What Our 4-Metric Completeness Gate Excludes and Why

The anti-thin content gate

Every city page on this site passes through a 'completeness gate' before it's allowed to be indexed by Google. The gate checks that the city has all four required cost-of-living metrics, each with a status of 'ok' (not quarantined) and a population above 5,000. If any required metric is missing or quarantined, the gate fails and the page is noindexed — it still renders for human visitors, but Google is told not to index it.

The four required metrics are: rent_1br (1-bedroom rent), median_household_income, grocery_index, and utilities_monthly. These are the core of the cost-of-living calculation and the affordability calculator. If any is missing, the calculator can't provide a meaningful answer, and the page becomes thin content — data-poor pages that Google's 2026 quality updates penalize at the site level.

In our live data run, 996 out of 1,000 cities pass the gate with all four metrics at 'ok' status. Only 4 cities fail — and those 4 pages are noindexed, not deleted. If the data is later completed (e.g., a new ACS vintage fills the gap), the page can be reindexed without any code changes.

Why not just show '—' for missing data?

The obvious alternative to noindexing: show the page with a '—' or 'data not available' for missing metrics. Many cost-of-living sites do this. We don't, for three reasons.

First, a page with a '—' in the rent field and a calculator that says 'enter your own rent' is not a cost-of-living page — it's a generic calculator with a city name attached. Google's helpful-content guidelines specifically target pages that 'provide little or no value to users because the content is not specific or comprehensive.' A page missing 25% of its core data falls into that category.

Second, site-level demotion. Google's 2026 algorithm updates made clear that quality is judged at the domain level, not just the page level. A site with 1,000 complete pages and 500 incomplete pages is weaker than one with 1,000 complete pages alone. The incomplete pages drag down the complete ones.

Third, our thesis. This site's value proposition is: 'every number cited and dated.' A '—' is the opposite of a cited number. It's an absence where a promise should be. Better to not show the page in search results than to show a broken promise.

The population floor (5,000)

The gate also requires a population above 5,000 (COL_MIN_POPULATION in our constants). Cities below this threshold are noindexed even if all four metrics are present.

The reason is statistical: the ACS 5-year estimates have wider margins of error for smaller populations. A town of 2,000 might have a rent estimate with a margin of error of ±$200 — meaning the 'real' rent could be anywhere from $900 to $1,300 on a reported $1,100. Publishing a page with that level of uncertainty, without prominently disclosing the margin of error, is misleading.

Below 5,000, the data is also less likely to be complete. Small towns are more likely to have missing or quarantined metrics because the ACS sample size is too small to produce reliable estimates for every variable.

The 5,000 floor isn't arbitrary — it's the level at which the ACS 5-year estimates have reasonable margins of error for the metrics we publish. The Census Bureau itself recommends not relying on 5-year estimates for areas below certain population thresholds for specific variables.

What this means for you

If you search for a small town and can't find it on our site, it's likely because the gate excluded it — either the population is below 5,000 or one of the four required metrics is missing. The page may still exist (you can navigate to it directly), but it won't appear in search results.

For cities that pass the gate, you can trust that every metric on the page is sourced, dated, and within the published range. The explainer, the data table, the calculator, and the FAQs all use real, complete data — not estimates of estimates.

This is a deliberate tradeoff: fewer pages, but every indexed page is complete and defensible. We'd rather be the site with 996 solid city pages than the one with 10,000 pages of varying quality. In 2026, Google agrees.