Five Stages. Working Software Every Two Weeks.
Discovery, Scope, Build, Launch, Operate. Written requirements before code. Demos every sprint. Fixed-price scope. 30 days of operational support included.
Five Stages
The Five Stages, in Depth
Every engagement runs this process. The duration of each stage scales with project size; the structure doesn't change.
Discovery
1–2 weeksWe sit with the people who'll use the system, map the workflows, and identify the genuine bottleneck (often not where you think it is). Most failed software projects fail at this step — we don't skip it to look fast.
Deliverables
- System + data inventory (every tool, every data shape, every owner)
- Workflow maps showing current state + proposed state
- Identification of must-haves vs nice-to-haves vs can't-be-there-on-day-ones
- Risk register: technical risks, vendor risks, data quality risks
Your time commitment
60-90 min of interviews with 3-5 operators. We schedule, you show up.
Scope + Architecture
3–5 daysWritten requirements doc, architecture decision record, and fixed-price proposal. You review and sign off before code starts. Scope changes after this point get a written change order — no quiet rate creep.
Deliverables
- Requirements document (the thing you sign)
- Architecture decision record: why this stack, why this data model, why this auth approach
- Diagrams: data flow, system architecture, error paths
- Fixed-price proposal with timeline + milestones + payment schedule
Your time commitment
60 min review + sign-off. We answer questions live; you don't need to research.
Build
Sprints of 2 weeks eachProduction code, not prototypes. Every sprint ends with a working demo of what shipped that sprint. You see real progress weekly, not a black box for 90 days. Code reviewed, tested where testable, deployed to your staging environment per sprint.
Deliverables
- Working software in your staging environment, end of each sprint
- Live demo (recorded if you can't make it)
- Sprint notes: what shipped, what's next, blockers, decisions
- Pull requests in your repo for your team to review (if you have one)
Your time commitment
30 min demo per sprint + ad-hoc Slack/email questions as they come up.
Launch
1 weekStaged rollout: internal users first, then a pilot cohort, then full release. Each stage catches issues before they hit everyone. Feature flag controls let us dual-run new behavior alongside old behavior during transition.
Deliverables
- Production deployment via your CI/CD (or ours, if you don't have one)
- Monitoring dashboards live before users hit the system
- Runbook: what to do when something breaks, who to call, how to roll back
- Cutover plan: dual-run period, success criteria, rollback triggers
Your time commitment
60 min cutover review + on-call awareness during launch window.
Operate
30 days included + optional retainerProduction launch is when most agencies disappear. Ours includes 30 days of free operational support to catch the edge cases that only surface under real load. After that, your team operates it — or we stay on optional retainer.
Deliverables
- Daily Slack channel monitoring for the first 30 days
- Bug fixes for any issues caused by our code, free, for 30 days
- Optional retainer for ongoing iteration ($X/month negotiated per engagement)
- Full hand-off documentation: how to run locally, deploy, operate, debug
Your time commitment
Whatever's needed when issues come up. Usually quiet.
Cadence
Your Weekly Rhythm
What you'll see from us between sprints. Designed to keep you informed without eating your week.
Sprint demo
Every 2 weeks (Friday afternoon)30 minutes. We show what shipped, you ask questions, we capture next-sprint priorities.
Sprint notes
End of each sprintWritten summary in your shared Notion / Linear / Google Doc: what shipped, what's next, blockers, key decisions.
Slack / async
ContinuousShared Slack channel for fast questions. We respond inside business hours, usually within 2-4 hours.
Architecture decisions
When they come upBig tradeoffs (which DB, which auth, which deploy target) get a written ADR before we commit. You see the reasoning, not just the choice.
Status update to the budget owner
WeeklyBrief one-page update for the person paying the bills: percent through scope, days to milestone, any flags. Skippable if your team already gets updates via demos.
Standards
The Quality Bar
What you can expect on every engagement, regardless of scope. These aren't upsells.
Tests
Unit tests for business logic, integration tests for the critical paths, regression tests for production bugs. Not 100% coverage — coverage of the things that would hurt to break.
CI/CD
Buildkite or GitHub Actions. Tests gate deploy. Migrations run in CI. You can ship Friday afternoon and sleep Friday night.
Monitoring
Sentry for errors, custom dashboards for the business metrics that matter. Alerts on the right severity threshold, not on every blip.
Documentation
Architecture decision records, operator runbooks, hand-off guides — written for your in-house team or the next firm. Not just for us.
Security
Secrets in a real secret manager (Doppler / Vault), not in source. Auth set up correctly the first time. PII handled appropriately. Audit logs where they matter.
Code review
If you have a dev team, your team reviews our PRs. If you don't, we self-review with code-review checklist + automated linting. Either way, code reaches main reviewed.
Tooling
The Tooling We Use
We adapt to your existing tools when you have them. When you don't, these are our defaults.
Project tracking
- Linear
- Notion
- GitHub Projects
- or your existing tool
Communication
- Slack (shared channel)
- Loom (async video)
- Zoom for demos
Code
- GitHub (your repo)
- Pull requests + reviews
- Conventional commits
Infrastructure
- Your existing cloud (DigitalOcean / AWS / GCP / Vercel)
- Doppler for secrets
- Buildkite / GitHub Actions CI
Differentiation
Why Purcell
How we compare against in-house teams, big consultancies, freelancers, and no-code platforms.
Pricing
Pricing
Fixed-price scope per service. What each tier includes and what it doesn't.
Services
All 7 Services
AI, system integration, process automation, custom development, analytics, NetSuite, programmatic SEO.
Ready to Start Stage 1?
Schedule a 30-minute discovery call. You'll get a written assessment within a week — including whether we're the right call for your project.