CASM flashcards

40 cards. Read the term and the definition here. The deck at the top of this page is optional.

Definition of Done

A shared, explicit quality bar that every product increment must meet before it is considered complete, applied uniformly to all backlog items rather than negotiated per item.

Pull system

A work management approach where new work is started only when downstream capacity becomes available, replacing forecast-based scheduling that builds queues and excess work in progress.

TIMWOOD wastes

The seven Lean wastes: Transport, Inventory, Motion, Waiting, Overproduction, Overprocessing, and Defects. Service contexts commonly add unused talent and unused capacity as additional wastes.

Empirical process control

Decision-making based on transparency, inspection, and adaptation, used when requirements and technology are too unpredictable for detailed up-front planning to remain accurate.

Transformation backlog

A prioritized, visible list of organizational improvement items, each with an owner and a Definition of Done, managed iteratively instead of as a fixed multi-year project plan.

Pilot value stream selection

Choosing a first adoption target that has high pain, few cross-team dependencies, a visible customer, and a sponsor able to shield the team from interrupt work.

Adoption readiness gate

A pre-scaling check confirming the team can deliver a thin slice to a real user each iteration, has a decision-capable product owner, and can measure lead time.

Behavior-based adoption metric

Evidence that working practice has changed, such as release frequency, honored WIP limits, or closed retrospective actions, rather than training attendance or tool installation counts.

INVEST criteria

A checklist for well-formed user stories: Independent, Negotiable, Valuable, Estimable, Small, and Testable, used to decide whether a story is ready for development.

Three amigos

A short collaborative review of a story by business, development, and testing perspectives, held before development to expose ambiguity and agree concrete examples.

Example mapping

A structured conversation technique that breaks a story into rules, then concrete examples under each rule, then open questions that must be resolved before development starts.

Service blueprint

A service design diagram showing frontstage customer actions, backstage employee actions, supporting processes, and tangible evidence at each step, revealing failure points and handoffs.

Deployment versus release

Deployment installs an artifact into an environment; release makes functionality visible to users. Separating them enables dark launches, feature flags, and canary exposure.

Canary release

A deployment pattern that routes a small share of live traffic to the new version and expands exposure only while health signals such as error rate stay within thresholds.

Standard change

A pre-authorized, low-risk, repeatable change executed by a documented procedure. It avoids per-instance approval but is still recorded and available for audit.

Blue-green deployment

Running two identical environments and switching traffic from the live one to the newly deployed one, allowing near-instant rollback by switching traffic back.

Flow efficiency

Process time divided by lead time, expressed as a percentage. It measures how much of the total elapsed time a work item spends in value-adding activity rather than waiting.

Little's Law

Average work in progress equals throughput multiplied by average cycle time. In a stable system it predicts any one of the three variables from the other two.

Kaizen versus kaikaku

Kaizen is continuous incremental improvement made by the people doing the work; kaikaku is radical process redesign. Both generate items for the same improvement backlog.

A3 problem solving

A one-page structured report covering background, current condition, target condition, root cause analysis, countermeasures, action plan, and follow-up, used to drive improvement.

Classes of service

Kanban policies that treat different work types differently, typically expedited, fixed date, standard, and intangible, each with its own rules and capacity allocation.

Cumulative flow diagram

A chart of work item counts per workflow state over time. Widening bands indicate growing work in progress; flat bands indicate blocked flow; band slope estimates throughput.

Cycle time percentiles

Reporting the 50th, 85th, and 95th percentile of cycle time rather than the mean, so delivery commitments express a service level and reflect real variability.

Scrumban

A hybrid combining Scrum's cadence, roles, and ceremonies with Kanban's explicit workflow, work-in-progress limits, and pull-based replenishment, often used for interrupt-heavy support work.

The Three Ways

DevOps principles of flow from development to operations, amplified feedback from operations back to development, and a culture of continual experimentation and learning.

Blameless postmortem

A structured review after an incident that examines timeline, contributing system conditions, detection and response, and assigns owned actions without attributing fault to individuals.

Toil

Manual, repetitive, automatable, tactical operational work that scales linearly with service growth. It is measured as a share of operational time and deliberately capped.

Shift left

Moving quality, security, and compliance verification earlier in the delivery lifecycle, typically by automating checks in the pipeline instead of performing late-stage manual reviews.

Error budget

One minus the service level objective, expressing the amount of unreliability permitted in a period. Spending it fast triggers reliability work; leaving it unspent can justify more releases.

Work item age

The elapsed time since a work item entered its current started state. It highlights items at risk of breaching service expectations before they are completed.

Goodhart's law

The observation that when a measure becomes a target it stops being a good measure, which is why each reported metric should be paired with a counter-metric.

DORA delivery metrics

Deployment frequency, lead time for changes, change failure rate, and time to restore service, used together to describe delivery performance without rewarding speed alone.

Policy as code

Expressing control requirements such as encryption, logging, and access rules as automated tests in the pipeline, producing continuous compliance evidence instead of periodic manual sampling.

Risk treatment options

The four responses to an identified risk: avoid the activity, reduce likelihood or impact, transfer it to another party, or accept it with a named owner and review date.

Continuous control monitoring

Automated, ongoing verification of control effectiveness that flags drift as it occurs, replacing point-in-time audit sampling and lowering the cost of collecting evidence.

Segregation of duties in pipelines

Separating the roles that author, approve, and deploy a change, with each step recorded as an automated attestation so the control is enforced without manual gates.

ADKAR model

An individual change model with five sequential elements: Awareness, Desire, Knowledge, Ability, and Reinforcement. Change stalls at the first missing element, which diagnosis must identify.

Change saturation

The condition where a group faces more concurrent changes than it can absorb, causing adoption to fall even when each individual change is well designed and communicated.

Kotter's eight steps

A sequence for organizational change: urgency, guiding coalition, vision, communication, empowerment, short-term wins, consolidating gains, and anchoring new practice in culture.

Anchoring change

Embedding new practice in formal mechanisms such as job descriptions, budgeting, performance review, and onboarding so it persists after the original sponsors depart.