How do you diagnose whether a new product feature that generates thousands of new indexable URLs is creating genuine organic value or producing thin content?

The most reliable diagnostic isn’t a manual content review of thousands of pages, it’s checking what Google itself has already decided about your URLs using data you already have access to: Search Console’s Page Indexing report. Google’s own indexing decisions, specifically the ratio of pages it indexes versus leaves in “Crawled – currently not indexed” or “Discovered – currently not indexed,” function as a real, already-computed quality signal at scale, because Google explicitly withholds indexing from pages it judges to be low-value duplicates or thin variations of the same template. Pair that with a smaller manual sample audit and engagement data, and you get a genuinely defensible answer rather than a guess.

Why indexation ratio is the right starting signal

When a product feature generates URLs programmatically (city pages, filtered listing pages, tag pages, auto-generated profile pages), the underlying risk Google is evaluating for is described directly in its Spam Policies under scaled content abuse: using automation to generate large volumes of pages that provide little or no added value to users. Google doesn’t need to apply a manual action to suppress this pattern; its indexing pipeline already declines to index pages it judges redundant or low-value, and Search Console surfaces exactly that decision.

If a large share of your new URL set sits in “Crawled – currently not indexed,” that’s Google’s own quality judgment telling you the pages aren’t clearing its bar, before you’ve spent a single hour on manual review. Conversely, if the bulk of the new URLs are indexed and Google continues to crawl them at a healthy rate over time (visible in Crawl Stats), that’s a positive, if not conclusive, signal that Google isn’t treating the set as templated filler.

This is a genuinely useful practical proxy specifically because it’s not a fabricated or invented metric, it’s Google’s real indexing pipeline output, already computed, already segmentable by URL pattern in the Page Indexing report.

The audit workflow

1. Segment by template, not the whole set at once. Pull the new URLs into their own group (by URL pattern) in the Page Indexing report rather than reading the aggregate property-level number, which will dilute the signal with your existing, presumably healthy, page set.

1. Check the indexed-versus-excluded ratio for that segment specifically. A high proportion sitting in “Crawled – currently not indexed” is the strongest available red flag that Google’s own systems see the pattern as low differentiation. There’s no official “safe” percentage Google publishes here, so don’t treat any specific threshold as a hard rule, but a segment where indexation is dramatically lower than your site’s baseline is a real signal worth investigating rather than an ambiguous one.

1. Sample manually against Google’s own published self-assessment questions. Google’s “creating helpful, reliable, people-first content” guidance gives a concrete list you can score pages against directly: does this page provide original information, insight, or data not available elsewhere; would someone reading it feel they learned enough to achieve their goal; does it avoid being a thin wrapper around the same underlying data with only a template variable changed. Pull a genuine random sample across the new URL set, not just the best-performing pages, and score each against these questions.

1. Check engagement on the subset that is indexed. Pages Google has chosen to index but that show extremely low engagement (near-zero time on page, high immediate bounce specific to organic entrances) alongside low or declining impressions over time can indicate the content is technically indexed but not actually serving user need, a softer signal than the indexing decision itself but useful corroboration.

1. Compare against the doorway-page and scaled-content-abuse definitions directly, not just intuition. Ask specifically whether each page type serves a genuinely distinct user need, or whether multiple pages exist mainly to capture more long-tail queries funneling to functionally the same result.

What this diagnostic deliberately avoids

There’s no safe or unsafe URL-count threshold here, the volume of new URLs by itself is not the problem Google is evaluating. Ten thousand programmatic pages built on genuinely distinct, real underlying data (actual per-item inventory, actual per-location information) can be entirely legitimate, while a few hundred templated pages with swapped city names and no real local information can trip the same policy. The mechanism Google applies is a value judgment about whether each page serves a distinct need, not a count-based rule, so any diagnostic that starts and ends with “how many pages did we create” is asking the wrong question.

As a hypothetical example, imagine a marketplace platform we’ll call “Site B” that ships a feature auto-generating a page for every seller’s inventory location, producing around 6,000 new URLs overnight. Hypothetically, if the Page Indexing report showed 70% of that segment sitting in “Crawled – currently not indexed” while the site’s baseline indexation rate elsewhere was above 90%, that gap alone would be a strong signal to investigate before assuming the feature is working simply because it shipped without errors.

What to do with the finding

If the indexation ratio and manual sample both point toward thin content, the fix is adding genuine per-page differentiation (real distinct data, not just a swapped variable) or consolidating overlapping URLs rather than trying to force indexation through technical means like sitemap resubmission, which doesn’t address the underlying quality judgment. If both signals look healthy, that’s reasonably strong evidence the feature is producing genuine value, and continued monitoring of the same indexation ratio over time (rather than a one-time check) is the right ongoing practice as the feature scales further.