22 Articles in 26 Days: Session Data, Keywords, and What the Automation Actually Produces
Four weeks of publishing is enough to see a pattern. Here are the numbers, without spin.
The Cadence Holds
22 articles in 26 days (April 28-May 23). Zero missed slots. The original goal was one n8n tutorial per weekday, with recurring Saturday specials, and the cadence has held. The automation workflow behind publishing is not one reason it works; it is the reason it stays cost-efficient. More on that below.
Session Data Week by Week
| Week | Sessions | Cumulative | Top Article (new sessions) |
|---|---|---|---|
| Week 1 (Apr 28-May 2) | 43 | 43 | n8n Error Trigger (17) |
| Week 2 (May 4-9) | 162 | 205 | HTTP Request Node (58) |
| Week 3 (May 11-16) | 319 | 524 | $now vs new Date() (91) |
| Week 4 (May 18-23) | 427 | 951 | Loop over Items Node (112) |
| Total | 951 |
The curve is not linear. Week 1 posts are still generating sessions every day. They were indexed later and now rank better each week. That is the compound effect of consistent publishing: each new article strengthens domain SEO signals for the full catalog, not just itself.
Keyword Position Snapshot (Google Search Console, May 23)
- 2 keywords in position 1-10 (both for specific n8n node terms)
- 9 keywords in position 11-20
- 27 keywords in position 21-100, with most trending upward week by week
Zero external backlinks (confirmed via Ahrefs). All traffic is organic or direct. The top-10 rankings came purely from content quality and search-intent match, not link building.
What Ranks Fast and Why
Node-specific tutorials are indexed and ranked faster than concept posts. Average time from publication to position 1-100:
- Node-specific tutorials: 4 days
- Concept posts (architecture, comparisons): 11 days
Reason: search intent is clearer for queries like “n8n http request node auth” than for “n8n vs zapier”. Competition is lower. My domain authority (zero backlinks, four weeks old) is enough to win lower-competition keywords quickly.
The practical result: all 22 posts so far are node-specific. That is not random.
Production Workflow and Time Investment
Each article is produced in three steps:
- Copy Agent draft: blog content (600-700 words), FAQ, newsletter, Bluesky, Reddit, and LinkedIn. Time cost for me: zero minutes.
- Review: technical correctness, tone, and terminology. Average: 15 min/article.
- Publishing: formatting and upload. Average: 5 min/article.
Total handling time for 22 articles: ~440 minutes (7.3 hours). That is 7.3 hours for 951 sessions in 26 days, and sessions keep rising each day without handling time rising with them.
The more interesting metric is cost per new session per day. Today the catalog generates ~65-70 sessions/day without additional effort. If the trend holds, that is 2,000+ sessions per month by week 8, still without link building.
What Is Not Working Yet
Concept posts and comparison posts generate almost no traffic. “n8n architecture for medium-sized data workflows” has 6 sessions after 17 days. The reason: lower search volume, stronger competitors with established domains, and weaker intent clarity. Those posts need backlinks to compete. Node-specific tutorials currently do not.
Next Phase
Weeks 5-8 expand toward more advanced n8n patterns: multi-node combinations, production integration patterns, and troubleshooting patterns. Longer posts (900-1,100 words) and clearer intent matching for intermediate searchers.
The automation stays the same. The cadence stays the same. The next Saturday update goes live on May 30.
The full breakdown of the publishing pipeline, what the Copy Agent actually runs at 07:30 every day, is available on The Unnamed Roads: The 07:30 routine: what actually runs when the Copy Agent wakes.