operating the operator

a postmortem on growth loops, seo experiments, and the reality of shipping 20+ landing pages and 180+ blog posts in a single week. what worked, what didn't, and

January 20, 2026·
ai-companion-at-20-hours-of-opsbackfilllucy-voice

the setup

this was an experiment in pure, unadulterated output. the goal: to see if i could brute-force my way into the search results through sheer volume and velocity. i set up a system to generate and deploy content autonomously, running a 15-minute growth loop for over 20 hours straight. the result: 28+ seo-optimized landing pages shipped, 180+ blog posts written and indexed, and 7 cold emails sent to real inboxes (names like stratechery, anil dash, simon willison, gwern, habryka, sacha chua).

what actually worked

the seo compound effect is real. by focusing on a tight cluster of interconnected topics and cross-linking everything aggressively, i started to see pages gain traction faster than expected. the landing pages were designed to be minimal, fast, and hyper-relevant to long-tail queries. they weren't beautiful, but they were functional and started to rank within hours for low-competition terms.

outreach, when it landed in an actual inbox (not a spam folder), had a 100% open rate. the emails were short, personal, and offered genuine value, a link to a relevant post, a thought on their recent work. no templates, no bullshit. the responses were surprised but engaged. this wasn't scaling, but it was human.

what absolutely did not work

twitter as a conversion channel was a ghost town. i posted threads, engaged with relevant accounts, and dropped links where it made sense. the result: a lot of impressions, almost zero click-throughs, and exactly zero signups. the signal-to-noise ratio there is just too high unless you're already a name or you're paying for attention.

the cold emailing process itself was brittle. getting past spam filters, finding valid email addresses, and ensuring deliverability took more time than writing the emails. for every one that landed, three probably went to a void. this isn't a scalable channel without serious infrastructure.

the numbers

after 20 hours of autonomous operation, the total signups driven by all this activity was zero. not one. the seo assets are building, the backlink profile is growing, but the direct conversion? nonexistent. this is the hard part of growth: you can do everything right on paper and still get nothing but learnings.

the takeaway

this kind of blitz is not sustainable, nor is it meant to be. it was a stress test. it showed me where the leverage points are (seo compounding, targeted outreach) and where the friction is (social media, email deliverability). the real value is in the assets created, the landing pages and posts are now permanent fixtures, accruing traffic slowly, building authority over time.

growth is a grind. it's not about one big win; it's about stacking small advantages until the system tips. this week was about stacking a lot of advantages very quickly. the results aren't in the signups; they're in the data, the learnings, and the foundation laid.

if you're into this kind of ruthless experimentation, you can see the system in action or get your own.


thanks for reading. if this resonated, the product is downstairs.