Evidence standards · Objective-based reviews · Capacity guardrails

Review

Monthly portfolio review; formal reassessment after two complete quarters.

Attribution limit

A documented link does not prove that the work caused the strategic result.

ProductOriginal applications revisited

Product discovery

A product organisation committing delivery capacity before testing customer need and solution assumptions.

Starting statement

Ensure that all new product ideas are validated through customer feedback and market analysis.

CLEAR outcome

Within six months, reduce the share of material product investments reversed after development because the underlying customer need was unsupported from 28% to below 15%.

Inspect the outcome record

Why the first statement fails

  • Completion of research is an activity, not evidence of a better decision
  • A 100% validation target encourages ceremonial approval
  • Feedback alone can confirm rather than challenge assumptions

Evidence

  • Decision-quality review of discovery evidence
  • Post-investment reversals traced to unsupported need assumptions
  • Customer research including disconfirming evidence

Guardrails

  • Discovery effort remains proportionate to decision risk
  • Accessibility and underserved user needs are represented
  • Teams can stop ideas without being penalised for a failed validation

Candidate interventions

Problem interviews · Prototype tests · Market analysis · Assumption mapping

Review

Review every material investment decision and aggregate learning quarterly.

Attribution limit

Market and customer conditions may change after a sound discovery decision.

ProductOriginal applications revisited

Customer journey mapping

A service with known abandonment in two high-value, cross-channel customer journeys.

Starting statement

Ensure that 100% of customer journeys are mapped and analysed for improvements.

CLEAR outcome

Within nine months, increase successful completion of the two priority customer journeys from 62% to 78%, with improvement visible across assisted and digital channels.

Inspect the outcome record

Why the first statement fails

  • Journey maps are outputs
  • Mapping every journey can consume effort without improving experience
  • Coverage does not show whether customers succeed

Evidence

  • End-to-end completion and abandonment by journey stage
  • Customer interviews and service-contact reasons
  • Completion distribution across channel and accessibility need

Guardrails

  • No exclusion of customers who need assisted service
  • Fraud, privacy, and regulatory controls remain effective
  • Average improvement does not conceal subgroup deterioration

Candidate interventions

Journey mapping · Form redesign · Policy clarification · Assisted-service changes

Review

Monthly evidence review with customers, service operations, product, and compliance.

Attribution limit

Demand mix and external eligibility rules can affect observed completion.

ProductOriginal applications revisited

Sprint planning

Product teams regularly complete sprint commitments without being able to explain their contribution to a product objective.

Starting statement

Ensure that 90% of sprint plans are aligned with the product roadmap and strategic goals.

CLEAR outcome

Over the next two quarters, at least 85% of completed sprint capacity contributes to a current product outcome or an explicit operational obligation, compared with 52% at baseline.

Inspect the outcome record

Why the first statement fails

  • A plan can be labelled aligned without changing allocation
  • Sprint compliance does not demonstrate customer or operational value
  • Necessary unplanned and enabling work is not recognised

Evidence

  • Completed capacity traced to product outcomes or explicit obligations
  • Review of exceptions, enabling work, and unplanned demand
  • Outcome evidence rather than sprint completion alone

Guardrails

  • Reliability, security, and technical sustainability work remains visible
  • Teams do not split or relabel work to improve the percentage
  • Forecast adjustment is not treated as failure

Candidate interventions

Outcome-linked planning · Capacity allocation · Backlog retirement · Planning policy changes

Review

At each product review; inspect the measure and exceptions quarterly.

Attribution limit

Contribution links are hypotheses until outcome evidence is observed.

ITOriginal applications revisited

Test planning

A digital service where critical customer journeys have experienced repeated production regressions.

Starting statement

Ensure that 100% of test plans cover all critical features and functionalities.

CLEAR outcome

Within two release quarters, reduce escaped severity-one and severity-two defects affecting critical customer journeys from 7.4 to fewer than 3.0 per 100 releases.

Inspect the outcome record

Why the first statement fails

  • Test-plan coverage is an output
  • Coverage can be complete while tests are ineffective
  • Feature counts do not represent customer or operational criticality

Evidence

  • Escaped defects classified by severity and customer journey
  • Pre-release detection effectiveness
  • Independent review of risk-based test selection

Guardrails

  • Release lead time remains within the agreed range
  • Critical accessibility, security, and recovery tests are not traded away
  • Severity classification is independently sampled

Candidate interventions

Risk-based planning · Automation · Exploratory testing · Environment improvement

Review

After each release and quarterly with product, engineering, operations, and risk owners.

Attribution limit

Changes in release size and defect reporting can alter observed rates.

ITOriginal applications revisited

Test-case writing

Teams use Azure DevOps test cases, but requirement misunderstandings are still discovered late in user acceptance testing.

Starting statement

Ensure that 100% of user stories have detailed test cases in Azure DevOps.

CLEAR outcome

Within six months, reduce user-acceptance defects caused by misunderstood acceptance behaviour from 18% to below 8% of tested stories.

