Takusen / Takusen technical self-audit
Internal example · not a client case study
A real review of this website’s technical readiness, with evidence boundaries made explicit. AI-answer visibility was not measured.
Baseline reviewed on September 24, 2026 (America/Juneau; public requests at September 25, 03:52 UTC). Scope: the homepage, former sample-audit page, stylesheet, public robots response, sitemap URL and a nonexistent route. No paid AI API or consumer-answer tests were run.
Source means inspected files. Production means the public HTTP responses captured on that date. Local revision means the revised static output; it does not establish that production has changed. This report is a dated snapshot.
Source and production HTML: the homepage promised a universal 0–100 score and called tier comparison unique. No scoring implementation or exclusivity evidence was supplied in the website repository. These claims could create misleading expectations.
Change: publish separate signal definitions, agreed scope and sampling limitations; remove the exclusivity claim. The local revision does this, without fabricating a replacement score.
Source and production HTML: the sample page had placeholder headings; the homepage promoted it with unfinished client/city text. A visitor could not assess the deliverable.
Change: replace it with this substantive internal example and redirect the old sample route when deployed.
Source: both pages lacked canonical and social metadata and JSON-LD. Production: the sitemap request returned 404. This left fewer explicit signals about preferred URLs and site structure.
Change: add per-page metadata, a sitemap, navigation and conservative structured data. These are technical improvements, not a guarantee of indexing or rich results.
Source: script added a hiding class before constructing IntersectionObserver. If that API failed, sections could stay hidden. Muted text used #5A6478 on navy.
Change: keep essential content visible by default, remove observer dependence, raise muted-text contrast and provide keyboard focus and skip navigation.
Production: HTTPS apex, HTTPS www and HTTP apex each returned 200 rather than redirecting to a single preferred origin.
Change: add self-referencing HTTPS apex canonicals locally; arrange a host-level redirect preserving path and query parameters. The redirect remains an owner action, not a completed production change.
Production: robots.txt returned a content-signal explanatory preamble with no user-agent rules in the captured response. A random nonexistent route correctly returned 404, with an empty body.
Change: provide an explicit crawler policy and a helpful 404 page. Recheck the combined CDN and origin robots response after any approved deployment.
The baseline used source inspection and seven bounded, read-only HTTP requests. Public HTML matched the pre-change source. No X-Robots-Tag was present in the captured homepage headers. Initial sandbox DNS failures were environment limitations; the subsequent public requests succeeded.
The local revision includes repeatable checks for metadata, internal links, sitemap inventory, JSON-LD references and deployment contents. Local checks passed for the 11 generated HTML pages, including metadata, links, schema references and sitemap inventory. Browser review covered 390px, 768px and 1440px layouts, no-JavaScript content, keyboard focus, form errors and reduced motion. The local server returned 200 for intended pages and 404 for a missing route. Its detailed validation record is maintained with the engineering documentation. Local test results do not establish live crawler access or search inclusion.
Search Console, Bing Webmaster Tools, CDN security settings and genuine provider crawler logs were not inspected. No AI mentions, citations, recommendation rates or before/after business outcomes are reported.
Under an agreed scope: versioned buyer questions, repeated responses from specified surfaces, dated citations and screenshots where available, competitor context, explicit counts and exclusions, and a fix list mapped to the customer’s pages. Sensitive response evidence belongs in an access-controlled delivery, not on a public sample page.
Read the proposed methodology or discuss an audit. Use your browser’s print command for a clean paper copy.