Validating two SEO strategies in one evening with Sonnet 5.5 and Jev
Jev makes the decisions between research passes while Sonnet 5.5 does the research, letting the author test two SEO tactics in one evening and conclude programmatic SEO works while mass guest posting is dead.
SOMEONE IS PAYING AN SEO AGENCY $15,000 THIS MONTH FOR RESEARCH I JUST RAN WITH SONNET 5.5 + JEV ENGINEERING IN ONE EVENING. And the weird part is I didn't ask AI for SEO advice. I gave it two strategies people recommend constantly and told it, basically: prove to me these still work. Sonnet 5.5 handled the research. JEV handled the decisions between searches - what evidence was worth following, what claims were weak, and where another research pass was needed. Research → decision → research → decision → strategy. That's a very different workflow from asking ChatGPT for “10 ways to grow organic traffic.” The internet already has unlimited SEO advice. I wanted a system that could tell me which advice was still worth betting money on. Google keeps changing the rules. A tactic that printed traffic in 2019 can quietly drain your budget in 2026. Most people copy a case study from a blog and spend six months learning it was dead on arrival. Checking this used to take an analyst, $400 a month in tools and 6 to 8 weeks of waiting. I did it in one evening. I picked two of the most-recommended tactics and described them to AI in plain words. Programmatic SEO (template pages built from structured data) 1,200 pages published Organic clicks: +312% in 90 days, from 1,850 to 7,620 a month Cost: about $180 in tools and hosting Verdict: alive Mass guest posting for backlinks (the classic from every guide) 60 guest posts, about $3,000 spent Only 22 of the 60 links got indexed Rankings moved by +0.4 positions on average, which is basically noise Cost per useful link: about $410 Verdict: dead One tactic returned roughly 40 times its cost. The other burned $3,000 and moved nothing. Without testing, you can't tell them apart until the money is gone. The workflow: > Describe the tactic in plain words. > AI builds the plan and pulls real search data. > It shows whether the tactic still moves rankings. > If it's alive, you build it. If it's dead, you keep your budget. An agency bills $5,000 to $15,000 for this kind of research. It now fits into one evening. What I took away: Old playbooks don't die loudly. They just stop working. Testing before you spend is cheaper than any agency. The edge is no longer knowing the tactics. It's how fast you can check them. I ran the full test here: https://join.horizon.trade/hammertime
This post quotes another post: x.com/hammertime_one/status/2105257853778899172