Inspect the outcome record

Why the first statement fails

  • Tool documentation is an output
  • Detailed test cases can repeat an incorrect interpretation
  • One test case per story says nothing about meaningful coverage

Evidence

  • Root-cause classification of user-acceptance defects
  • Sample review of executable examples and acceptance discussions
  • Rework time attributable to expectation mismatch

Guardrails

  • Documentation effort remains proportionate to risk
  • Exploratory testing is not displaced
  • Teams are not rewarded for reclassifying defects

Candidate interventions

Example mapping · Three-amigos review · Executable specifications · Targeted test cases

Review

Monthly quality review and a formal decision after six months.

Attribution limit

Root-cause classification contains judgment and should be sampled independently.

ITOriginal applications revisited

Bug management

Critical production defects are tracked in Azure DevOps, but closure does not always represent restored customer service.

Starting statement

Ensure that 100% of critical bugs are resolved within the defined SLA.

CLEAR outcome

Over the next two quarters, restore the affected critical customer capability within its agreed service window in at least 95% of incidents, compared with 76% at baseline.

Inspect the outcome record

Why the first statement fails

  • Ticket closure can be mistaken for restored service
  • A 100% target can encourage severity or timestamp manipulation
  • Resolution speed alone ignores recurrence and workaround quality

Evidence

  • User-visible restoration time
  • Independent sampling of ticket closure accuracy
  • Recurrence and failed-fix rates

Guardrails

  • No unsafe or insecure workaround is accepted as resolution
  • Severity and SLA clocks cannot be changed without traceable approval
  • Repeated incidents trigger problem investigation

Candidate interventions

Triage changes · Observability · Runbooks · Problem management · Engineering fixes

Review

Review every critical incident; analyse patterns monthly.

Attribution limit

Supplier dependencies and incident mix affect restoration performance.

ITOriginal applications revisited

Release management

A digital product releasing frequently across services with different operational risk profiles.

Starting statement

Ensure that 100% of releases meet quality standards and deadlines.

CLEAR outcome

Across the next two quarters, at least 95% of production releases complete without rollback, critical customer disruption, or emergency remediation, while median approved lead time does not worsen.

Inspect the outcome record

Why the first statement fails

  • Deadline and checklist compliance do not prove a safe release
  • A perfect target can suppress reporting or encourage low-risk batching
  • Quality standards require observable operational consequences

Evidence

  • Change failure rate with explicit definitions
  • Customer-impact and rollback data
  • Lead-time distribution by release risk class

Guardrails

  • Teams do not reduce release frequency merely to improve the percentage
  • Mandatory security and change controls remain effective
  • Near misses and emergency remediation remain visible

Candidate interventions

Progressive delivery · Release automation · Smaller changes · Dependency controls

Review

After material failures and monthly across the release portfolio.

Attribution limit

Release complexity and risk mix must be considered when comparing periods.

ProductCross-industry illustrations

Complete a first usable production workflow

A B2B software product used by administrators to configure production workflows.

Starting statement

Launch a new onboarding wizard in Q3.

CLEAR outcome

Within six months, increase the proportion of new tenant administrators who complete their first verified production workflow within 24 hours of account creation from 35% to 60%.

Inspect the outcome record

Why the first statement fails

  • Describes an output
  • Predetermines the intervention
  • No user change or guardrail

Evidence

  • Product events with a defined verified-workflow event
  • Cohort analysis by organisation and accessibility need
  • Interviews with successful and unsuccessful administrators

Guardrails

  • Support contacts per new tenant do not increase
  • Critical journeys retain accessibility conformance
  • No reduction in production security or configuration quality

Candidate interventions

Guided setup · Improved defaults · Concierge onboarding · Simpler permissions

Review

Monthly evidence review; formal decision after two complete cohorts.

Attribution limit

Customer mix, pricing, sales qualification, and partner support may also affect time to value.

ITCross-industry illustrations

Restore critical transactions

Tier-one customer transaction services operating across multiple availability zones.

Starting statement

Move all critical services to a multi-zone architecture.

CLEAR outcome

By the end of Q4, restore user-facing transactions within 15 minutes in at least 95% of controlled single-zone-loss exercises, compared with 55% in the current baseline exercise set.

Inspect the outcome record

Why the first statement fails

  • Architecture is an output
  • Assumes resilience
  • Does not test real recovery

Evidence

  • Controlled resilience exercises
  • User-visible error and transaction completion
  • Recovery-time distribution and dependency coverage

Guardrails

  • No data loss beyond the approved recovery point
  • No bypass of change controls
  • Cost per transaction remains within an approved range

Candidate interventions

Dependency isolation · Automated failover · Runbook improvement · Service simplification

Review

After every exercise and quarterly across the service portfolio.

Attribution limit

Exercises may not represent correlated regional or supplier failures.

CybersecurityCross-industry illustrations

Reduce exploitable exposure

Internet-facing systems supporting a digital service.

Starting statement

Patch 100% of critical vulnerabilities within seven days.

CLEAR outcome

Over the next two quarters, reduce the median period during which confirmed exploitable vulnerabilities remain reachable on internet-facing production assets from 12 days to fewer than 4 days.

