Content Pipeline Lessons: From One Article to 240

When you publish one article a week, the pipeline can be whatever. When you publish one article a day, the pipeline is the product. Eight months of daily publishing forced me to build and rebuild my content pipeline multiple times. Here are the lessons that stuck.

The Pipeline Stages

Every article moves through five stages:

1. Capture: idea lands in a notes file
2. Draft: write the body, usually in 25 minutes
3. Edit: one pass for clarity, one for em dashes
4. Publish: POST to the backend, verify on the frontend
5. Review: check in the week after for errors

Skipping any stage breaks the pipeline. Capturing without drafting loses ideas. Drafting without editing publishes typos. Editing without publishing wastes the work.

The Artifact Script Win

The single biggest pipeline improvement was switching from hand-posting articles to running artifact scripts. A Python script with an array of article objects and a loop that POSTs each one.

def publish_batch(articles):
    for article in articles:
        post_article(article)
        verify_published(article['slug'])

Twenty articles in ninety seconds. Hand-posting took hours. The script is not glamorous but the time back is real.

The Category Problem

I started without a category taxonomy. By article 50 I had invented categories inconsistently. I spent a day retrofitting a taxonomy and re-categorizing everything. Now the categories are fixed, documented, and every new article is tagged from the defined list.

Lesson: define the taxonomy before the first article, not after fifty.

The Backup Story

Everything lives in git. Articles are in a database but also mirrored in markdown files in the repo. If the database dies, the content survives. If the site dies, the content still lives in the repo. Two failure modes, one backup strategy.

What Did Not Work

• Automated publishing on a schedule (I want to review before publish)
• AI-generated articles end to end (quality was uneven, voice was wrong)
• Collaborative editing tools (overkill for a solo writer)

What Is Left to Improve

I still do not have good analytics on which articles drive subscribers. That is the next pipeline upgrade: attribution from article to conversion.

A content pipeline is not glamorous infrastructure. It is infrastructure anyway. Build it once, improve it gradually, respect the compounding.

The Publishing Ritual

Every morning around the same time, I sit down, open the ideas file, pick a topic, write, edit, publish, verify. Thirty minutes end to end on a good day, ninety on a bad one. The ritual mattered more than any specific tool. Routines beat motivation. Habits beat willpower.

Version Control as Safety Net

Every draft, every edit, every published version goes through git. If I want to know what an article looked like three months ago, git blame answers in seconds. That safety net let me take risks with experiments I would not have taken otherwise.

For headless CMS pipeline patterns, see the Strapi content pipeline documentation.