Impressions move before clicks do. A sustained dip here is the first quiet warning that a ranking change, an indexing problem, or a seasonal swing is working its way towards your organic traffic.
At a glance
Impressions Trend plots the daily count of Google organic-search impressions for the connected Search Console property across a rolling 90-day window. An impression is logged each time any URL from your site appears in a search result a user could see (whether or not they scroll to it). Because impressions react to ranking and indexing shifts a step earlier than clicks, this card is an early-warning trend, not a revenue number. It fires a sensitivity alert when impressions drop more than 20% versus the prior comparable period.
| What it tracks | Daily total impressions from the Search Console Search Analytics API (searchType=web, no query/page dimension), summed across all queries, pages, countries and devices for the property, then drawn as a 90-day area trend. |
| Data source | Google Search Console Search Analytics API (searchanalytics.query), the same dataset behind the Performance report’s Impressions metric. Vortex IQ pulls the daily series and caches it on the standard refresh. |
| An impression means | Your URL appeared in a results page the user reached. For most results that is on the rendered page; for results below the initial viewport, an impression is counted when the user scrolls them into view or the result is otherwise present. One URL can earn at most one impression per results page per search. |
| Time window | 90D rolling. The card compares the latest period against the prior equal-length period (vsP) to compute the trend and alert. |
| Alert trigger | drop >20% vsP. Sentiment key gsc_impressions_trend. A 20% floor is set above normal day-to-day organic noise (which is typically 5 to 15%) so the alert flags structural change, not weekday seasonality. |
| Why 20%, not 15%? | Impressions are noisier than clicks because they include deep-SERP appearances that fluctuate with every minor ranking wobble. Clicks Trend alerts at 15%; impressions need a wider band to avoid false alarms. |
| Data freshness | Search Console data is typically 2 to 3 days behind real time. The most recent 2 to 3 days of the trend will be partial or absent and should not be read as a drop. |
| Roles | owner, marketing |
Calculation
The card requests a date-dimensioned Search Analytics query for the property over the trailing 90 days:{ date, clicks, impressions, ctr, position }. The card keeps the impressions value per day and renders the series as an area chart. No query or page dimension is requested, so the totals are not subject to the row-anonymisation that affects query-level reports: a property-level impressions total includes the long tail of rare queries that would be hidden if you broke the same query down by search term.
The trend and alert are derived by splitting the window in two: the latest period (for a 90-day card, typically the most recent 7 to 14 days, configurable per profile) is compared against the immediately preceding equal-length period. The percentage change is (latest − prior) ÷ prior. When that change is a fall of more than 20%, the card raises its sensitivity flag. See the worked example for a typical reading.
Worked example
A UK home-fragrance brand on Shopify, Search Console connected to the domain propertysc-domain:scenthaus.co.uk. The team is watching organic visibility after a site re-platform on 12 May 26.
| Period | Dates | Total impressions | Daily avg | vs prior |
|---|---|---|---|---|
| Prior 14 days | 28 Apr 26 to 11 May 26 | 1,184,000 | 84,571 | baseline |
| Latest 14 days | 12 May 26 to 25 May 26 | 902,000 | 64,429 | down 23.8% |
- The alert fired on the morning of 28 May 26. The 14-day window ending 25 May 26 (the freshest fully settled data, three days behind) showed a 23.8% fall, clearing the 20% floor. The card’s area chart shows a clean step-down starting 13 May 26, the day after the re-platform went live.
- Impressions fell before clicks. The companion Clicks Trend card showed only a 9% dip over the same window, still under its 15% alert. This is the textbook pattern: pages slipped from page 1 positions 6 to 10 down to page 2, which strips most of their impressions but only a few clicks (those deep positions earned little CTR anyway). Reading impressions first gave the team a two-day head start.
- The cause was a canonical regression, not an algorithm update. The re-platform emitted self-referencing canonicals pointing at the old URL structure for collection pages. Google began consolidating those pages onto URLs that 404ed, so they dropped out of the results they used to appear in. Confirming the trend against Index Coverage Trend showed a matching rise in “Excluded by canonical” pages, which pinned the root cause to the deploy rather than to a wider ranking shift.