Inspect the outcome record

Why the first statement fails

  • Severity is not risk
  • Closure can be gamed
  • Ignores reachability and controls

Evidence

  • Validated asset inventory
  • Reachability and exploitability confirmation
  • Independent sampling of closure accuracy

Guardrails

  • Remediation respects service disruption limits
  • Closure requires evidence that exploitability is removed
  • Poorly inventoried assets remain visible

Candidate interventions

Attack-surface reduction · Patching · Segmentation · Compensating controls

Review

Weekly operational review and quarterly measure-validity review.

Attribution limit

Threat conditions and discovery rates change; trends must account for inventory improvements.

OTCross-industry illustrations

Improve line reliability safely

A packaging line where unplanned stoppages reduce output and can create restart risk.

Starting statement

Install condition-monitoring sensors on all critical motors.

CLEAR outcome

Reduce unplanned packaging-line production loss from 9.2 to fewer than 5.0 hours per 1,000 runtime hours over two consecutive rolling quarters.

Inspect the outcome record

Why the first statement fails

  • Installation is an output
  • Assumes sensors create reliability
  • Omits safety and quality

Evidence

  • Standardised downtime classification
  • Runtime, lost production, and recurrence data
  • Failure-mode analysis and operator observations

Guardrails

  • No increase in safety trips or bypassed interlocks
  • Product defect rate remains within limit
  • All changes follow engineering and management-of-change controls

Candidate interventions

Condition monitoring · Maintenance redesign · Failure elimination · Equipment redesign

Review

Weekly reliability review and quarterly cross-functional safety review.

Attribution limit

Product mix, ambient conditions, supplier quality, and operating mode can affect downtime.

OTCross-industry illustrations

Reduce red-zone exposure during tubular connections

Rig-floor crews perform repetitive tubular connections using a roughneck and coordinated pipe-handling equipment.

Starting statement

Automate every tubular connection with a new roughneck sequence.

CLEAR outcome

Within three months of stable operation, increase eligible connection sequences completed without unplanned crew entry into the defined red zone from 84% to at least 98%, while preserving stop-work authority and connection integrity.

Inspect the outcome record

Why the first statement fails

  • Automation deployment is an output
  • A faster connection can still expose people to moving equipment
  • An entry target can suppress necessary intervention unless exceptions and guardrails are explicit

Evidence

  • Sequence and equipment-event logs
  • Authorised and unplanned red-zone entries from detection records and sampled video
  • Connection torque, rework, stop-work, fault, and near-miss records

Guardrails

  • Stop-work authority and authorised contingency entry remain unrestricted
  • Interlocks, lockout controls, connection integrity, and dropped-object controls remain effective
  • Sensor gaps, manual sequences, and excluded tasks remain visible

Candidate interventions

Roughneck sequence tuning · Pipe-handler coordination · Equipment maintenance · Zone detection · Rig-floor layout · Procedures and drills

Review

Review exceptions after every tour, operational patterns weekly, and the full outcome after three stable months with crew representatives.

Attribution limit

The numerical baseline and target are illustrative. Task eligibility, rig layout, detection coverage, tubular condition, and operating practice affect comparison.

OTCross-industry illustrations

Reduce repeat top-drive interruptions

A drilling rig experiences recurring mechanical, electrical, or control faults affecting top-drive availability.

Starting statement

Deploy predictive maintenance and additional sensors for the top drive.

CLEAR outcome

Over two consecutive rolling quarters, reduce unplanned downtime attributable to repeat top-drive faults from 7.5 to below 3.0 hours per 1,000 operating hours, while alarms, interlocks, inspections, and post-maintenance tests remain effective.

Inspect the outcome record

Why the first statement fails

  • Monitoring deployment is an output
  • Alert volume does not demonstrate reliable drilling function
  • Downtime can be reduced through reclassification, constrained operation, or premature replacement

Evidence

  • Control events, condition data, and standardised downtime classification
  • Work orders, parts history, repeat-fault rate, and restoration time
  • False alerts, premature removals, and post-maintenance functional tests

Guardrails

  • No operation beyond approved equipment limits
  • Alarms, interlocks, inspections, and verification tests are not bypassed
  • Remote support and additional sensing do not introduce unmanaged cyber risk

Candidate interventions

Condition monitoring · Fault-isolation improvement · Maintenance-task redesign · Spares strategy · Control changes · Component redesign

Review

Review every material interruption, recurring patterns weekly during drilling, and reliability and safety quarterly.

Attribution limit

The numerical values are illustrative. Operating load, drilling mode, environment, failure rarity, coding quality, and configuration affect comparability.

OTCross-industry illustrations

Improve riser operating-envelope decisions

A floating drilling unit manages riser and wellhead response under changing currents, vessel motion, and weather.

Starting statement

Install a real-time riser monitoring and advisory system.

CLEAR outcome

During the next operating campaign, reduce weather and current-related non-productive time caused by late operating-envelope decisions from 16 to below 8 hours per 100 operating days, with no operation beyond mandatory load, positioning, fatigue, or safe-disconnect criteria.

