SEO Studio: seven days of publishing observed across four sites

Wix published on five of the seven days, Shopify on three, Webflow on three and Oryon on four. No site covered all seven days. No site-day contained two publications in the closing observations.

A missed day means that a complete observation contained no matching publication receipt for that local date. It does not mean that nothing was prepared: several preparations stopped before publication. Any later success does not change this count.

Times are in Paris, shown to the second. “Loaded” means that the image actually displayed during the check, with a single cover and a distinct inline illustration. A dash means there was no published article to inspect. On a small screen, the table can be scrolled horizontally.

* “None recorded” refers to generation interventions logged in the stages examined. The available evidence does not independently identify the initial caller through a journal. For the published rows on the 19th, the column counts corrections to the delivered article. An additional recovery of Webflow’s preparation for the 18th, performed on the 19th and unsuccessful, is described below.

Wix was the only pilot configured for 09:00 Europe/Paris. All five publications arrived after that time, but before 09:05. They were therefore late relative to the exact configured time. Their precise delays are preserved below rather than rounded into “on-time” publications. The 17th and 18th remain missed days.

Shopify, Webflow and Oryon had no configured daily publication time. We do not apply Wix’s time to them or label them late against 09:00. Delays after an individual article was scheduled are separate from adherence to a daily publication slot.

For all 15 published articles, the retained evidence compares the title and body with the delivered revision. It documents one loaded cover and a distinct inline illustration loaded after scrolling, with captions and credits. Navigation cards and different sizes of the same image are not counted as additional illustrations.

On 15 October, Shopify’s portrait cover appeared oversized and blurry on desktop. Loading and uniqueness were verified; this presentation caveat remains in the report. On 17 October, Webflow’s inline illustration showed two roller skates: it matched the recorded content, but its relevance to the topic was not validated. Technical correspondence alone does not establish editorial quality.

The Shopify store remained password protected. Its three rendered articles were inspected through the authorised owner preview, not as pages available to anonymous visitors. We publish no screenshots of that preview. These checks describe the retained views; they are not an exhaustive audit of every screen.

Some preparations stopped at plan, length or factual checks. Some had automatic recoveries; these are separate from manual interventions. An attempted check does not turn a blocked preparation into a publication.

On 19 October, two manual corrections were recorded for the Shopify preparation: a plan correction followed by a correction of claims requiring review. The article was eventually published at 13:56:18. One manual correction of Webflow’s preparation for that day led to publication at 13:41:21. Both deliveries are successes with intervention and count only on the 19th.

A fourth manual recovery, also on Webflow on the 19th, concerned a preparation for the 18th. It failed without producing an article. The 18th remains a missed day. The journal records an administrator account for these operations; that alone does not independently identify the person who performed them. For this fourth recovery, the observation record does not establish the provenance of human authorisation.

The operating environment also changed during the week: fixes for Shopify service access and blog selection, interface language and daily reliability were deployed with the necessary migrations. These are operational interventions separate from the four generation recoveries. The reliability fix on the 19th applied to eligible new preparations created after activation; installing it did not establish success for already blocked preparations or for future days.

These two screenshots come from our Webflow pilot’s public page after its publication on 19 October. They show the title and cover, then the inline illustration after scrolling. They document that specific delivery, achieved after a manual correction. Zeal Consult is the name displayed by the test site’s theme, not a client reference.

The count uses each published article’s unique identity and remote receipt, its actual publication date, the matching version and the automatic review of that exact revision. Oryon’s native page copy was also compared with the article. Recovery and modification records were examined separately.

A receipt and automatic review do not independently prove every generation step. A separate length verdict is unavailable for all 15 published articles, and some final visual-review verdicts for images remain unknown. These are not reported as all checks passed. Likewise, the absence of a recorded intervention is not absolute proof of autonomy.

Observations are discrete and include gaps, notably from 16 to 19 October, then from the afternoon of the 19th until closure. The final observation was on 20 October at 11:21:51 Paris time; no closing check at 00:05 is evidenced. Receipts support the count through the end of the 19th. Previously documented rendered views are reused only for the same unchanged revisions.

The 13–19 October window is retained. Raw evidence, internal identifiers and account screenshots remain private. These results describe seven days across four test sites; they do not guarantee the cadence of future days, perfect accuracy of every claim or any search-performance effect.

The concrete result is verified publication and rendering for 15 articles, while the expected cadence missed 13 site-days. Corrections enabled two deliveries on the last day without erasing previous failures. Useful monitoring preserves these distinctions instead of reducing the report to a total article count.

The connection guides explain Wix and Webflow prerequisites. SEO Studio describes preparation, checks and publishing; SEO consultancy is a separate service. No traffic, ranking or sales gain is inferred from this test.