Client weeks will eat your products if you let them. I learned that the hard way while shipping Shopify work by day and trying to grow thin tools at night.
The fix was not “wake up earlier.” It was a weekly review for side projects that fits in one hour: look at metrics, ship one thing, cut one thing. That is how MetaGrid, BulkFix, and VisaChamp-shaped ideas stay alive when invoices and kickoffs stack up.
Why side projects die on consulting calendars
It is rarely a lack of ideas. It is context switching. Monday you sketched a metafield bulk editor. Wednesday a Plus migration blew up. Friday you opened the repo and felt behind, so you closed it.
Without a ritual, guilt replaces progress. Guilt does not merge PRs.
The one-hour shape
I block it the same way every week — usually Friday afternoon Karachi time, or Sunday night before the client week starts. Phone down. One doc. Timer on.
Minutes 0–15: metrics without theatre
I only track numbers I will act on:
- Signups or waitlist adds (if any)
- Active use this week (even if N is small)
- Support or feedback themes (three bullets max)
- Revenue if there is any — no vanity MRR storytelling
If a metric did not change and I will not change behavior, I stop staring at it. Fake dashboards are how you burn the hour.
Minutes 15–35: one ship
Not a roadmap. One ship. Examples that count:
- Fix the confusing empty state
- Publish the pricing FAQ
- Add export for the one format a user asked for twice
- Write the README section future-you needs
- Ship a landing paragraph that tells the truth
I define “ship” as something a user can see or something that reduces my ops load. Refactors only count if they unblock the next ship.
Minutes 35–50: one cut
This is the part people skip. Cut means:
- Delete a half-built feature branch you will not finish this quarter
- Remove a pricing tier nobody chooses
- Stop a channel that brings curiosity clicks and zero activation
- Say no to a custom request that would turn the product into a services firm
Cutting creates the space the next ship needs. Without it, the backlog becomes a museum of good intentions.
Minutes 50–60: schedule the next wedge
I write one sentence for next week: “Ship X so that Y.” If I cannot finish the sentence, the idea is not ready. I also note any client conflict (travel, launch week) so I do not pretend the hour will magically double.
Rules that keep the review honest
- No opening Twitter “for inspiration” inside the hour.
- No rewriting the vision doc every week.
- Client fire drills can steal a day; they do not automatically cancel the review. Move it, do not delete it.
- If I skipped two weeks, the review starts with a cut, not a pep talk.
How this interacts with client work
I steal patterns from products into proposals and the other way around. Bulk metafield pain showed up in client admin tooling and in product sketches. Document-QA work informed how I refuse empty retrieval in support bots. The weekly review is where I notice the repetition and decide: thin tool, client playbook, or ignore.
I do not wait for a platform. I wait for the third time I swear at the same problem — then the review’s “one ship” is often a tiny public utility or a scoped MVP.
When the hour says stop
Sometimes the honest review is: park the project. Write the reason. Archive the repo note. That is still a successful review. Zombie side projects tax attention worse than a clean pause.
Other times the hour says: raise price, narrow ICP, or turn on a waitlist. Those are product decisions, not grind decisions.
What “metrics” look like when N is embarrassing
Early products do not have pretty charts. That is fine. My metrics block has looked like:
- “7 people used the formatter this week; 2 came back”
- “3 support emails, all about CSV encoding”
- “0 revenue; 1 inbound that asked for a custom feature we will not build”
Honesty beats vanity. If nobody returned, the ship for next week is usually onboarding or clarity — not a new module. The weekly review is where I admit that without turning it into a brand crisis.
Pairing the review with client boundaries
I keep a hard rule: client emergencies can move the review; they do not own the entire weekend by default. Writing that into my own calendar notes sounds soft. It is how MetaGrid-shaped work survives. When a launch week is real, I pre-announce a pause in the previous review (“no ship this Friday; cut the experimental branch instead”) so I do not carry guilt plus fatigue.
Shipping in public without oversharing clients
Some ships are public changelog lines. Some are private refactors. The review does not require a LinkedIn post. When I do write publicly, I talk about the problem class — bulk metafields, document QA, API cost — not confidential merchant data. The playbook exists so the learning compounds without burning trust.
A failure mode: the productive-feeling fake review
Reorganizing Notion, rewriting the vision, and renaming the repo feel like care. They are not ships or cuts. If my note from last week is only taxonomy changes, I failed the ritual. The timer helps: fifteen minutes of metrics, then a diff that users can feel.
Quarterly cousin of the weekly review
Once a quarter I zoom out: is this product still the third repetition of a pain I charge for? Did client work move on so the wedge is stale? Keep, kill, or merge into a service offering. The weekly review keeps motion; the quarterly review keeps honesty about direction.
Tools I keep open during the hour
A terminal on the repo, the production metrics or a simple analytics query, and the notes doc. That is it. No Figma rabbit holes unless the ship is literally a UI empty state. Constraining tools constrains scope.
I also keep a running “idea parking lot” outside the review. Ideas are welcome; they do not get airtime inside the hour unless they beat the scheduled ship. Parking lots prevent the review from becoming a brainstorm that produces nothing mergeable.
If a client week destroyed deep work, the ship can be documentation or a pricing clarification. Small still counts. The streak matters more than the hero feature — especially when you are building from Karachi beside billable delivery.
A template you can copy
Title the note with the date. Three headings: Metrics, Ship, Cut. Four lines under each, maximum. Link the PR or commit. Stop.
If you are a consultant building in public from a timezone that does not overlap your buyers, the weekly review is how you stay a builder without pretending you have a venture-backed calendar. One hour. One ship. One cut. Repeat until the product either earns a real slot or earns a graceful archive.