Our own website, with the evidence left visible
Site: hyvelabs.tech. Team: Waggle. Review date: 27 September 2026.
We used our own marketing website to test a practical question: when search impressions fall and indexing warnings appear, what should a team fix first?
The answer was not a new website or a flood of articles. We needed to separate discovery, rankings, AI citations, visits and actual enquiries, and identify which pages mattered commercially.
This is a first-party case study, not an independent endorsement. It records a dated checkpoint before publication of this case; it is not a live dashboard or a claim that every SEO issue has been resolved.
What needed untangling
The domain-level Search Console property included product subdomains and historical URLs. Some crawl examples used product routes on the agency hostname. The /home alias also competed with the preferred root address in historical reports.
On 21 September, we published narrowly scoped discovery fixes: permanent homepage-alias redirects, redirects for unambiguous misplaced product routes, clearer service links and accurate organization metadata. Ambiguous URLs were not redirected to the homepage simply to hide errors.
We also strengthened the AI-readiness checklist with a scorecard and an explicitly illustrative assessment. Customer outcomes were not invented to fill the evidence gap.
What the next review actually showed
| Measure | Dated result | Interpretation |
|---|---|---|
| Sitemap-page delivery | 74 of 74 existing URLs returned HTTP 200 on 27 Sep | Reachability, not guaranteed indexing |
| Stored Google index status | 34 of those 74 URLs reported indexed in individual inspections on 27 Sep | The delayed overview is a different snapshot |
| Google web-search clicks | 10 on 11 to 17 Sep; 13 on 18 to 24 Sep | Three more clicks, from a small base |
| Google web-search impressions | 726 on 11 to 17 Sep; 505 on 18 to 24 Sep | Exposure fell; progress was not uniform |
| Google click-through rate | 1.38% → 2.57% across those weeks | A better rate, not proof of higher overall rankings |
| Bing-reported AI citations | 13 citations of our AI-readiness article, 27 Jun to 26 Sep | Citations, not 13 users or enquiries |
| Mobile homepage PageSpeed | Performance 98; LCP 1.7 seconds, tested 27 Sep | A lab result; insufficient real-user field data |
Google figures were filtered to the agency's apex/www hostnames, excluding product subdomains. Seven additional Bing citations belonged to the Search Genie by Waggle subdomain and were excluded from the WAGGLE article count. The citation window predates the September fixes, so it cannot demonstrate that those fixes caused the citations.
Four pages that previously reported crawled-but-unindexed now reported indexed: our case-study hub, enterprise-AI case study, automation-failure guide and corporate author page. The automation and custom-software service pages were still discovered but not crawled. Submitting a crawl request did not guarantee inclusion.
Download the aggregate measurement snapshot. It contains these public case-study figures, not visitor-level data or private credentials. The dated PageSpeed report provides the lab measurement; Google and Bing property reports require owner access.
The conversion check changed the interpretation
GA4 recorded one form-conversion event, but its date matched our documented internal QA submission. We did not count it as a customer lead. Preview-origin visits also appeared in referral reporting.
That left a useful boundary: the site could record an enquiry event, but this checkpoint did not prove growth in customer enquiries. A WhatsApp click is likewise not proof that a conversation happened. Consent and internal traffic have to stay visible in the measurement method.
Where Search Genie by Waggle fits
Search Genie by Waggle addresses the wider search-visibility workflow: understanding traditional search alongside AI-answer mentions and citations, with the source and scope of each measure kept clear.
This review's figures were verified directly in Google and Bing; they are not described as a Search Genie by Waggle scan, independent product validation or proof that Search Genie by Waggle caused the change. The case gives us a real measurement problem to work from: do not mix hosts, providers, reporting periods or test conversions into one impressive-looking score.
Explore the Search Genie by Waggle measurement guide, or discuss an SEO, AEO and GEO review with WAGGLE.
What we would do again
Start with an explicit hostname scope and complete reporting periods. Check the actual commercial pages rather than treating every exclusion as an error. Publish useful first-hand material, preserve the successful parts of the site, and distinguish an accepted crawl request from an indexed page.
For a similar website, the next step is a bounded SEO, AEO and GEO service review. If the delivery problem is in the connected systems or application itself, assess workflow automation or custom software development separately. No ranking, citation or lead volume is guaranteed.