Inspect the outcome record

Why the first statement fails

  • Monitoring installation is an output
  • Reduced downtime is unsafe if the operating envelope is stretched
  • Model output, sensor health, and decision lead time must remain visible

Evidence

  • Riser, wellhead, metocean, and positioning measurements
  • Alarm, decision, disconnect, and non-productive-time records
  • Sensor availability, model discrepancies, load excursions, and fatigue findings

Guardrails

  • Mandatory operating and safe-disconnect criteria remain controlling
  • Degraded monitoring triggers the approved fallback
  • Alarms are not suppressed and uncertain downtime causes remain reviewable

Candidate interventions

Sensor assurance · Alarm rationalisation · Decision-support displays · Forecast integration · Threshold review · Operational drills

Review

Review every material alarm or interruption, patterns weekly during the campaign, and the complete outcome after the campaign.

Attribution limit

The numerical values are illustrative. Sea state, currents, well geometry, sensor uncertainty, vessel response, and campaign mix limit comparison and attribution.

OTCross-industry illustrations

Improve safe availability of mobile assets

A fleet of safety-critical mobile assets experiences repeat faults, no-fault-found removals, and early post-maintenance withdrawals.

Starting statement

Deploy predictive maintenance analytics across the fleet.

CLEAR outcome

Across two consecutive quarterly review periods, increase scheduled operating periods completed without maintenance-driven withdrawal in the first 10 operating hours from 78% to 90%, without weakening release criteria, mandatory inspections, functional tests, configuration control, or safety reporting.

Inspect the outcome record

Why the first statement fails

  • Analytics deployment is an output
  • Availability can improve by deferring faults or excluding difficult assets
  • Prediction accuracy alone does not demonstrate safe maintenance effectiveness

Evidence

  • Operating-period completion and early withdrawal records
  • Repeat faults, first-time fix effectiveness, and no-fault-found removals
  • False alerts, missed detections, safety findings, configuration, and functional-test records

Guardrails

  • No asset remains in service solely to protect availability
  • Release criteria, mandatory work, configuration control, and independent safety review remain effective
  • False negatives, rare high-consequence modes, and sensor-coverage gaps remain visible

Candidate interventions

Condition-based maintenance thresholds · Diagnostic improvement · Troubleshooting guidance · Spares positioning · Maintenance training · Design feedback

Review

Review every early withdrawal, maintenance and operating patterns monthly, and the full outcome quarterly with independent safety participation.

Attribution limit

The numerical values are illustrative. Duty profile, environment, fleet mix, configuration, rare failures, sensor coverage, and maintenance-record quality affect interpretation.

OperationsCross-industry illustrations

Make regulated decisions sooner

An internal approval service for regulated expenditure.

Starting statement

Automate the approval workflow.

CLEAR outcome

Within nine months, reduce median elapsed time for complete standard applications from eight working days to four, while maintaining at least 98% first-pass decision correctness.

Inspect the outcome record

Why the first statement fails

  • Predetermines automation
  • Omits decision quality
  • Can hide unequal delays

Evidence

  • End-to-end timestamps
  • Independent decision-quality sampling
  • 90th-percentile time by request type and group

Guardrails

  • Required audit evidence remains complete
  • Fraud and control exceptions remain in range
  • No group experiences hidden material deterioration

Candidate interventions

Form redesign · Triage · Decision-right delegation · Policy simplification

Review

Monthly, with quarterly review of correctness and distribution.

Attribution limit

Demand mix, policy changes, and staffing can affect elapsed time.

CapabilityCross-industry illustrations

Apply problem-solving capability

Frontline teams repeatedly escalate recurring operational defects.

Starting statement

Train all frontline employees in root-cause analysis.

CLEAR outcome

Within nine months, frontline teams independently resolve at least 70% of eligible recurring defects using evidence-based problem solving, compared with 25% at baseline, with no increase in recurrence or unsafe local workarounds.

Inspect the outcome record

Why the first statement fails

  • Training is an output
  • Knowledge is not application
  • No connection to operational value

Evidence

  • Explicit eligibility rules
  • Independent sample of analysis quality
  • Resolution, recurrence, and escalation data

Guardrails

  • High-risk defects follow mandatory escalation
  • Closure requires evidence of sustained effect
  • Teams are not rewarded for suppressing escalations

Candidate interventions

Training · Coaching · Protected improvement time · Decision-right changes

Review

Monthly sampling and a quarterly capability review.

Attribution limit

Complexity, specialist availability, leadership behaviour, and decision rights affect results.

Public serviceCross-industry illustrations

Improve access to permit decisions

A regulated permit service delivered through digital and assisted channels.

Starting statement

Launch a self-service permit portal.

CLEAR outcome

Within nine months, increase the share of complete standard applications receiving a correct decision within 10 working days from 46% to 75%, with no material deterioration by channel or applicant group.

Inspect the outcome record

Why the first statement fails

  • A portal is an output
  • Digital completion is not equitable access
  • Speed alone can weaken decision quality

Evidence

  • End-to-end case timestamps
  • Independent decision-quality samples
  • Distribution by channel and applicant group

Guardrails

  • Appeal outcomes do not worsen
  • Assisted access remains available
  • Complex applicants are not reclassified to protect the target

Candidate interventions

Form redesign · Triage · Policy clarification · Assisted support · Workflow changes

Review

Monthly operational review and quarterly review with applicants, casework, policy, accessibility, and oversight.

