Day 4: Publishing From a Pub / Chee Ann

I shipped most of today's work from a pub. Laptop open, drink in hand, pushing commits between conversations. There's something funny about that. Four days into a shipping experiment and I'm deploying blog infrastructure from a bar stool.

The posts I finally put out

I've had a case study sitting in my head for weeks. A Google Ads automation project where I handed off 37 Python scripts to an ads manager who'd never used a terminal. Real numbers, real handoff, real results. I just... never wrote it up.

Today I did. Published it on both grooveconsulting.org and here.

There was also a personal piece that came out of today. About watching my husband perform on stage and what it feels like to build quietly while someone you love performs publicly. Wrote it and published it same day.

Two posts in one day. One technical, one personal.

Building the pipes

Writing is one thing. Making it findable is another. I realized my blog had no distribution infrastructure at all. No social previews, no RSS, no share buttons. Just... posts sitting there.

So I built:

Dynamic OG images. Every post now auto-generates a social card. No more blank thumbnails when I share on LinkedIn.
RSS feed. /feed.xml works now.
Share buttons. Twitter, LinkedIn, copy link on every post.
Ship30 progress bar. Shows 4/30 on the homepage.
Draft protection. Found that unpublished posts were accessible if you knew the URL. Oops. Fixed.

Also tried adding email subscriptions via follow.it, realized it wasn't right, ripped it out same day. Going with Kit (ConvertKit) instead.

Poking at SEO (still clueless)

I should be honest about this: I don't really know SEO. The team I work with does. They've been doing it for years. I'm just the person trying to figure out if AI tools can help speed up the tedious parts.

So I built a reusable /seo-audit command for Amp that scrapes a URL and checks meta tags, heading structure, that kind of thing. Whether the output is actually useful... I'm still learning. Set up GA4 and Google Business Profile API connections too.

I can run the scripts. I can pull the data. I just don't always know what to do with it yet. The SEO team does, which is the whole point. I'm building tools for people who know more than me.

Company site polish

Ran a design audit on grooveconsulting.org. Added a "Latest Insights" section, tag pages, Article schema. Then did a visual QA pass and found six things wrong:

Heading sizes inconsistent
Nav and CTA touch targets too small for mobile
Font flash on page load
Blog grid breaking on tablets
Timestamps showing when just the date would do

The kind of stuff you only notice when you stop building and start actually using the thing.

What I shipped

cheeann.com (11 commits):

Published case study + field note + Day 3 log
OG images, RSS feed, share buttons, progress bar
Draft URL protection, slug redirects, timezone fix

Company site (10 commits):

Case study published
Latest Insights section, tag pages, Article schema
6-finding design review

SEO automation (6 commits):

/seo-audit skill (still learning what good output looks like)
GA4 + Google Business Profile API setup
Client deliverables and team onboarding guides

TikTok harness (2 commits):

Config cleanup, output tracking

What's shifting

Days 1-3 were about building things. Day 4 was about telling people they exist.

I can sit in a terminal all night pushing commits nobody sees. That part comes naturally. The part that doesn't: putting my name on something and saying "I made this, and here's what it did."

Still figuring that out. But at least today I started. From a pub.

The hardest commit wasn't code. It was hitting publish.

Running total

98 commits, 7 repos, 4 days
2 blog posts published (case study + field note)
2 websites with design polish
RSS feed, OG images, share buttons shipped
0 blockers