Software Delivery Glossary
51 clear definitions of the concepts, frameworks, metrics, and terms we use to diagnose and fix broken delivery models.
3
30-Day Satisfaction Guarantee
Sonatafy's risk-reversal commitment: if a placed engineer is not the right fit within the first 30 days, the engagement is renegotiated or refunded. Designed to remove hiring risk from the buying decision.
A
AI Agent
An autonomous or semi-autonomous AI system that can plan, execute, and iterate on multi-step tasks (e.g., writing code, running tests, opening pull requests) under human supervision. Distinct from a single-shot LLM completion.
AI in Delivery
The systematic embedding of AI tools (GitHub Copilot, Claude, Cursor, internal agents) across all 7 phases of the SDLC, from sprint planning and code review to testing and deployment, under a single accountable delivery owner. Distinct from ad-hoc developer use of AI.
Learn more →AI Process Automation
Replacing manual handoffs between systems (CRM, ERP, ticketing, internal tools) with AI-driven workflows. Focused on eliminating the 'swivel-chair' work that consumes hours of operational capacity each week.
Learn more →B
Backlog
The ordered list of work items (features, bugs, technical debt) waiting to be delivered. A healthy backlog is groomed, prioritized, and sized to roughly 2 to 3 sprints of work, not 200.
Backlog Illusion
The false sense of progress created by a growing backlog. A full backlog is not a healthy backlog. When items accumulate faster than they ship, the backlog becomes a liability disguised as a pipeline, masking the fact that delivery throughput cannot keep pace with intake.
Read the concept →Bench
A pool of pre-vetted engineers under retainer who can be deployed to client engagements quickly. The bench is what enables a 3 to 5 week match window for senior LATAM talent.
C
Change Failure Rate
The percentage of deployments to production that result in degraded service and require remediation (rollback, hotfix, patch). Elite teams operate below 5%.
CI/CD
Continuous Integration / Continuous Delivery. The automated practice of merging code changes frequently and deploying them to production through repeatable pipelines. CI/CD maturity is a core indicator of delivery health.
Code Review
The structured peer evaluation of code changes before merge. Effective code review catches defects early, spreads architectural knowledge, and is increasingly augmented (not replaced) by AI tooling.
Coordination Tax
The invisible overhead that compounds as engineering teams grow without clear delivery ownership. Every new engineer adds communication paths, and each path adds latency, context switching, and decision overhead. A 30-person team can lose $900K to $1.5M annually to coordination tax.
Read the concept →Coordination Tax Calculator
A 5-step interactive diagnostic that estimates the dollar value of coordination overhead in your engineering organization based on team size, meeting load, average comp, and decision latency.
Run the calculator →Copilot Coverage
The percentage of an engineering team actively using AI coding assistants in their daily workflow. High copilot coverage is necessary but not sufficient, measurable delivery impact requires integrating AI across the full SDLC.
Cross-Functional Team
A team containing all the disciplines needed to deliver an outcome end-to-end (engineering, design, QA, product, ops). Cross-functional structure reduces handoff latency and is foundational to high-performing delivery.
Cycle Time
The elapsed time from when work is started until it reaches production. Cycle time isolates execution efficiency from queue/wait time and is one of the four DORA metrics for high-performing teams.
D
Data Transformation & AI
An engagement model that establishes the data foundation (warehouse, lineage, governance, quality) required before reliable AI workloads can be deployed. 'Data first, AI second', most failed AI initiatives fail at the data layer, not the model layer.
Learn more →Definition of Done
The explicit, shared checklist of conditions a work item must meet before it is considered complete (e.g., tests written, code reviewed, deployed to staging, documented). Ambiguous DoD is a leading cause of missed commitments.
Delivery Model
The organizational structure, processes, and accountability mechanisms that determine how software moves from idea to production. A broken delivery model, not bad engineers, is the most common cause of missed roadmap commitments.
See how we work →Deployment Frequency
How often a team releases changes to production. Elite teams deploy on demand (multiple times per day); low-performing teams deploy monthly or less. One of the four DORA metrics.
Discovery Call
The first conversation between a prospective client and a Sonatafy delivery expert. Used to understand the delivery problem, current state, constraints, and recommend an engagement model, not a sales pitch.
Book a discovery call →DORA Metrics
Four research-backed indicators of software delivery performance defined by the DevOps Research and Assessment team: deployment frequency, lead time for changes, change failure rate, and mean time to recovery (MTTR).
E
Engagement Quiz
A 5-question diagnostic that recommends the right Sonatafy engagement model (Consulting, Managed POD, Staff Augmentation, or Data Transformation) based on team size, budget, timeline, and delivery maturity.
Take the quiz →Engineering Delivery Maturity
A four-stage model measuring an organization's progression from ad hoc execution (Stage 1) through structured delivery (Stage 4: predictable, instrumented, accountable). Most organizations believe they are at Stage 3 but are operationally stuck at Stage 2.
View the framework →F
Feature Flag
A configuration mechanism that allows code to be deployed to production but selectively enabled for users. Feature flags decouple deployment from release and are essential for trunk-based development and progressive rollouts.
Fractional CTO
Executive-level engineering leadership provided on a part-time or project basis. Includes delivery system design, engineering org assessment, process architecture, and hands-on execution support, not just a strategy deck.
