To manage multiple side projects, stop treating every project as equally active. Give each project a portfolio role, choose one weekly focus, limit work in progress, and review evidence before allocating another week.

The goal is not to touch every project. It is to move the portfolio toward durable value.

Give every project a role

A solo portfolio usually contains several kinds of work:

  • Core product: the clearest path to customers or revenue.
  • Experiment: a bounded test of demand, technology, or distribution.
  • Infrastructure: shared capability that lowers cost across products.
  • Content or channel: an asset that compounds discovery and trust.
  • Maintenance: stable work that protects existing users.
  • Archive: a project intentionally kept but not actively funded with time.

Roles make tradeoffs explicit. An experiment should not consume core-product reliability time without new evidence.

Choose one weekly focus

Pick one project that receives the best uninterrupted hours. Other projects may receive maintenance or a small unblock, but they do not compete for equal focus.

Choose with evidence:

  • customer urgency;
  • revenue or adoption signal;
  • blocked release;
  • learning value;
  • reuse across projects;
  • cost of delay;
  • unresolved risk.

A dashboard that shows everything but cannot name the next action is inventory, not strategy.

Limit active outcomes

Each active project gets one meaningful outcome, not a long list.

For example:

  • Core product: reduce first-use setup below five minutes.
  • Experiment: interview five users who abandoned onboarding.
  • Infrastructure: make one shared release path reliable.
  • Content: publish the guide supporting a search opportunity.

If a new urgent outcome enters, pause another. Unbounded work in progress turns AI speed into more context switching.

Track movement with evidence

Use signals that reflect the project's role:

  • core product: activation, retention, revenue, support;
  • experiment: hypothesis tested and decision reached;
  • infrastructure: consumers adopted and failures reduced;
  • content: impressions, qualified visits, signups, backlinks;
  • maintenance: incidents, freshness, dependency risk.

Do not compare unlike metrics as one universal “score.”

Look for capability compounding

The strongest portfolio work helps several projects:

  • one authentication integration;
  • one reliable deployment path;
  • one content distribution system;
  • one verified agent workflow;
  • one reusable design pattern.

But shared infrastructure is valuable only when real products consume it. Otherwise it is another side project wearing an architectural label.

Run a Friday portfolio review

Ask:

What moved?

Which user, delivery, or learning result changed?

What stalled?

Was the blocker direction, implementation, distribution, or attention?

What compounds?

Which work benefits more than one active product?

What stops?

Which project or task will receive no time next week?

What is next?

Name one primary project and one outcome.

When to pause a project

Pause when:

  • the hypothesis has been disproved;
  • no target user will commit;
  • the project duplicates a stronger path;
  • maintenance exceeds strategic value;
  • another project has materially better evidence;
  • the next step is only “keep building.”

A pause is a portfolio decision, not a personal failure. Preserve the state and reason so the project can be evaluated later without rediscovery.

Use AI to reduce coordination, not increase it

AI agents can advance several projects quickly, but a solo developer still owns prioritization and review. Keep project state, evidence, and next actions visible in one portfolio view; let agents execute bounded work inside each project.

The core principle

A healthy solo portfolio has many options but few active bets. Concentrate attention, demand evidence, reuse what compounds, and stop funding work that cannot name its next user result.