Everyone says SEO is dead. Last year my sites pulled 200k+ clicks and 10M+ impressions out of Google.
It is not dead. It stopped forgiving one mistake, and I made it for a whole quarter.
I wrote 28 pages for one of those sites. Three months later: zero impressions. Not low. Zero.
The pages were fine. I picked the topics by thinking about them instead of checking.
Mistake one: people search in lists.
"best running shoes for flat feet" gets searched.
"asics gel kayano for flat feet" does not.
I shipped a page per item while the demand sat in the aggregate query above them.
Mistake two: difficulty scores lie.
Two keywords both look easy. One has an open SERP. The other is held top to bottom by publishers you will never outrank. Opening the actual results takes 60 seconds per keyword, and it decides whether your next month of writing returns anything.
So I turned the method into a Claude Code skill.
9 steps from "no traffic" to "we know what we are building", each with a command that verifies it. Steps 3 and 4, find real demand and drop what you cannot win, decide the outcome. Everything else is prep or cleanup.
Two scripts ship with it, standard library only, no API key:
audit_site.py crawls your sitemap and finds what breaks silently: hreflang groups, x-default pointing at your homepage, duplicate meta, orphan pages, invisible unicode that models leave in your copy.
preflight_page.py checks a single page before you publish it, including parsing its JSON-LD and following the images the markup claims exist.
Free and MIT.
What is the page you wrote that got nothing?
The most interesting martech in YC's Spring 2026 batch isn't adtech — it's AI operators that run growth.
Broke down @CharacterQuilt , @kuli_ai , @manicule and @InstaAgentAI , with the traction behind each.
1/ I went through YC's Spring 2026 batch (X26/P26) looking for adtech.
Almost none. No DSP, no SSP, no retargeting.
Instead: AI operators that take whole chunks of work off marketing teams.
1/ I went through YC's Spring 2026 batch (X26/P26) looking for adtech.
Almost none. No DSP, no SSP, no retargeting.
Instead: AI operators that take whole chunks of work off marketing teams.
6/ My take: YC isn't betting on one more AI copywriter. It's betting on full-stack growth agents.
The winners won't generate content — they'll close the loop:
hypothesis → creative → launch → performance data → next experiment.