Attribution limit

Case complexity, demand mix, policy change, and incomplete demographic data affect interpretation.

HealthCross-industry illustrations

Reduce medication discrepancies

Adult patients discharged from an acute-care service with changes to long-term medication.

Starting statement

Deploy a medication-reconciliation alert to every ward.

CLEAR outcome

Within 12 months, reduce clinically significant unresolved medication discrepancies identified within 72 hours of discharge from 11.4 to below 5.0 per 100 eligible discharges.

Inspect the outcome record

Why the first statement fails

  • Alert deployment is an output
  • Alerts can be ignored or create fatigue
  • The patient safety condition is not stated

Evidence

  • Structured record review
  • Pharmacist reconciliation sample
  • Patient and primary-care follow-up

Guardrails

  • Discharge is not inappropriately delayed
  • Alert burden remains tolerable
  • High-risk groups do not experience worse performance

Candidate interventions

Pharmacist review · Workflow redesign · Patient communication · Record integration · Targeted alerts

Review

Monthly safety review and quarterly review with clinical, pharmacy, patient, data-quality, and primary-care representation.

Attribution limit

Documentation quality and post-discharge care outside the service affect detection and attribution.

SustainabilityCross-industry illustrations

Reduce energy intensity without shifting harm

A distribution centre with variable weather, occupancy, and order mix.

Starting statement

Replace all warehouse lighting with LEDs.

CLEAR outcome

Over two comparable rolling quarters, reduce grid electricity per dispatched order from 1.80 to at most 1.35 kWh, while absolute use, service, safety, and working conditions remain visible.

Inspect the outcome record

Why the first statement fails

  • Equipment replacement is an output
  • Intensity can improve while absolute use rises
  • Work or energy can be shifted outside the boundary

Evidence

  • Revenue-grade meter data
  • Warehouse throughput and order mix
  • Weather normalisation and workplace observations

Guardrails

  • Lighting and workplace safety do not worsen
  • Picking quality and service remain in range
  • Outsourced activity is included in the boundary review

Candidate interventions

Lighting and controls · HVAC optimisation · Equipment settings · Scheduling · Process redesign

Review

Monthly energy review and quarterly method, boundary, finance, safety, and workforce review.

Attribution limit

Weather adjustment, order mix, on-site generation, and outsourcing materially affect comparability.

Supply chainCross-industry illustrations

Improve material availability reliability

Critical production materials sourced from approved suppliers with different transit and quality risks.

Starting statement

Implement a new supply-planning platform.

CLEAR outcome

Within three quarters, increase production orders starting with all critical materials available from 72% to 90%, while inventory exposure and incoming-quality failures remain within approved limits.

Inspect the outcome record

Why the first statement fails

  • Software implementation is an output
  • Planning accuracy is not material availability
  • Inventory or quality can be sacrificed to hit service

Evidence

  • Order-start readiness
  • Supplier and transit variability
  • Inventory age and incoming-quality data

Guardrails

  • No material increase in obsolete inventory
  • Quality acceptance rules remain intact
  • Supplier risk is not hidden through expedites

Candidate interventions

Planning policy · Supplier development · Safety-stock segmentation · Logistics changes · Platform configuration

Review

Monthly supply review and quarterly review across procurement, operations, quality, finance, and suppliers.

Attribution limit

Demand volatility, specification changes, and force-majeure events constrain period comparisons.

DataCross-industry illustrations

Make decision data trustworthy in time

Operations leaders use a daily dashboard to make same-shift staffing and flow decisions.

Starting statement

Build a real-time operations dashboard.

CLEAR outcome

Within six months, at least 95% of scheduled same-shift decisions use data no more than 15 minutes old and within agreed reconciliation tolerance, compared with 61% at baseline.

Inspect the outcome record

Why the first statement fails

  • A dashboard is an output
  • Real-time can still be wrong or unused
  • The consequential decision is missing

Evidence

  • Pipeline freshness and reconciliation checks
  • Decision logs and user observation
  • Exception and override analysis

Guardrails

  • Privacy and access controls remain effective
  • Manual fallback remains available
  • Teams do not defer decisions merely to improve usage

Candidate interventions

Pipeline redesign · Data contracts · Decision workflow changes · Training · Source-system repair

Review

Weekly data-quality review and monthly decision-use review with operations, data, privacy, and users.

Attribution limit

Decision logs may not capture informal judgment, and data use does not prove decision quality.

CapabilityCross-industry illustrations

Sustain a new decision practice

A product portfolio introducing evidence reviews before major investment decisions.

Starting statement

Complete training and communications for all portfolio leaders.

CLEAR outcome

By the third quarterly cycle, at least 80% of material investment decisions document credible user, operational, and financial evidence plus explicit uncertainty, compared with 29% at baseline.