Learn more →H
Hallucination
When an AI model generates output that is fluent and plausible but factually wrong or fabricated. Mitigation requires retrieval grounding, structured outputs, evaluation harnesses, and human review at decision points.
Handoff
The transfer of work, context, or responsibility from one person or team to another. Each handoff introduces latency, information loss, and accountability dilution. Reducing handoffs is a core delivery acceleration lever.
L
Lead Time
The elapsed time from when work is committed to until it reaches production. Sonatafy clients typically see 30 to 50% lead time reduction within the first 90 days of a managed delivery engagement.
See results →LLM (Large Language Model)
A neural network trained on very large text corpora that can generate, summarize, classify, and reason over natural language and code. Examples: GPT, Claude, Gemini, Llama.
M
Managed Delivery POD
A self-contained engineering team led by a US-based principal engineer managing 4 senior LATAM engineers. The POD owns sprint execution from backlog to production, providing full delivery accountability rather than just staffing capacity.
Learn more →Mean Time to Recovery (MTTR)
The average time it takes to restore service after a production failure. Elite teams recover in under one hour; low-performing teams take days or weeks.
N
Nearshore
Engineering talent based in the same or adjacent time zones as the client (typically LATAM for US-based companies). Nearshore eliminates the asynchronous handoff penalty of offshore models while preserving cost advantages over onshore staffing.
Compare nearshore vs offshore →O
Offshore
Engineering talent based in time zones with limited overlap with the client (typically Asia or Eastern Europe for US clients). Offshore can lower direct cost but introduces handoff latency, communication friction, and ownership dilution.
Compare nearshore vs offshore →Ownership Gap
The structural gap between engineering effort and delivery outcomes. When no single person owns the end-to-end delivery result, teams stay busy but roadmap commitments slip. The ownership gap is the root cause of most delivery failures.
Read the concept →P
Platform Enablement
Targeted engineering engagements focused on CI/CD pipelines, infrastructure automation, QA maturity, and DevOps modernization. Includes documentation, pair programming, and structured handoff so internal teams can sustain the improvements.
Learn more →Pod
A small, persistent, cross-functional team aligned to a product area or outcome. The pod model reduces coordination overhead and increases ownership compared to functional silos.
Principal Engineer
A senior US-based technical leader (typically 12+ years experience) who manages a Managed Delivery POD. Responsible for sprint outcomes, code quality, architectural decisions, and the day-to-day client relationship.
R
RAG (Retrieval-Augmented Generation)
An architecture where an LLM is given relevant external context (retrieved from a vector store, database, or API) at inference time. RAG is the standard pattern for grounding LLM outputs in proprietary or current data.
Ramp Time
The elapsed time between an engineer joining a team and reaching full productive contribution. Sonatafy's vetted senior engineers and structured onboarding compress typical ramp time from 8 to 12 weeks down to 2 to 4 weeks.
S
SDLC
Software Development Life Cycle, the end-to-end process of planning, designing, building, testing, deploying, and operating software. Modern AI-in-delivery practice instruments AI across all 7 SDLC phases.
Software Delivery Done Right
Sonatafy's operating slogan. Captures the commitment to predictable delivery, full accountability, and senior LATAM engineering as a managed system rather than a staffing transaction.
Sprint
A fixed-length iteration (typically 1 to 2 weeks) during which a team commits to and completes a defined set of work. The sprint is the basic unit of cadence in Scrum and most agile delivery models.
Sprint Velocity
The amount of work a team completes in a single sprint iteration, measured in story points or completed items. Velocity is a lagging vanity metric in isolation, true delivery health requires tracking lead time, cycle time, and deployment frequency together.
Staff Augmentation
The integration of senior external engineers (8 to 10+ years experience) directly into an existing engineering team. Unlike traditional staffing, Sonatafy's model includes vetting for production readiness, communication skills, and a 30-day satisfaction guarantee.
Learn more →T
Talent Marketplace
A two-sided platform that connects companies with freelance or contract engineers (matching, contracting, payments). Marketplaces transfer screening risk to the buyer and provide no delivery accountability, fundamentally different from a managed delivery model.
See the comparison →Technical Debt
The implied cost of rework caused by choosing an expedient solution now over a better solution that would take longer. Unmanaged technical debt compounds, eventually consuming the majority of engineering capacity.
Throughput
The number of work items a team completes per unit of time (typically items per week or per sprint). Unlike velocity, throughput is normalized across teams and is harder to game.
Trunk-Based Development
A source-control workflow where developers integrate small, frequent changes directly into a single shared branch (the trunk), typically behind feature flags. Strongly correlated with elite DORA performance.
V
Vector Database
A database optimized for storing and querying high-dimensional embeddings (numerical representations of text, images, or other data). The retrieval layer underneath most production RAG systems.
Velocity Trap
The anti-pattern of optimizing for the velocity number rather than for delivered outcomes. Teams in the velocity trap inflate estimates, ship low-value work, and avoid hard problems to keep the chart up and to the right.
W
Work in Progress (WIP)
The number of items actively being worked on at any moment. High WIP correlates with longer cycle times, more context switching, and lower throughput. Limiting WIP is a core lean principle.
Have a Term You Want Defined? Let Us Know.
Our team publishes new frameworks and definitions based on what we see across 60+ delivery engagements.