Our Delivery Framework
SPARK™ Framework
Structured Delivery That Works
SPARK™ is Salt's delivery framework for building, modernizing, and scaling software, data, and AI systems using cross-functional Pods. Five stages that bring discipline to agile without the bureaucracy.
The 5 Stages
From Alignment to Continuous Improvement
Each stage has clear goals, activities, and deliverables. No ambiguity, no surprises—just structured progress.
Scope & Shape
Week 1-2
Goal: Align on what we're doing and why.
Before writing a single line of code, we ensure everyone is on the same page. We dive deep into understanding your problem, context, stakeholders, and constraints. This stage sets the foundation for everything that follows.
Key Activities
Problem & Context Analysis
We work with your team to understand the core problem you're solving, the business context, and who the key stakeholders are.
Define Success Metrics
What does success look like? We establish clear, measurable outcomes so everyone knows what we're working toward.
Identify Risks & Dependencies
We surface potential blockers early—technical constraints, third-party dependencies, compliance requirements, and organizational risks.
High-Level Solution Outline
We sketch the approach: technology choices, architecture direction, and a rough delivery roadmap.
Deliverables & Artifacts
- Problem statement document
- Success metrics & KPIs
- High-level solution outline
- Initial roadmap & priority list
- Risk register
Stage Outcome
Clear alignment on goals, approach, and success criteria
Plan & Pilot
Week 3-6
Goal: Prove fit, velocity, and working model quickly.
Instead of lengthy planning phases, we get to work. A small Pilot Pod delivers a thin slice of value so you can see how we collaborate, how fast we move, and what quality looks like—before committing to a long-term engagement.
Key Activities
Build Realistic Delivery Plan
We break down the work into sprints with clear deliverables. No waterfall estimates—real, achievable milestones.
Form Pilot Pod
We assemble a small, focused team (typically 3-5 people) tailored to your needs: developers, QA, and a tech lead.
Deliver MVP or PoC
The team ships something real—a working feature, a proof of concept, or a meaningful improvement to your existing system.
Validate Collaboration
You experience our communication style, delivery cadence, and quality standards firsthand. If it's not a fit, we part ways.
Deliverables & Artifacts
- Pilot plan & scope document
- Pod composition & roles
- Definition of done for pilot
- Working software / demo
- Pilot retrospective & feedback
Stage Outcome
Validated approach with real results and mutual confidence
Architect & Automate
Week 5-8
Goal: Make delivery repeatable, safe, and fast.
This is what separates Salt from 'just hire some devs' vendors. We invest in the foundation that enables sustainable velocity: solid architecture, automated pipelines, comprehensive testing, and built-in security.
Key Activities
Finalize Architecture Decisions
We document key decisions using Architecture Decision Records (ADRs). No tribal knowledge—everything is explicit and justified.
Set Up CI/CD Pipelines
Automated build, test, and deployment pipelines. Every commit is validated. Deployments are push-button, not prayers.
Implement Automated Testing
Unit tests, integration tests, and basic E2E tests. We target 80%+ code coverage on critical paths.
Add Security & Observability
Security scanning in pipelines, dependency vulnerability checks, logging, metrics, and alerting foundations.
Deliverables & Artifacts
- Architecture Decision Records (ADRs)
- CI/CD pipeline documentation
- Testing strategy overview
- Security & observability checklist
- Infrastructure-as-code templates
Stage Outcome
Scalable foundation ready for sustained growth
Release & Run
Ongoing
Goal: Ship and operate with confidence.
Stable, predictable releases without drama. We ship frequently in small increments, monitor everything, and handle incidents professionally. Your users get value continuously, not in risky 'big bang' releases.
Key Activities
Frequent Small Releases
We deploy multiple times per week (sometimes per day). Small changes mean small risks and fast feedback.
Monitor Performance & Errors
Real-time dashboards track system health, error rates, and user experience. We catch problems before users do.
Handle Incidents & Support
When issues arise, we respond quickly with documented runbooks. Post-mortems turn incidents into improvements.
Keep Environments Healthy
Regular maintenance, dependency updates, and infrastructure hygiene. No 'set it and forget it' mentality.
Deliverables & Artifacts
- Release notes & changelogs
- Runbooks & incident handling docs
- Monitoring dashboards
- SLA/SLO tracking reports
- Post-incident reviews
Stage Outcome
Stable, predictable releases without drama
Keep Improving
Continuous
Goal: Continuously improve speed, quality, and outcomes.
Your product and platform should get better every quarter, not just 'stay alive'. We track metrics, run retrospectives, and proactively propose improvements. This isn't maintenance mode—it's continuous evolution.
Key Activities
Track Key Metrics (DORA)
Lead time, deployment frequency, change failure rate, MTTR. We measure what matters and make it visible.
Regular Retrospectives
Every sprint ends with honest reflection. What worked? What didn't? How do we get better?
Optimize Performance & UX
Continuous profiling, load testing, and UX refinement. We find and fix bottlenecks before they become problems.
Propose New Improvements
We don't just execute—we think. The team actively suggests features, refactors, and optimizations.
Deliverables & Artifacts
- Metrics & trend reports
- Improvement backlog items
- Periodic health check summaries
- Technical debt inventory
- Quarterly roadmap recommendations
Stage Outcome
Continuous growth and optimization, not just maintenance
Why It Works
SPARK™ Delivers Predictable Results
SPARK isn't just process for process's sake. Every stage exists to reduce risk, accelerate delivery, and ensure long-term success.
Reduces Risk
The Pilot Pod validates fit before you commit. You see real results, not slide decks.
Accelerates Time-to-Value
No 3-month discovery phases. You get working software within weeks, not quarters.
Ensures Quality
Architecture, automation, and testing are built-in—not afterthoughts bolted on later.
Enables Continuous Improvement
SPARK doesn't end at launch. The 'K' ensures your product keeps getting better.
Ready to Experience SPARK™?
Start with a Pilot Pod to see how SPARK™ works in practice. Experience our methodology, meet your team, and get real results—all within 4-6 weeks.