TriRank's own-site SEO baseline: what we fixed and what remains unproven

The starting point

We use trirankai.com, our own site, to check the path from connecting data to choosing a next action. This is a process case, not proof of a traffic or revenue increase. Search Console clicks and AI citations answer different questions and stay separate in the report.

October 13: connect and verify the data

The first Search Console import failed because the main Worker lacked its connection-encryption binding. We restored the existing configuration and reran the import. The recorded run completed at 23:41 Shanghai time on October 13, 2026, processing 1,337 rows. The Google data ended on October 11. Row count measures import work, not visitors, clicks or revenue.

We also added the question “TriRank for Shopify SEO.” It was registered for the next scheduled check. Older results for “trirank” were historical observations; they were not answers to the newly added question. The dated walkthrough shows the connection, imported rankings and existing citation evidence. Waiting periods were shortened in editing.

October 14–15: audit and make the next changes

The public-page audit found a robots rule blocking Next.js rendering assets, sitemap-index modification dates changing on every request, login and registration URLs included in the sitemap, and an n8n page still describing a published template as under review.

This iteration changes those rules and dates, adds clearer steps after a free check, and separates the published Slack workflow from the Email template awaiting review. These are repository changes prepared for release. The page publication date does not establish when every fix reached production.

What we will measure after release

Record the actual deployment time first. Check that Google can fetch the new robots file and rendering resources; submission alone is not proof of recrawling. At 7, 14 and 28 days, compare the same query, page, country and device filters in Search Console. Keep data before the October 13 release separate from data after this technical release.

For AI checks, save the question, engine, checked time, cited sources and source position. Pending, unavailable and measured but not cited are different outcomes. A changed citation reading does not prove that an earlier content edit caused it.

Try the same path

Start with the AI visibility checker, use the methodology to interpret its scope, and look at the read-only product demo. For a full deliverable example, the sample audit uses a different site we own; it is not the trirankai.com baseline. You can use the n8n integration to send notifications when stored readings change.