· fractional-coo · claude-code · operations · consulting · workflow
How a fractional COO uses Claude Code on Sunday afternoons
A real Sunday-afternoon prep cadence for a multi-client fractional operator: three clients, a follow-up queue, an invoice chase, a fresh SoW to draft.
Sunday at 2pm. Coffee, a closed door, and a week ahead with three retainer clients, a stalled prospect who needs a fourth follow-up nudge, an invoice that has gone twenty-eight days past due, and a fresh statement of work to draft for a kickoff on Tuesday. This is what the prep block looks like for one fractional COO we work with, mapped to the actual cadence rather than the toolchain.
Names anonymised, numbers real. Two and a half hours start to finish, give or take. The point of writing it out is not the heroism of a Sunday work block; the point is that without the prep block, Monday morning becomes a triage session that eats the first client’s strategy hour. Sunday afternoon buys back Monday morning.
Inventory the week first
The first fifteen minutes are not Claude Code. They are a notebook and a pen. List every client, every open thread, every artefact owed back. Today’s list:
- Client A (Series B SaaS, two days a week): Q2 OKR review on Wednesday, hiring panel Thursday, a contractor offboarding email pending since Friday.
- Client B (consumer brand, one day a week, retainer): monthly board pack due Tuesday, two team 1:1s on Monday, the founder wants a view on a candidate CV by tomorrow.
- Client C (early-stage, half a day a week): kickoff Tuesday for a new SoW that hasn’t been drafted yet.
- Stalled prospect: discovery call three weeks ago, three follow-ups sent, no reply. Needs either a final nudge or a walk-away.
- Overdue invoice: a past retainer client, twenty-eight days past due. T+30 escalation tier.
Inventory is the bit that breaks if you skip it. Every Sunday block we’ve watched go sideways started with the operator opening a tool before they’d written the list. The list is the input. The tools just execute against it.
Draft the SoW first because the calendar forces it
The Tuesday kickoff is the only thing on the list with a hard external deadline. It goes first.
The proposal for Client C went out a week ago and came back accepted on Friday with two minor scope tweaks. The job now is to convert that proposal into a Statement of Work: same structural fields (parties, scope, deliverables, fees) but with the contractual gaps the proposal didn’t cover, including trading entities, governing law, IP ownership, termination, liability cap, and acceptance criteria for every deliverable. SoWs are where most engagements quietly go wrong, because the deliverables get described loosely enough that “done” becomes a matter of opinion later.
Run sow-templater against the accepted proposal. The skill inherits the structure from the proposal, asks for the six contractual gaps in a single grouped question, and writes explicit acceptance criteria deliverable by deliverable. Twenty-five minutes including a careful read-through. Default length is 4–6 pages, deliberately not enterprise boilerplate. Send the draft to Client C with a short cover note: “signing version Tuesday morning, flag anything before then.”
The reason this step matters: the operator who skips this and turns up to a Tuesday kickoff with a verbally-agreed scope and no signed SoW is the operator who, in week six, has a polite disagreement about whether the dashboard rebuild was in scope. We’ve watched this play out enough times to make it the second-most important skill in the Solo Consultant Ops pack. Proposals win the work; SoWs protect it.
Clear the follow-up queue
Three follow-ups owed. A stalled prospect. A check-in for Client A’s CFO who flagged a concern in Wednesday’s board prep last fortnight. A testimonial ask for a former client whose engagement closed cleanly six weeks ago.
These are different shapes of the same problem: a short, specific email that references a moment from the prior interaction and ends with one clear action. They are not different in cadence; they are different in tone, target, and what counts as a useful response. The mistake most operators make on Sundays is batching them as if they were the same job and writing three near-identical openings.
follow-up-sequencer handles the differentiation. Each output references a specific moment from the prior interaction, stays under 100 words, and ends with one clear action. The walk-away rules are baked in: there is no chase email four. If the stalled prospect doesn’t respond to the third nudge, the right move is a clean walk-away message that leaves the door open without begging. The skill writes that one too. It will not write a fourth.
The testimonial ask is the most under-done of the three. Most consultants don’t ask, because asking feels needy. The skill produces a request that opens with the specific outcome from the engagement and asks for a one-paragraph quote: not a case study, not a video. One paragraph. The reply rate on a tight, specific testimonial ask is materially higher than on a vague one. We’ve watched the difference enough times to make it a default mode in the skill rather than an afterthought.
Twenty minutes for all three. The bottleneck is not drafting; it is remembering the specifics of each prior interaction, which is why running this on Sunday with the inventory list in front of you matters.
The invoice chase nobody wants to write
Twenty-eight days past due. T+30 tier. The previous chase emails went out at T+0, T+7, and T+14, each escalating in tone. The next one is the formal final notice citing the SoW’s late-payment clause and stating a specific consequence the operator is actually willing to execute.
This is the email everyone writes badly because it’s emotionally costly. The two failure modes are equally common: too soft, in which case the email doesn’t move the invoice; or too aggressive, in which case the relationship is damaged for no gain. The right tone is firm, professional, and grounded in the contract.
invoice-chaser picks the tier from the days-overdue figure and refuses to skip tiers, because skipping a tier weakens any subsequent legal position. It also refuses to bluff consequences the operator wouldn’t actually execute. “We will refer this to a debt collection agency” is not a sentence to write unless you’d actually do it; the skill pushes back if the operator tries.
Ten minutes. Send. Then the relationship-recovery follow-up, queued for after the payment lands, drafted now while the context is fresh. The relationship-recovery sequence is the bit that stops a late-payment incident from becoming a lost client; it resets the working relationship without litigating the lateness. Two emails over a fortnight after payment, neither of which mentions the chase.
Why the cadence matters more than the tools
The whole block runs about two hours and twenty minutes today: inventory (15), SoW (25), three follow-ups (20), invoice chase plus recovery sequence (15), board pack notes for Client B’s Tuesday deadline (35), buffer (30) for the things that always come up: a Slack message that needs a thoughtful response, a candidate CV review for Client B’s founder, fifteen minutes spent staring out of the window because the week starts tomorrow.
The skills are scaffolding. They make the cadence possible by removing the decision cost of how to write each artefact. The cadence itself (inventory, then deadline-ordered execution, then the emotionally-costly task last while the easier ones build momentum) is the operator’s craft. The pack does not replace that. It removes the friction that stops it being sustainable on the second Sunday in a row.
The same week with no Sunday block looks different. Monday morning becomes Client A’s strategy hour with a side of triage. The Tuesday SoW gets drafted in the cab on the way to the kickoff. The invoice chase doesn’t get sent at all this week, slipping to T+35 and weakening the position further. The testimonial never gets asked.
For a structural look at which Claude Code skills are worth writing for solo consulting practice and which aren’t, see the real list. For the proposal-writing failure mode that produces most of the work that lands on Sunday afternoons in the first place, see stop writing the same proposal six times a year.
What this looks like with Solo Consultant Ops
The four skills above (proposal-writer, sow-templater, follow-up-sequencer, invoice-chaser) ship in the Solo Consultant Ops pack for £99 lifetime, no subscription. The pack is the prep stack for the Sunday block. It does not replace the block; it makes it survivable for the operator who has to run it forty Sundays a year.