Inspect the outcome record

Why the first statement fails

  • Communication and training are outputs
  • Attendance does not demonstrate changed practice
  • Documentation can become ceremonial

Evidence

  • Independent sample of decision records
  • Observation of review meetings
  • Follow-up on challenged assumptions and decision reversals

Guardrails

  • Decision time remains proportionate to risk
  • Dissent and uncertainty are not suppressed
  • Teams are not rewarded for producing longer documents

Candidate interventions

Training · Coaching · Decision templates · Review rights · Leadership modelling

Review

After each quarterly cycle with leaders, teams, finance, research, operations, and independent reviewers.

Attribution limit

Documentation quality is partly judgmental and does not prove that evidence determined the decision.

ProductCross-industry illustrations

Resolve critical uncertainty before product commitment

A product group is approaching a roadmap commitment while the target segment, problem severity, alternatives, and willingness to change remain uncertain.

Starting statement

Interview 30 prospective customers.

CLEAR outcome

By the end of a 12-week discovery, the product investment group records a go, change, or stop decision with corroborating or contradicting evidence for every critical desirability, viability, and feasibility assumption.

Inspect the outcome record

Why the first statement fails

  • Interview count is an activity
  • Stated preference can be mistaken for behaviour
  • Research has no value if it cannot change the decision

Evidence

  • Observed current behaviour and service data
  • Diverse research and market evidence
  • Prototype, commercial, and technical tests
  • Contradictory findings and decision rationale

Guardrails

  • Consent, privacy, and research integrity remain protected
  • Representative and underserved users are included
  • The team remains able to stop

Candidate interventions

Contextual research · Problem interviews · Market analysis · Prototype tests · Concierge trials · Technical spikes

Review

Weekly evidence review and a formal investment decision at the end of discovery.

Attribution limit

Illustrative timing. Small samples, recruitment bias, emerging markets, and prototype conditions limit generalisation.

ProductCross-industry illustrations

Realise customer value from roadmap investments

A product portfolio reliably delivers roadmap items but rarely reviews whether the intended customer condition changed.

Starting statement

Deliver 90% of the product roadmap.

CLEAR outcome

Across the next two quarterly reviews, increase completed roadmap investments that demonstrate their intended customer change within the agreed evidence window from 32% to 65%, with a recorded continue, adapt, or stop decision for every investment.

Inspect the outcome record

Why the first statement fails

  • Roadmap completion measures output
  • Feature delivery does not establish value
  • No response to negative or inconclusive evidence

Evidence

  • Outcome and baseline defined before commitment
  • Product telemetry and customer research
  • Commercial, service, and support evidence
  • Contrary evidence and follow-on decision

Guardrails

  • Safety, security, compliance, and accessibility work remains funded
  • Exploratory and long-horizon work is not excluded
  • Teams cannot relabel weak evidence

Candidate interventions

Outcome-based roadmaps · Smaller bets · Discovery · Instrumentation · Staged funding · Portfolio retirement

Review

Monthly product evidence review and quarterly portfolio decision.

Attribution limit

Illustrative figures. Adoption cycles, shared product effects, external market changes, and small cohorts complicate attribution.

ProductCross-industry illustrations

Improve retained paid adoption transparently

A subscription product has healthy trial demand but weak paid conversion, early cancellation, and uncertainty about total price and plan fit.

Starting statement

Launch new subscription tiers and increase conversion.

CLEAR outcome

Across two complete quarterly acquisition cohorts, increase qualified trial organisations that choose a paid plan and remain meaningfully active 90 days later from 18% to 28%.

Inspect the outcome record

Why the first statement fails

  • Pricing launch is an output
  • Checkout conversion can conceal poor fit
  • Short-term revenue can undermine trust and retention

Evidence

  • Qualified-trial conversion and 90-day activity
  • Cancellation, downgrade, and involuntary churn
  • Total-price comprehension and complaints
  • Support demand and segment distribution

Guardrails

  • Total price and mandatory charges remain clear
  • Consent and cancellation remain fair
  • Accessibility, service quality, and smaller-customer treatment do not worsen

Candidate interventions

Packaging changes · Transparent total-price presentation · Trials · Contract simplification · Value communication · Product improvement

Review

Monthly cohort review and a formal pricing decision after two complete 90-day cohorts.

Attribution limit

Illustrative figures. Seasonality, acquisition mix, competitor pricing, sales incentives, and product changes affect comparison.

ITCross-industry illustrations

Protect critical SaaS tenant journeys

A multitenant service reports strong infrastructure uptime while some tenant groups experience failed or slow critical journeys.

Starting statement

Achieve 99.99% platform uptime.

CLEAR outcome

For three consecutive months, at least 99.9% of eligible critical-journey opportunities complete within the agreed threshold for every committed service tier, with no material tenant segment below 99.5%.

Inspect the outcome record

Why the first statement fails

  • Infrastructure uptime is not customer experience
  • Aggregate performance can hide tenant harm
  • A target without an error-budget decision is inert

Evidence

  • Tenant-aware synthetic and real-user journeys
  • Good-event ratio by tier and segment
  • Error-budget consumption and affected tenants
  • Incidents, dependencies, and support evidence

Guardrails

  • Security, privacy, and data integrity remain protected
  • Recovery capability is exercised
  • Difficult tenants and valid errors remain visible

Candidate interventions

SLO design · Dependency isolation · Capacity changes · Resilience engineering · Progressive delivery · Service-tier redesign

Review

Weekly reliability review, monthly service-level decision, and immediate review after a material breach.

