All writing
April 10, 2026/7 min readmethodshippingportfolio

Shipping a product fleet: four products in three weeks

Cormac Kerins
Cormac Kerins

Product consultant & builder

Between February 10 and March 1, 2026, four consumer products went from empty repo to live: a property marketplace, a recruitment screening tool, a family activities guide, and a paid walking club. The commit history shows exactly how, so this piece walks through it.

The method in one line: start from a template, ship the whole thing on day one, fix forward in production, and roll every platform improvement across all products in one wave.

Timeline

Feb 10Moo-ving initial commit — full platform, first production fixes land the same day
Feb 16Vetted landing page live
Feb 198tot8 restyled and rebuilt in a single day of rapid commits
Mar 1Walk with Toogi payment landing page live
Mar 8Wave: GA4 analytics rolled across the fleet in one sitting
Mar 14–19Wave: Vercel Analytics, consent handling, public /report pages
Apr 8Wave: sitemaps, structured data, security headers, funnel events — five repos in one day

Start at 80%: templates and napkins

8tot8 starts with a commit named after the template that produced it: a Vite, React, shadcn, TypeScript starter with testing already wired in. The name in the history records which starting point produced which product. A good template means launch day is mostly styling and content.

Each repo also has a napkin file: a short note holding the product idea, its current state, and the open decisions. Commits like "docs: update napkin to reflect shipped MVP state" show how it works — the napkin changes when reality changes, and it stays short enough to read in a minute. Big spec documents go stale. A napkin stays honest.

Ship whole, fix forward

Moo-ving went live in its first commit — the whole platform. The commits that follow, hours apart on the same day, are production fixes: a lazy email client to survive a missing API key, a Next.js upgrade to clear an Edge Runtime crash, middleware brought back for session refresh. Production found the weak points fast, and each one was fixed within hours.

The bet is simple. Predicting every failure before launch costs weeks. Meeting failures in production costs hours, as long as deploys are fast and every fix is small. Vetted made the same bet: the full MVP — intake, interview, review, bilingual UI — landed as one commit on March 16, then hardened step by step.

The wave: prove once, roll out everywhere

The clearest pattern in the history is the improvement wave. March 8: GA4 analytics lands in Walk with Toogi and Moo-ving with near-identical commits. March 14: Vercel Web Analytics, fleet-wide. March 15 and 19: consent handling arrives, then gets simplified, with cookie policies aligned to actual behaviour — every product, same day. April 8 is the best example: sitemap fixes, structured data, security headers, GA4 funnel events, and a blog landed across five repos in one working session.

The reason is cost. Proving a pattern on the first product carries all the research and all the mistakes. Rolling it out to the rest costs a fraction of that, and with agents doing the rollout, each extra product costs about as much as a code review. So every product stays current on analytics, consent, SEO, and security in the same week the pattern is proven.

Platform work happens in waves. Product work stays focused on one product at a time. The two rhythms run side by side in the history and rarely collide.

Instrument early

Every product gets the same measurement setup: GA4 with funnel events, Vercel Analytics as a second opinion, and public /report pages showing live traffic. It all lands early, so when a product question comes up — which step loses people? — the data is already there.

The detail work backs it up. Toogi’s purchase events moved server-side once client-side gaps showed up. Vetted got six funnel events across intake, registration, and review in one commit. Analytics gets the same fix-forward care as payments.

The takeaway

Templates put launch day at 80%. Shipping whole and fixing forward turns prediction time into iteration time. Waves keep the whole fleet current for the price of one proof. Instrumentation lands before opinions form. This is the same discipline I bring to client work.

My time is for hire.

Two ways to work with me: a full project build, where I design, build, and ship your product with these methods, or consulting inside your team.