Sibling cards to read alongside
| Card | Why pair it with Impressions Trend |
|---|---|
| Total Impressions | The single-number version of this trend, with today / 7D / 30D comparisons. Use it for the headline figure; use this card for the shape. |
| Clicks Trend | The downstream signal. Impressions usually move first; if clicks follow within a few days, the ranking change is real and reaching users. |
| Average Position | Explains why impressions moved. A rising (worsening) position alongside falling impressions confirms a ranking slip rather than reduced search demand. |
| Average CTR | Disambiguates demand from packaging. Flat CTR with falling impressions means fewer appearances, not worse snippets. |
| Index Coverage Trend | The indexing cross-check. A concurrent fall in indexed pages turns “impressions dropped” into “pages left the index”. |
| Impressions by Country | The geographic breakdown. Tells you whether a drop is global or concentrated in one market. |
| Ranking-Drop Alert | The real-time companion. If a top-10 query has tumbled, this is often the query behind a property-wide impressions dip. |
Reconciling against the source
Where to look in Google Search Console:Performance → Search results, with the Impressions metric toggled on and the date range set to Last 3 months (Search Console’s nearest preset to a 90-day window). The chart there is the canonical reference for this card.To match Vortex IQ cleanly:
- Set the Search Console date range to the same start and end dates the card uses (remember the card stops 2 to 3 days short of today).
- Leave the search-type filter on Web if the card is reading web results; Image, Video and News are separate impression pools and are excluded here.
- Do not apply a query or page filter in Search Console; this card is the property-wide total.
| Reason | Direction | What to do |
|---|---|---|
| Data delay. Search Console lags 2 to 3 days. If you compare our latest day against the UI on the same calendar day, both may be partial. | Both low at the trailing edge | Compare fully settled days only (4+ days old). |
| Anonymised queries. The UI’s query-level total omits rare, personally-identifying queries; the property-level total (which this card uses) includes them. | UI query view lower than this card | Reconcile against the unfiltered Performance total, not the summed query rows. |
| 1,000-row UI cap. The Performance report shows at most 1,000 rows per dimension; summing visible rows understates the true total. | UI sum lower | Use the headline metric, or the Search Analytics API, not the row sum. |
| Search-type scope. The UI defaults to Web; if you switch to All or a different type the impression pool changes. | Variable | Match the search type. |
| Property type. Domain properties aggregate all subdomains and protocols; a URL-prefix property does not. | Domain property higher | Confirm both sides point at the same property. |
dimensions: ["date"] and no filters, and sum the impressions column over the same dates. That is the same call Vortex IQ makes.
Known limitations / FAQs
Why did the alert fire when my rankings did not change? Impressions also fall when search demand falls. A seasonal product, a holiday lull, or a one-off news spike rolling out of the prior period can all cut impressions without any ranking movement. Cross-check Average Position: if position held steady while impressions dropped, the cause is demand, not ranking. My impressions jumped overnight with no extra clicks. Is that good? Not necessarily. A sudden impressions spike with flat clicks usually means you started ranking for a batch of low-relevance or deep-SERP queries (often page 2 or 3). It inflates impressions and depresses CTR without adding traffic. Check Average CTR and CTR by Position Bucket. The last two or three days of the chart dip every time. Is my site declining daily? No. That is the standard Search Console reporting delay. The most recent 2 to 3 days are always partial because Google has not finished processing them. Read the trend up to the last fully settled day. Does this include image, video, and news appearances? No. This card reads web search results only (type=web). Image, Video, News and Discover are separate impression pools in Search Console and are not summed in. If you publish heavily to those surfaces, treat this card as web-only visibility.
Why is the property total higher than the sum of my top queries?
The Performance report caps at 1,000 rows and anonymises rare queries, so summing query rows always understates the true total. This card reads the unfiltered property-level series, which includes the entire long tail. The gap between the two is normal and often large for sites with broad keyword coverage.
Can I change the 20% alert threshold?
Yes. The sensitivity floor is configurable per profile in the Sensitivity tab. Brands with volatile, demand-driven catalogues (fashion, seasonal gifting) often widen it; stable B2B sites often tighten it. Tune it to your own baseline noise rather than the generic default.