Case study · engagement 2026-07-11 to 2026-08-02 · published August 31, 2026
The cleanup PruneQ ran on its own website
This is the first PruneQ case study, and the client is PruneQ. It includes the part most case studies leave out: for roughly four weeks, the work was worth nothing, because it had never actually reached the live site.
By Christopher Buchner, founder of PruneQ
Where it started, 2026-07-11
- Pages crawled
- 11 public URLs, 2026-07-11.
- Technical SEO
- 100/100 on this audit's own checks. These are internal diagnostic scores, not Google signals.
- AI-search readiness
- 92/100: crawler access 30/30, schema and entity 22/30, answer-shaped content 25/25, discoverability 15/15.
- Findings
- 0 critical, 0 high, 1 medium, 0 low. The single medium: the organization graph carried no sameAs profile links.
- AI crawler access
- GPTBot, OAI-SearchBot, ClaudeBot, Claude-SearchBot, PerplexityBot and bingbot each returned 200 on a live fetch, not merely 'allowed' in robots.txt.
A near-perfect internal score on a site nobody had found yet is worth stating plainly: it meant the technical layer was not the constraint. That is the same thing PruneQ tells clients when an audit comes back clean, and it is unwelcome news roughly as often as it is good news.
What was actually wrong
A second audit on 2026-07-30 compared the live HTML against the source. The live site was serving a build from roughly 2026-07-01. Every on-page, schema, sitemap and content improvement made in the intervening month existed only in the working tree. The evidence library returned 404 in production.
No crawl block, no indexation trap, no duplicate content, no schema error. The constraint was publication and time. A month of clean work had shipped to nobody, and the audit that found it was pointed at PruneQ's own site.
This is in the case study on purpose. A vendor who only publishes wins is telling you what they want you to believe, not what they found.
What changed
- Added the missing sameAs link after the founder profile was owner-confirmed on 2026-07-12. Nothing was invented to fill the gap; the field stayed empty until there was a real profile to point at.
- Published the evidence library in three dated batches: two on 2026-07-31 and one on 2026-08-01, nine articles in total.
- Built explicit publication gating, so an unfinished page returns 404 by design instead of shipping thin. Four drafted articles are still deliberately withheld.
- Replaced deploy-time sitemap freshness with explicit per-page change dates, so the sitemap cannot claim a page changed when it did not.
- Kept a dated self-audit on the site instead of a manufactured case study, until there was an engagement honest enough to write up.
Verified on 2026-08-02
A read-only status check of 22 public URLs on pruneq.com — the commercial pages, the nine published articles, the insights hub, the sitemap, robots.txt and llms.txt — returned 200 on 21 of them. The one 404 is /insights/structured-data-clarity-not-promises, a drafted article that is deliberately withheld. That 404 is the publication gate working, not a broken link, and it is reported here rather than quietly excluded from the count.
What this does not prove
- Rankings. No position, impression, or click data was opened for this write-up. Search results are decided by Google, not by PruneQ.
- Traffic. No GA4 session data is cited here.
- AI citations or recommendations. Crawler access was verified; being cited is a separate thing nobody can promise.
- Leads or revenue. This engagement produced no attributable inquiry, and saying otherwise would be the exact claim PruneQ tells clients to distrust.
The reasoning behind that list is written out in what nobody can promise. The starting measurements are in the dated website baseline.