Zhihe GrowthA real case is valuable because it explains the client's problem, visible work completed, verifiable figures, and the test conditions to which results apply, not because it displays an attractive percentage. Separate AI brand mentions, official-site citations, and factual accuracy: they are different outcomes.
Five Layers of Case Disclosure
First, identify the client and permission scope: named or anonymized, approved brand name, and permitted materials. Second, describe initial conditions: existing site, product structure, and information gaps. Third, explain implementation roles for technical access, fact governance, long-form content, FAQs, evidence, and internal links. Fourth, present results: public asset counts first, then scoped test values. Fifth, state whether platform, region, time, question set, and valid-answer definitions permit comparisons.
| Metric Type | What It Can Establish | Common misuse |
|---|---|---|
| Asset Counts | Published products, articles, FAQs, and evidence pages | Treating quantity as quality or a citation guarantee |
| Technical inspection | Whether crawling, Schema, and links pass defined checks | Equating valid syntax with search indexing |
| AI Answer Metrics | Performance of a fixed question set during the recorded period | Treating an interim percentage as lasting organic traffic |
| Business Outcomes | Subsequent effects on inquiries, sales, and similar outcomes | Attributing all growth to GEO without an attribution design |
Public Scope of Two Named Projects
Zhihe Growth has permission to identify SuperPDR and ELEREIN by name. According to this site's current case pages, SuperPDR's public scope includes 692 product entries, 37 categories, 28 technical articles, 16 FAQ hubs, and 2 Trust entry points. Its stage-specific official-site citation rate for the defined question scope is 17%. ELEREIN's public scope includes 271 dynamic sitemap URLs, 166 product detail pages, 13 categories, 52 technical articles, 72 short Q&A entries in 6 FAQ hubs, 6 Evidence topic pages, and 225/225 target Schema checks. Its stage-specific official-site citation rate is 15%.
These percentages come from different project question sets and recording periods, not a same-denominator comparison or ranking. ELEREIN's page also reports pre-optimization results for non-branded questions Official-site citations were approximately 0.6% for that narrower set of questions. This is not an overall baseline with the same denominator as 15%, so it cannot be used to calculate a like-for-like multiplier. If a fully comparable baseline is disclosed later, report its date, denominator, and question set separately while retaining the source for this version.
Keep Results Traceable
For each AI test, retain the original question, platform and mode, date and region, complete answer, final clickable source URLs, mention and citation classifications, and whether cited content supports the claim. If a source appears only as a title chip, click or hover to verify its URL. Failure to copy a URL does not mean the platform provided none. Public case studies need not expose full conversations, but internal evidence must support the statistics. See Citation Rate Definitions.
Products, specifications, and decision boundaries in a case should be verifiable on the client's actual website: SuperPDR Repairability Matrix shows the structure of technical decisions; the ELEREIN Equipment Selector and Evidence Center show a path from scenarios to evidence. They establish visible content structure, not the underlying performance-test records.
Update and Correction Rules
When asset counts grow, project scope changes, or platforms update, publish a new case version with an as-of date and metric definitions. Do not silently present old results as current. If a client withdraws disclosure permission, remove the affected facts and media, then check Schema, cached pages, and articles citing them. Where older reports conflict with the current approved website scope, use the current page confirmed in this review for public statements, retaining older reports internally for traceability.
Scope and limitations: A case is not a performance guarantee. Crawling, model updates, question wording, and regional changes can affect later results. Zhihe Growth presents these two real projects so clients can examine methods and evidence, not as promises for every industry.