I Automated Half My Business With 200 Lines of Python -- Here's What I Learned

The most valuable automation I’ve built is a 200-line Python script called domain-daily-pipeline. It checks 20,000 expired domains, filters them by quality, appraises the top candidates, and emails me a report.

It takes 45 seconds to run. It runs automatically every morning at 8 AM. It saves me about 5 hours of manual checking per week.

The code isn’t clever. It doesn’t use AI or machine learning. It’s simple: fetch a list, check each domain against filters, score the survivors, rank them, output the result. That’s it.

The Problem: Automation Feels Like Engineering

Most people think automation requires:

• Complex infrastructure (Docker, Kubernetes, cloud deployments)
• Expensive tools (Zapier, Make, n8n)
• Deep technical knowledge (APIs, microservices, webhooks)

The reality: most useful automations are simple scripts that:

• Read a file
• Process each line
• Write the result somewhere
• Run on a schedule

That’s it. 50-200 lines of code, a cron job, and you’ve eliminated a weekly chore.

What My 200-Line Script Does

1. Fetches expired domain list from a GitHub repo (WhoisFreaks, 20K domains)
2. Filters by: .com only, no hyphens/numbers, 12 chars max, English words
3. Checks DNS (fast, 20-30 at a time)
4. Scores each domain using the 3-criteria framework
5. Ranks by score
6. Writes a markdown report
7. Saves to cache

That’s it. No database, no API calls to paid services, no machine learning. Just a for loop and some if statements.

The hardest part was the first hour: fetching the domain list and parsing it. Everything else was adding one filter at a time.

The Second Best Script: Email Deduplication

Another 100-line script: checks a list of companies against my outreach tracker (3,600+ records) and removes duplicates before sending.

Before I wrote this, I emailed the same company 5 times about the same domain because I didn’t remember contacting them. After, I can’t accidentally duplicate because the script checks and blocks it.

This script doesn’t look impressive. It reads a CSV, compares columns, writes a filtered CSV. But it saved me from looking unprofessional to dozens of companies.

How You Can Do It Too

Pick one task you do that’s:

• Repetitive (same steps every time)
• Rule-based (no judgment required)
• Takes 15+ minutes per week

Write a script that:

1. Reads input (file, API, clipboard)
2. Does the processing (filters, calculations, lookups)
3. Writes output (file, email, notification)

Schedule it:

• crontab -e
• 0 8 * * * /usr/bin/python3 /home/you/scripts/daily-task.py

For non-technical people: Describe your manual process to an AI agent and ask it to write the script. You don’t need to be a developer to automate your workflow.

When This Doesn’t Work

If your task involves human judgment (deciding which client to call first, which domain name is better), a script can’t replace you. Use it to prepare data, not make decisions.

If the data source changes format every month, the script will break constantly. Only automate stable inputs.

Related: Build Your Own Tools

##s

Q: What’s the easiest first automation?

A: A daily email report. Fetch data from a source, format it, send it to yourself. One less thing to remember.

Q: Do I need to know Python?

A: Python is the easiest for this kind of thing. But bash scripts work too. Pick what you know.

Q: How do I make sure the script doesn’t crash silently?

A: Add notifications. Cron sends errors by email by default. Or have the script send a Telegram message when it finishes.

Q: What’s the best automation you haven’t built yet?

A: Automated invoice reconciliation. Still manual, still painful, still on the list.

Related: I Automate Everything That Doesn’t Need Judgment – the broader automation philosophy. Building OmniPost – a more complex automation example.