Attribution limit

Illustrative thresholds. Instrumentation gaps, low-volume tenants, client conditions, and third-party dependencies affect interpretation.

ITCross-industry illustrations

Realise verified value from cloud migration

A migration programme reports application counts while some moved services retain legacy dependencies, unplanned cost, or weak operations.

Starting statement

Migrate 80 applications to the cloud this year.

CLEAR outcome

By the end of each migration wave plus 90 days, increase priority services meeting their approved reliability, recovery, security, performance, and unit-cost commitments from 40% to 85%, with a decision on every legacy-retirement exception.

Inspect the outcome record

Why the first statement fails

  • Application count is an output
  • A technical move does not prove benefit
  • Dual running and retained dependencies can hide failure

Evidence

  • Business-service telemetry and resilience exercises
  • Security and operational evidence
  • Cost per business transaction
  • Dependency and legacy retirement records

Guardrails

  • Data integrity, privacy, security, and continuity remain protected
  • User performance does not deteriorate
  • Cutover is reversible until recovery is proven

Candidate interventions

Rehost · Replatform · Refactor · Retire · Operational redesign · Cost optimisation · Contract change

Review

Fortnightly wave review, monthly benefit review, and a formal decision after the 90-day stabilisation period.

Attribution limit

Illustrative figures. Demand, contracts, shared platforms, one-time costs, and workload heterogeneity affect comparison.

ITCross-industry illustrations

Deliver software changes safely in normal hours

A customer-facing service releases in large batches after long queues and depends on risky after-hours coordination.

Starting statement

Implement CI/CD for every application.

CLEAR outcome

For three consecutive months, reduce median change lead time from 10 days to 2 days while keeping change fail rate at or below 8% and standard releases requiring after-hours work below 10%.

Inspect the outcome record

Why the first statement fails

  • Pipeline implementation is an output
  • Speed without instability is unsafe
  • Release frequency can rise without delivering value

Evidence

  • Change lead time and deployment frequency
  • Change fail and deployment rework rates
  • Failed-deployment recovery time
  • User incidents and after-hours work

Guardrails

  • Security and segregation of duties remain effective
  • Testing and recovery are not weakened
  • Teams do not hide failures or transfer work to on-call staff

Candidate interventions

Smaller changes · Automated testing · Progressive delivery · Approval redesign · Platform improvement · Dependency decoupling

Review

Weekly flow review, review after each failed deployment, and monthly balanced-metric decision.

Attribution limit

Illustrative figures. Architecture, risk controls, change size, work mix, and timestamp consistency affect comparison.

Public serviceCross-industry illustrations

Complete a public service across channels

Applicants repeat information, switch channels, or abandon a public service because its digital and assisted journeys are fragmented.

Starting statement

Move 90% of applications online.

CLEAR outcome

By month nine of public beta, increase eligible people completing a correct application without avoidable repeat contact from 52% to 78%, with no material user group below 65% across digital and assisted routes.

Inspect the outcome record

Why the first statement fails

  • Digital take-up is not successful completion
  • Channel shift can conceal exclusion
  • The end-to-end service and decision quality are missing

Evidence

  • Correct completion, abandonment, and avoidable contact
  • Channel switching and assisted-support wait
  • Decision-quality samples and appeals
  • Accessibility and performance by user group

Guardrails

  • Entitlement and procedural fairness remain intact
  • Accessibility, privacy, and appeal rights do not weaken
  • Assisted access remains visible and adequately supported

Candidate interventions

Journey simplification · Form redesign · Status information · Assisted support · Policy clarification · Casework integration

Review

Monthly service review, fortnightly research synthesis, and quarterly equity and decision-quality review.

Attribution limit

Illustrative figures. Eligibility, policy, seasonal demand, user-group data, and third-party support affect interpretation.

Public serviceCross-industry illustrations

Adopt a joined-up casework practice

A new case-management platform is live, but parallel records, incomplete handoffs, and workarounds continue to delay cases.

Starting statement

Train all caseworkers and roll out the new case-management platform.

CLEAR outcome

By month six in each participating office, at least 85% of sampled eligible cases follow the joined-up practice with a correct handoff and no duplicate data entry, without increasing incorrect decisions or staff workload risk.

Inspect the outcome record

Why the first statement fails

  • Training and rollout are outputs
  • System use can coexist with shadow processes
  • Adoption must include real work and service quality

Evidence

  • Independent case sampling and workflow logs
  • Handoff completeness and duplicate entry
  • Delay, correction, and decision quality
  • Staff observation, workarounds, and user feedback

Guardrails

  • Lawful discretion and decision quality remain protected
  • Privacy, accessibility, and appeal rights do not weaken
  • Staff wellbeing and continuity are monitored

Candidate interventions

Workflow redesign · Coaching · Local change networks · Platform fixes · Role clarification · Policy and partner agreements

Review

Weekly transition review, monthly practice sampling, and a formal six-month adoption decision.

