The Optiview Measurement Methodology

Publish reviewed, machine-readable brand context through supported integrations.

Payload size is a mechanical measurement, not proof of improved AI visibility. We separate delivery, retention, and comprehension so a successful conversion is not mistaken for a successful outcome.

1. Delivery verification

Available today: recorded payload deliveries, with separate Enterprise API and proxy metrics. These records do not establish verified bot identity or model ingestion. A User-Agent label or Markdown preference is not identity verification.

Benchmark method: record the request format preference, response outcome, and independently validated client identity where available. Report verified, unverified, and test traffic separately. Do not infer crawler adoption from a manual curl request.

2. Retention rate

Planned benchmark: establish a reviewed reference set of product facts, category destinations, and differentiators before extraction. Count which remain accurate in the output and document omissions or changes. Evaluate static fetching and browser rendering separately.

Report retained reference facts divided by total reference facts, alongside the page sample, extraction settings, and failures. Current pruning tests are regression checks; they do not establish universal taxonomy retention.

3. Comprehension lift

Planned benchmark: compare baseline content and the Optiview payload using the same questions, model version, prompt, and inference settings. Use repeated trials and blind scoring against the reviewed reference set.

Measure factual correctness, unsupported assertions, and unanswered questions. Evaluate extraction alone separately from extraction plus approved curation. Controlled answer quality does not prove real-world crawling, indexing, or citation lift.

Benchmark reports

No controlled benchmark reports are published yet. This page describes our evaluation methodology, not completed research or a claim that testing is already underway.

Inspect a simulated payload → · Review the integration architecture → · Discuss an evaluation pilot →