The common belief is that domain properties and URL-prefix properties provide the same data at different scope levels, making cross-property comparison straightforward. This is wrong. Domain properties and URL-prefix properties apply fundamentally different URL resolution, subdomain inclusion, and protocol handling rules that cause the same query to return different impression counts, click totals, and average positions. Programmatic comparisons across property types produce reconciliation errors that range from 5% to over 40% depending on site architecture, and these errors compound when used for trend analysis or performance benchmarking.
How Domain Properties and URL-Prefix Properties Mechanistically Scope and Aggregate Data Differently
A domain property aggregates data across every URL variation under a root domain: all subdomains (www, blog, shop, app), all protocols (HTTP and HTTPS), and all URL paths. When Google records an impression or click for any URL matching that root domain, the metric appears in the domain property’s data regardless of which specific subdomain or protocol served the result.
A URL-prefix property scopes data to a single protocol-subdomain-path combination. A property configured for https://www.example.com/ captures only data for URLs matching that exact prefix. Impressions generated by http://www.example.com/, https://blog.example.com/, or https://example.com/ (without www) are excluded entirely.
The aggregation difference manifests most clearly in impression deduplication. When Google shows two URLs from the same domain in a single search result page, the domain property counts this as one impression (property-level aggregation). Two separate URL-prefix properties for those URLs each count one impression independently. Summing impressions across URL-prefix properties therefore double-counts queries where multiple subdomains appear in the same SERP.
Click attribution follows different rules. A domain property attributes all clicks on any URL under the root domain to a single property-level metric. URL-prefix properties attribute clicks only to the specific prefix that matches the clicked URL. If Google’s canonical selection assigns a click to a URL under a different prefix than expected (for example, crediting http:// instead of https://), the click appears in the domain property but not in the URL-prefix property the analyst expected.
For sites with simple architectures (single subdomain, HTTPS only, no protocol variants), the practical divergence between property types is small, typically under 5%. For sites with multiple active subdomains, ongoing protocol migrations, or legacy HTTP pages still indexed, divergence routinely exceeds 20%.
Specific Data Reconciliation Failures When Comparing Metrics Across Property Types
The intuitive expectation that summing all URL-prefix property metrics should equal the domain property total fails for three mechanical reasons.
Impression deduplication asymmetry. The domain property deduplicates impressions at the property level. If www.example.com/page-a and blog.example.com/page-b both appear for the same query, the domain property records one impression. The www prefix property records one impression. The blog prefix property records one impression. The sum is two, not one. The deduplication gap grows proportionally with the number of subdomains that rank for overlapping queries.
Canonical URL attribution mismatch. Google’s indexing system selects a canonical URL for each page, and the selected canonical determines which prefix property receives the data. If Google canonicalizes https://www.example.com/page to https://example.com/page (dropping www), clicks and impressions for that page appear under the non-www prefix property. The www prefix property shows no data for that page. The domain property shows data for both canonical versions because it encompasses all prefixes. This creates situations where the domain property shows clicks for a page that no URL-prefix property claims.
Position averaging discrepancies. Average position calculations differ between property types because the denominator (the set of impressions being averaged) differs. The domain property averages position across all impressions for all URLs under the domain for a given query. A URL-prefix property averages position only across impressions for URLs matching that prefix. When a domain ranks multiple URLs from different subdomains for the same query at different positions, the domain property reports the highest (best) position, while individual prefix properties report their own respective positions.
These reconciliation failures are not data errors. They are mathematical consequences of different scoping rules applied to the same underlying search event data. Attempting to force reconciliation introduces systematic bias.
The Subdomain and Protocol Edge Cases That Produce the Largest Cross-Property Divergences
Several site architecture patterns maximize the divergence between property types, making cross-property comparison particularly unreliable.
Active www and non-www variants. Sites where both www.example.com and example.com are indexed produce the largest divergences. Google may canonicalize some pages to the www version and others to the non-www version based on internal linking patterns and crawl history. A URL-prefix property for either variant misses the pages canonicalized to the other. The domain property captures both, producing totals that can exceed either prefix property by 30-40%.
HTTP to HTTPS migration residue. After an HTTPS migration, legacy HTTP URLs may remain in Google’s index for months or years. The domain property continues to aggregate data from these HTTP impressions. An HTTPS URL-prefix property does not see this traffic. For sites with incomplete migrations or sites where Google is slow to recrawl and update canonical selections, this residue can account for 10-20% of domain property impressions.
