Sprint Cadence & Sustainability

Trovella's planning cadence is adapted for a solo founder building part-time alongside a full-time job. The system prioritizes sustainability and capacity learning over commitment-based sprints.

Planning Units

Milestones (Primary)

Milestones (Linear Projects) are the primary planning unit. Each milestone defines an outcome: "Auth complete," "Core feature done," "Beta-ready." Milestones have target dates and child issues. Release communication is tied to milestones, not weekly sprints.

Weekly Cycles (Secondary)

Weekly Cycles (Monday--Sunday, matching the weekend work pattern) are data-collection check-ins. Issues are pulled from milestone backlogs into each week's cycle. There is no guilt on low-output weeks -- the data is for learning capacity, not accountability.

Daily Focus (Tactical)

The "Big 3" daily system: identify three things that would make the day a success. This is a personal prioritization tool, not a tracked metric.

Weekly Planning Session

When: Sunday evening or Monday morning, 30--45 minutes.

Steps:

1. Review last week's Cycle -- what shipped, what carried over, what was blocked
2. Pull issues from milestone backlogs into the new Cycle
3. Flag blockers and dependencies
4. Prep specifications for morning agent dispatch
5. Run 15-minute sustainability check-in (see below)

Throughput Tracking

Throughput is tracked for capacity learning, not performance measurement:

Metric	Source	Purpose
Issues closed per week	Linear Cycles	Primary velocity signal
Trailing 4-week average	Linear data	Forecasting milestone completion
Milestone % complete	Linear Project progress	Phase-level progress
Session and token data	ccusage CLI (open source)	Understanding AI cost and efficiency

No story points. Just count issues. Issue sizes are roughly calibrated by convention (one week's scope per feature issue), making counts a reasonable velocity proxy.

No burndown charts. Trailing 4-week average divided into remaining work provides the same timeline estimate with less overhead.

First six weeks are calibration. Do not commit to milestone dates until 4--6 weeks of throughput data exists. AI accelerates some tasks 5x and barely helps others.

Monthly Self-Assessment

Once per month, 15 minutes:

• Completion rate trend (improving, stable, declining)
• Quality trend (are CI failures increasing?)
• Energy level (sustainable, stretched, depleted)
• Business impact (is the work moving toward revenue?)
• Capacity learning (what types of tasks take longer than expected?)

Update milestone target dates based on trailing 4-week velocity.

Deployment Cadence

Continuous deployment. Every merged PR deploys via the CI/CD pipeline. No batched releases. Feature flags (Statsig + OpenFeature) control user visibility. This removes "release day" pressure and makes each deployment small and low-risk.

Changelog is deferred until beta users exist. A simple CHANGELOG.md will be added when external communication of changes is needed.

Sustainability Guardrails

Burnout prevention is treated as infrastructure -- guardrails are established on day one, not added after symptoms appear. 72% of founders report mental health impacts; 65% of startup failures stem from founder burnout. These numbers are higher for founders stacking a side project on a full-time job.

Firm Commitments

Rule	Detail
Friday evenings off	Non-negotiable
One additional weekday evening off	Aligned to recurring activity with wife
11pm hard stop	All weekday and Sunday evenings
Weekend family time	At least one weekend day includes a 6--10 hour family activity block, zero SaaS work
Ultradian blocks	90 minutes focused work, 15-minute break, then remaining time. Not 3 hours straight.

AI-Assisted Sustainability Check-In

Built into the weekly planning session (15 minutes). The check-in reviews tracking data (schedule adherence, hard stops followed, evenings off taken, family time logged) and asks structured questions to identify Stage 2 burnout signals:

• Skipping exercise
• Dreading opening the laptop
• Declining social invitations
• Sleep quality degradation
• "Just until MVP launches" rationalization

If 2+ signals are detected: Reduce scope, take 2--3 days fully off, reassess milestone timelines.

Good Week / Hard Week Flexibility

Pre-defined low-energy task list for weeks when energy is low: PR reviews, triage, specs, docs, refactoring. Not every week needs to produce new features. Low-energy weeks still produce value by reducing future friction.

The AI Paradox Rule

AI-saved hours are not automatically filled with more work. At least some of the time saved by agent automation goes to rest and family. The goal is sustainable pace, not maximum output.

Scope Control

Milestones define what is in scope. The rule: resist adding more work until the current milestone ships. Infinite backlogs create infinite stress. When a good idea surfaces mid-milestone, it goes into a "next milestone" bucket, not the current one.

Cross-Domain References

• Agent Workflow (SPTI) -- the two work modes that define the weekly rhythm
• Phase Approach -- the multi-week phases that milestones decompose
• Product Vision -- what the milestones are working toward
• Conventions -- ticket format and commit conventions