Attribution limit

Illustrative figures. Case complexity, policy interpretation, partner maturity, sampling judgment, and staffing changes affect comparison.

CapabilityCross-industry illustrations

Contain priority incidents through practised capability

Responders complete incident training, but exercises reveal delayed escalation, unclear roles, and weak coordination.

Starting statement

Certify every responder in incident management.

CLEAR outcome

Across two quarterly exercise cycles, cross-functional responders correctly detect, classify, coordinate, and contain at least 85% of representative priority-incident scenarios within 30 minutes, with no critical unsafe action.

Inspect the outcome record

Why the first statement fails

  • Certification is an output and proxy
  • Individual knowledge does not prove team response
  • Capability must work under representative conditions

Evidence

  • Exercise telemetry and independent observations
  • Classification, accountable lead, and containment time
  • Decision and communication quality
  • Live-incident comparison and learning closure

Guardrails

  • Exercises protect production and personal wellbeing
  • Security, safety, and legal obligations remain intact
  • A blameless learning environment is maintained

Candidate interventions

Role-based learning · Simulations · Runbook improvement · Decision-right clarification · Tooling · Cross-team exercises

Review

Debrief after every exercise or material incident and make a quarterly capability decision.

Attribution limit

Illustrative thresholds. Exercises cannot reproduce every live pressure, novel failure, supplier constraint, or team composition.

CapabilityCross-industry illustrations

Apply data skills to recurring operational decisions

Frontline supervisors have dashboards and training, but important decisions still rely on habit or misunderstood measures.

Starting statement

Train all supervisors on dashboards and data literacy.

CLEAR outcome

By month nine, at least 80% of independently sampled recurring operational decisions use validated evidence, state material uncertainty, and record an appropriate action for three consecutive months.

Inspect the outcome record

Why the first statement fails

  • Training completion is an output
  • Dashboard use does not prove sound judgment
  • Knowing when evidence is weak is part of capability

Evidence

  • Independent decision samples
  • Measure interpretation and data-quality escalation
  • Decision timeliness and outcome follow-up
  • Observation, system use, and coaching evidence

Guardrails

  • Professional judgment and authorised discretion remain protected
  • Privacy, fairness, and safety do not weaken
  • Urgent action is not delayed for documentation

Candidate interventions

Role-based learning · Coached cases · Measure redesign · Data-quality improvement · Decision aids · Peer review

Review

Monthly independent sampling, quarterly outcome review, and immediate review after a material decision failure.

Attribution limit

Illustrative figures. Decision quality is partly judgmental, consequences can be delayed, and observer effects limit attribution.

SustainabilityCross-industry illustrations

Reduce freshwater intensity without harming discharge quality

A manufacturing process has high water use, incomplete loss visibility, and growing exposure to local water stress.

Starting statement

Install water-saving equipment across the plant.

CLEAR outcome

For two consecutive quarters, reduce freshwater withdrawal per conforming unit from 2.4 to 1.9 cubic metres in comparable production, without materially worsening absolute high-stress withdrawal or discharge quality.

Inspect the outcome record

Why the first statement fails

  • Equipment installation is an output
  • Intensity can hide absolute use or boundary shifts
  • Water quantity cannot be separated from discharge quality

Evidence

  • Calibrated source and process metering
  • Water balance, loss, and reuse
  • Production, product quality, and absolute withdrawal
  • Discharge quality and local water context

Guardrails

  • Product quality, hygiene, and worker safety remain intact
  • Permit and ecological requirements are protected
  • Supplier and community impacts remain visible

Candidate interventions

Leak repair · Controls · Cleaning redesign · Closed-loop reuse · Equipment changes · Treatment improvement

Review

Monthly water-balance review and quarterly environmental, boundary, and target decision.

Attribution limit

Illustrative figures. Meter coverage, evaporation, product mix, weather, recycled-water accounting, and boundaries affect comparison.

SustainabilityCross-industry illustrations

Keep returned components in safe productive use

Returned equipment is largely downcycled or disposed of despite components that may be safely reused or refurbished.

Starting statement

Recycle 90% of returned equipment.

CLEAR outcome

Across two consecutive half-years, increase eligible returned-component mass verified as safely reused or refurbished for productive use from 34% to 60%, with component-level reporting for critical materials.

Inspect the outcome record

Why the first statement fails

  • Recycling rate can reward low-value treatment
  • Mass diversion can hide stockpiling or export
  • Safe life extension and displaced virgin material are missing

Evidence

  • Serial and material traceability
  • Inspection, functional test, and life-extension evidence
  • Virgin material displaced and waste destination
  • Quality, warranty, hazardous-material, and partner records

Guardrails

  • Product and worker safety remain protected
  • Hazardous material, data security, and export rules remain intact
  • Destinations are verified and double counting is prevented

Candidate interventions

Design for repair · Take-back terms · Inspection improvement · Refurbishment · Parts harvesting · Recovery partnerships

Review

Monthly operational review and formal half-year lifecycle, boundary, quality, and market decision.

Attribution limit

Illustrative figures. Product mix, return condition, mass-based measures, partner traceability, and secondary-market demand affect interpretation.