Annual planning, sprint planning, weekly reviews — most planning systems pick one timeframe and grind. The frame I actually use is three calendars stacked together: year, season, sprint. The trick is keeping them in conversation with each other, not in competition.

Calendar 1 — the year

Once a year, in January, I write a single document: the thesis for the next 12 months. Not OKRs, not a list of goals. A thesis: “This is the year I scale Raw Studios past 10 staff and ship a body of music that earns the next chapter.” One paragraph. The thesis doesn’t change for 12 months no matter what happens. Big bets I’d be willing to bet a year on.

Calendar 2 — the season

Every quarter, three months at a time, I write a season plan. Three to five outcomes that — if they happened — would meaningfully advance the year-thesis. The season plan can shift based on what the previous quarter taught me. Mine for Q2 2026 was: “Ship the new website. Release X. Close the second MLO retainer worth $200K+. Rest two full weekends.”

Calendar 3 — the sprint

Every two weeks, the sprint plan: what specifically gets done in the next 10 working days, with names and deadlines. Sprint plans live in the same Notion as the team’s. They get reviewed every Monday and closed out every other Friday. Anything not finished moves to the next sprint or gets killed.

Why three layers, not one

Single-layer systems all fail in the same direction:

  • Year-only drifts. By March you’ve forgotten what January-you was thinking.
  • Sprint-only wins battles, loses war. Two-week velocity with no compass.
  • Quarter-only is the most popular and the worst — too short to be a thesis, too long to feel real.

The three layers create conversation between scales. Sprints serve seasons, seasons serve the year, the year is the why. Every Friday review asks: did this week move the season? Is the season still serving the thesis?

The artefacts I actually keep

  • A one-page year doc, written in January, re-read every Sunday
  • A one-page season doc, written quarterly, re-read every Monday
  • A sprint board with named tasks and dates

Three artefacts. Total maintenance: ~30 minutes per week.

What broke for me before this

I used to plan in single-layer modes — annual goals or sprint backlogs. Both eventually drifted into noise. Running five companies across music, voice, film, AI and fashion forced the discipline. The three-calendar system is the only thing I’ve found that holds across that breadth without collapsing.

The quiet superpower

This system makes saying no easier. Any opportunity gets evaluated against the season and the year. “Does this serve the season? Does the season serve the year?” The answer is usually no, and the no is fast and confident.

Try it

Spend 90 minutes this Sunday writing one paragraph for each layer for the rest of 2026. Don’t optimise; just write. Update them on a fixed cadence. The compound effect is enormous.

Want to swap planning docs? Send yours, I’ll send mine.