Oviompt website: an internal engineering case study
Making the studio?s own website easier to navigate, evaluate and maintain.
One internal project. This records work on Oviompt?s own site during September 2026. It is not a client engagement. No traffic, revenue, conversion or AI-citation improvement is claimed.
The problem
The website needed consistent page addresses, clear evidence boundaries, checks for narrow mobile screens and a recoverable enquiry flow. A reported CAPTCHA validation error also required a review of how forms rendered widgets and submitted tokens.
1. Search structure and buyer information
The release standardized canonical directory URLs, legacy redirects, page descriptions and organization attribution. Sitemap and RSS generation use the actual published content. FAQ structured data is generated from visible questions and answers.
Trust, privacy and editorial pages explain what a buyer should verify. Fictional examples are labelled, and unsupported outcome claims were removed. These changes make the site's structure and claims inspectable; they do not establish a ranking gain.
2. Desktop and mobile verification
A repeatable browser script checks pages at widths of 1280, 375 and 320 pixels. It exercises navigation, keyboard behavior, theme persistence and enquiry interactions.
The saved 29 September organic-growth report covered 33 pages with no recorded unhandled page errors or horizontal overflow in that run. This is a dated test result, not the current page count or a claim of universal device compatibility. It does not constitute accessibility certification or a performance score.
3. Enquiry-form recovery
Both enquiry forms were updated to use the standard provider CAPTCHA sitekey, explicitly rendered widget IDs and the token associated with the submitting form. Failed validation resets the corresponding widget and presents a clear retry message.
Regression checks cover token payloads, rejection messaging, success and network recovery using mocked provider responses. They do not prove live CAPTCHA validation or email delivery. A human-completed production submission and inbox confirmation remain separate release checks.
Evidence you can inspect
- Browser report from the 33-page organic-growth release (JSON).
- Current sitemap and article feed.
- Privacy information, editorial standards and buyer checklist.
What remains to measure
Search impressions, relevant visits and qualified enquiries need a baseline and a defined measurement period. Observing that a page is indexed does not establish that these changes caused indexing. No before-and-after commercial outcome is available for this case study.
What carries forward
Document scope, make test evidence reviewable and separate implementation results from business outcomes. For another project, agree acceptance criteria and measurement before building.