> the contento playbook
The /articles/ section of this site is generated by the same CLI we sell. Same niche file, same schemas, same chrome, same on-domain paths. We publish the GSC numbers monthly. If we wouldn't ship it on our own domain, we wouldn't sell it.
If you've been burned by AI content — or watched someone else get burned — you should not take a vendor's word that their tool is "safe." You should ask whether they trust it on their own apex domain.
We do. contento.sh/articles/ is built by the public CLI, published under the same Helpful Content rules we ship in our docs, and held to the operator-side checklist from our own Google Policies page. Every page passes a human review — mine — before publish. Every batch reports back to GSC, and every month the numbers go on this page.
If the experiment underperforms, you'll see that here too. The whole point is that the receipts are public.
setup
niche
A custom niche, hand-authored from inside the trade. 30 high-intent subtopics, real pain points (Helpful Content fear, agency cost, token surprises), real audience description. Lives at marketing/niches/pseo-for-indie-saas-founders.json in the repo.
schemas
Three of the ten bundled schemas. how-to for operational pages ("how to audit pSEO pages for HCU risk"), comparison for decision pages ("BYO-key vs. proprietary AI content platforms"), faq for quick-answer pages.
paths
/articles/No subdomain. No third-party host. Pages build into marketing/public/articles/ and ship with the rest of the Vite build to contento.sh. Chrome (header, footer, CSS) is the live site's chrome — rebuilt on every Tailwind hash change.
cadence
First batch is 25 pages. Pause for 30 days. Watch indexing, query mix, and click-through in GSC. Expand only if the numbers are healthy; prune underperformers; iterate the niche if the AI output is flat.
review
Pro editorial reviewer first, then me. No batch publishes without a manual read-through for the first three batches. After that, we'll evaluate whether to relax the gate — and report it here.
reporting
Impressions, clicks, indexed count, top queries, and underperformer count from the live contento analytics commands. Updated on the 1st of every month.
the receipts
The experiment hasn't started publishing yet. These are the placeholders the monthly update will fill. Each row is pulled from contento analytics summary --json and committed to this page.
PAGES PUBLISHED
0
target: 25 in batch 1
INDEXED IN GSC
—
measured 14 days post-publish
IMPRESSIONS / 30D
—
first month is noise; month 2+ matters
CLICKS / 30D
—
honest number, not adjusted
Last updated: not yet. First update on the 1st of the month after batch 1 publishes.
Likely causes: the niche overlaps too closely with our existing docs (cannibalization), the chrome is bloated and slows page weight, or the subtopics are too generic. Response: prune duplicates against /docs/pseo-guide.html and /docs/google-policies.html, audit page weight, tighten the niche.
Means the title pattern or meta-description is failing the SERP. Response: rerun affected pages with adjusted prompts and re-publish. The CLI's contento analytics zero-traffic command surfaces these directly.
The catastrophic outcome. If it happens, we will publish the post-mortem on this page within 7 days, prune the offending pages, and explain what we'd do differently. The honest version of "we got it wrong" is more useful to a reader than a quiet rollback.
If batch 1 fails the editorial bar, we don't publish — we ship a niche revision (more specific pain points, better subtopics) and rerun. The point is to test what Contento can produce on a serious niche, not to publish anything that fits in the schema.
read the live articles, read the GSC numbers above, then decide. We don't do free trials of the CLI because it's already free. We do offer a White-Glove engagement if you want this entire program run for you, on your domain, by the people running it on ours.