Site Impression vs URL Impression in Looker Studio

When you connect Google Search Console to Looker Studio, the connector asks you to choose between two tables: Site Impression and URL Impression. The labels can make this look like a minor setup choice, but each option represents a different way of aggregating the same search performance data.

Site Impression groups performance at property level. URL Impression groups it at page level. This affects how impressions, clicks, click-through rate and average position are calculated when more than one page from your site appears for the same query.

Both can power top-level scorecards, but I prefer URL Impression as the main reporting source. It shows the contribution of individual pages, usually returns larger impression totals because each qualifying URL can be counted, and is the only one of the two options that supports Discover and Google News.

Site Impression aggregates results by Search Console property. If two or more URLs from the same property appear on one results page for the same query, the property records one impression rather than a separate impression for every URL.

Position is based on the highest-ranking result from the property. If your pages appear in positions two, five and eight, the property-level position for that appearance is two. This makes Site Impression useful when the specific reporting question is how often the property appeared, with multiple URLs from that property deliberately deduplicated.

Because the page is not the unit of aggregation, this table is not designed to explain the full contribution of individual landing pages. Its lower totals are not necessarily wrong; they represent a narrower property-level counting method.

URL Impression aggregates performance by canonical page. When several pages from the property appear in the same set of results, each unique URL can record its own impression, click and position.

This table exposes the landing-page dimension, making it possible to evaluate individual pages, templates, directories and content groups. It can power top-level scorecards as well as detailed page tables, so it is not limited to landing-page reporting.

I consider URL Impression the more accurate representation of how the site's content is performing because it preserves the contribution of each page instead of deduplicating the property. It also supports Discover and Google News as separate selectable search types, extending the analysis beyond the four search types available with Site Impression.

Search Console generally assigns performance to the canonical URL selected by Google. The URL shown in the data may therefore differ from the exact URL a user saw or the URL you expected to receive credit.

Both tables provide the familiar Search Console metrics: clicks, impressions, CTR and average position. The difference is the level at which those metrics are counted.

A common mistake is to place Site Impression and URL Impression scorecards next to each other and assume that one is wrong when the numbers differ. A difference can be expected because the two sources are answering different counting questions.

Imagine one search results page contains three URLs from the same property. At property level, that appearance can count as one impression and use the highest position. At page level, it can count as one impression for each of the three URLs, with a separate position assigned to each page.

This is why URL Impression totals are usually higher and why CTR or average position can differ from Site Impression. The gap is not a Looker Studio calculation error. It is a consequence of property aggregation versus page aggregation.

I also validate this by exporting the Performance report from Google Search Console to Google Sheets. When I review the page-level data in the Pages sheet, the larger impression total aligns with the URL Impression result in Looker Studio. The property-level Chart sheet can still show the lower, deduplicated total, so it is important to compare like with like.

I prefer URL Impression because it gives me a more complete and actionable representation of performance. I can use the same aggregation for top-level metrics and then move into the pages responsible for those results without changing the meaning of the data.

The impression total is usually bigger because multiple qualifying URLs are counted separately. For me, that is valuable information rather than duplication to remove: it shows how much visibility the site's individual pages collectively earned.

URL Impression also supports Discover and Google News. These are separate search types rather than figures automatically added to Web, but their availability makes URL Impression the stronger foundation for reporting content performance across more Google surfaces.

Site Impression still has a valid specialised use. I use it when I specifically want a deduplicated property-level view or need to reconcile a figure with the property-level chart in Search Console.

I use URL Impression as the primary data source for scorecards, trend charts and landing-page analysis. If the report also needs a deduplicated property-level comparison, I add Site Impression as a second, clearly labelled source rather than treating it as the default.

Names such as “GSC — URL Impression — Web” and “GSC — Site Impression — Web” make the aggregation visible to anyone editing the report later. Discover and Google News should also be identified clearly because each is a separate search type.

Site Impression and URL Impression are not interchangeable versions of the same table. Site Impression explains how the property appeared in search. URL Impression explains how individual pages appeared.

My preference is URL Impression because it provides the more complete view of content performance, its usually larger totals reflect the visibility earned by each page, and it supports Discover and Google News reporting. Exporting page-level Search Console data to Google Sheets provides a useful way to validate those figures.

Site Impression remains useful for the narrower question of deduplicated property visibility. Whichever source is used, label the aggregation and search type clearly so readers understand exactly what the metrics count.