The client-proof workflow → Sampling methodology

Sampling vs Full Checks: The Index Audit Methodology

Checking all 80,000 URLs every week is a budget decision pretending to be a diligence decision. Checking 40 random ones and calling it monitoring is the opposite mistake. The working answer sits in between, and it has rules.

The core idea: templates fail together

Pages built from one template share their indexing fate — same internal linking pattern, same content skeleton, same directives. When deindexation happens at scale, it happens to a template, not to random URLs. A deploy adds noindex to pagination; a variant rollout thins out product descriptions; a category redesign orphans a level. Which means a modest sample per template detects the failure almost as fast as a full check would.

The sampling rules

When only a full check will do

  1. Baselines. The first audit of a new client runs the whole scope — samples need a reference point.
  2. Milestones. Migration, replatforming, template rollout, domain consolidation. Google's own sitemap documentation is the reminder that discovery and indexing are separate steps — after big changes you verify both, URL by URL.
  3. Disputes. Vendor and client arguments are settled with complete files, never with samples.

Batch capacity makes the full runs painless — up to 100,000 links per task through the bulk index checker, with a per-URL file back. The economics of the audit come from spending that capacity where it changes decisions, and sampling everywhere else. For collecting the URL lists themselves, the free sitemap extractor covers the sitemap side.

Rule of thumb: full check at day 0 and at milestones; stable per-template samples in between; any sample delta beyond your comfort band triggers a full check of that one template — not of the whole site.
Related playbooks: E-commerce catalogs · The monthly health check · The full agency workflow