Hugo Martin

Friday Weekly Review That Builds Next Week's Plan With Real Math

Turn this week's done list and open loops into a capacity-checked plan for next week: triage every loose end, pick three outcomes, and block the time.

At a glance

Best for: Freelancers, managers and anyone juggling several clients or projects who wants a weekly reset that respects the hours they actually have.
Tested on: Claude · Opus 5.5
You fill in: done_listopen_loopsnext_week_calendar
You get: Wins Brightloop's Q4 calendar was delivered and invoiced in the same week. Halden is a warm lead with a clear ask and deadline. Tern posts are half… (full result below)

productivity

business-strategy

decision-making

Prompt

Run my weekly review and build next week's plan.

What I finished this week: {{done_list}}

Open loops (anything unfinished, owed or nagging me), with rough hour estimates: {{open_loops}}

Next week's availability and fixed commitments: {{next_week_calendar}}

Steps:

1. Wins: 2-3 bullets, one line each. No padding.

1. Triage every open loop into Do next week / Delegate or ask / Defer to a named date / Drop. Give a reason of under 12 words for each.

1. Capacity check: available hours minus 20% buffer = plannable hours. Sum the estimates for everything in "Do". If it doesn't fit, move items out until it does, and say what you moved.

1. Top 3 outcomes for next week, each phrased as a finished state ("proposal sent", not "work on proposal").

1. Day-by-day blocks, placing the hardest work early in the week.

1. One decision I've been postponing that this plan forces me to make.

Show the capacity math. If an estimate looks unrealistic, flag it instead of silently changing it.

Fill in

{{done_list}}: What you completed this week.e.g. Delivered Q4 content calendar to Brightloop; sent invoice...
{{open_loops}}: Everything unfinished or owed, each with a rough hour estimate.e.g. 2 blog posts for Tern Coffee due Tue (6h); Halden proposal due Wed (6h)...
{{next_week_calendar}}: Working hours per day and any fixed meetings or time off.e.g. Mon-Thu 9-5 with 1h lunch; Fri 9-12 only. Mon 10-11 client call...

Example input

done_list: Delivered Q4 content calendar to Brightloop; wrote 2 of 4 blog posts for Tern Coffee; sent Brightloop invoice; sales call with Halden Studio (they want a proposal by next Wednesday).

open_loops: 2 remaining Tern Coffee posts, due Tue (6h total); Halden proposal, due Wed (6h); feedback on Mosaic HR's newsletter draft, they've waited since Monday (1h); 3 LinkedIn posts Brightloop asked for by Fri (3h); add Brightloop case study to portfolio (3h); Mosaic invoice is 18 days overdue, chase it (15 min); reply to podcast guest invite (15 min); learn Claude Code to automate my monthly client reports (4h).

next_week_calendar: Mon-Thu 9:00-17:00 with 1h lunch; Friday 9:00-12:00 only (school event). Fixed: Mon 10-11 Brightloop call; Wed 14-15 Tern check-in; Thu 9:30-10:00 Mosaic call.

1.0k0

Generated Outputs (1)

2 hours ago

Claude

Opus 5.5

2 hours ago

Generated Output

1. Wins

• Brightloop's Q4 calendar was delivered and invoiced in the same week.

• Halden is a warm lead with a clear ask and deadline.

• Tern posts are half done, ahead of Tuesday.

2. Triage

• Do: Tern posts (6h). Due Tue, and paid work.

• Do: Halden proposal (6h). New revenue with a hard Wednesday deadline.

• Do: Mosaic feedback (1h). They've waited a week, which is on you.

• Do: Brightloop LinkedIn posts (3h). The client asked, due Friday.

• Do: chase the Mosaic invoice (15m). 18 days overdue; send it separately, not in the feedback email.

• Do: podcast reply (15m). Faster to answer than to carry.

• Defer to Mon 5 Oct: portfolio case study (3h). Useful, but nobody is waiting on it.

• Defer to Mon 5 Oct, cut to 2h: Claude Code reporting automation. It's an investment that pays off after the busy week.

3. Capacity math
Available: Mon 6h (7 minus the call), Tue 7h, Wed 6h, Thu 6.5h, Fri 3h = 28.5h
Minus 20% buffer (5.7h) = 22.8h plannable
All loops as listed: 23.5h, 0.7h over.
After deferring the portfolio (3h) and the automation (4h): 16.5h in Do, leaving 6.3h of slack. That slack will get eaten by client email; let it.

Flag: 6h for a first proposal to a new client is tight if they want pricing options. Decide early whether you're sending one price or three.

4. Top 3 outcomes

1. Tern posts delivered by Tue 12:00.

1. Halden proposal sent by Wed 12:00, a half-day early.

1. Mosaic feedback sent and invoice chased by Mon 17:00.

5. Blocks

• Mon: 9-10 Tern post 3; 11-12 Mosaic feedback + invoice chase; 13-16 Tern post 4 (draft); 16-17 podcast reply + admin.

• Tue: 9-11 finish Tern posts and send; 11-17 Halden proposal (4h), with gaps for email.

• Wed: 9-11 Halden final pass, send by 12; 14-15 Tern check-in; 15-17 LinkedIn post 1.

• Thu: 10-12 LinkedIn posts 2-3; afternoon is the buffer.

• Fri: 9-11 send LinkedIn posts; 11-12 next week's review.

6. Decision this forces
Whether Halden gets one fixed price or tiered options. Decide it Monday so Tuesday's 4h is writing, not deliberating.

Why it works

Most weekly-review prompts produce a pleasant reflection and an overstuffed to-do list. This one has a hard constraint in step 3: available hours, minus a 20% buffer, compared with the summed estimates. The plan has to fit or something visibly moves out.

Forcing every open loop into Do/Delegate/Defer/Drop stops loose ends quietly rolling forward. "Defer to a named date" stops "later" from meaning "never". Outcomes phrased as finished states make Friday's review checkable. "Flag it instead of silently changing it" keeps your estimates yours. The final "one decision" line surfaces the thing the list has been hiding.

When not to use it

It assumes you can estimate hours roughly. If your week is mostly reactive (support rotas, on-call, retail shifts), a capacity plan will be fiction, so review the week but skip the time-blocking. It also can't see your real calendar unless you paste it, and a plan built from a half-remembered calendar will collide with meetings you forgot.

Comments (0)

No comments yet. Be the first to share your thoughts!