Multiple active subdomains. Sites operating subdomains for blog content (blog.example.com), documentation (docs.example.com), regional variants (uk.example.com), or applications (app.example.com) create unique deduplication scenarios. Each subdomain requires its own URL-prefix property. The domain property’s impression deduplication across subdomains means that query-level comparison between any single prefix property and the domain property will show the domain property reporting fewer impressions per query than the prefix (because the domain deduplicates while the prefix does not).
Internationalized domain variants. Sites using subdomain-based internationalization (de.example.com, fr.example.com) with hreflang tags can trigger cross-subdomain impression sharing. The domain property aggregates these shared impressions under a single metric, while individual country-subdomain prefix properties each record their localized subset.
Safe Comparison Methodologies for Cross-Property Type Analysis
When cross-property comparison is unavoidable, specific controls reduce the error magnitude to acceptable levels.
Trend comparison instead of absolute value comparison. Percentage changes over time are more comparable across property types than absolute click or impression counts. If the domain property shows a 15% impression increase and the primary URL-prefix property shows a 13% increase, the directional signal is consistent even though the absolute numbers differ. Trend comparison neutralizes the constant offset introduced by scoping differences.
Dimension-level matching. When comparing query-level data between property types, filter both requests to the same country, device, and search type dimensions. Dimension filtering reduces the scope of data that can diverge between property types, narrowing the comparison to more comparable subsets.
Click-through rate as a normalized metric. CTR calculated from the same property type is internally consistent regardless of how impression deduplication works, because the numerator (clicks) and denominator (impressions) are both affected by the same scoping rules. Comparing CTR across property types is more reliable than comparing raw click or impression counts.
Avoid mixing property types in a single analytical dataset. Never combine query rows from a domain property extraction with query rows from a URL-prefix extraction in the same analysis table. The different aggregation rules make the rows incompatible. Build analyses on a single property type throughout.
When to Standardize on a Single Property Type and Which Type to Choose
The cleanest approach to eliminating cross-property comparison errors is standardizing all programmatic analysis on a single property type and using the other only for diagnostic purposes.
Choose the domain property when: the site operates multiple active subdomains, protocol migrations are ongoing or recently completed, the analytical priority is comprehensive visibility across the entire domain, and subdomain-level analysis can be performed by filtering the domain property data by URL prefix patterns.
Choose a URL-prefix property when: the site uses a single subdomain and protocol with no migration residue, the analytical priority is precision for the primary site experience, integration with Google Analytics requires the URL-prefix property (GA4 linking is available only for URL-prefix properties in some configurations), and backlink disavow functionality is required (this feature only operates on URL-prefix properties).
For most enterprise SEO operations, the domain property is the better default because it prevents data blind spots created by canonical selection surprises. The domain property reveals when Google is attributing traffic to unexpected URL variants, providing diagnostic information that URL-prefix properties structurally conceal. Supplement the domain property with URL-prefix properties only when specific functional requirements (GA4 linking, disavow) demand them, and never mix the two in quantitative analysis.
Can you filter a domain property by URL prefix to replicate URL-prefix property data exactly?
Not exactly. Filtering a domain property by URL prefix pattern narrows the data scope but does not replicate the impression deduplication behavior of a URL-prefix property. The domain property applies property-level deduplication before the filter, while a URL-prefix property applies prefix-level scoping before deduplication. The filtered domain data and the prefix property data will still diverge for queries where multiple subdomains rank simultaneously.
Does switching from a URL-prefix property to a domain property cause historical data loss?
No historical data is lost in either property. Both property types retain their own independent data histories. However, the domain property only has data from the date it was verified, not retroactive data from before verification. Organizations switching to domain property analysis should verify and begin extracting from the domain property as early as possible to build historical coverage alongside existing URL-prefix data.
How does Google’s canonical URL selection affect which property type receives click attribution?
Google attributes clicks to the canonical URL it selects during indexing, which determines which URL-prefix property records the data. If Google canonicalizes a page to a protocol or subdomain variant different from the expected prefix, that prefix property shows no clicks for the page. The domain property captures all clicks regardless of canonical selection because it encompasses every URL variant under the root domain.
