Search Console's Discover Report: No Queries, No Position, and Traffic That Arrives in Spikes
The Discover report only appears once you have Discover traffic, and it drops three fields the Search report has. What each absence means and how to read the rest.
Google added the Discover performance report to Search Console in April 2019, and it behaves unlike every other report in the product. What to know before you read a row of it:
- It appears only if you have earned it: The report shows up once a property has meaningful Discover impressions in the trailing 16 months, and disappears again if they roll off. Its absence is a traffic fact, not a setup problem.
- Three fields are missing, for one reason: No queries, no average position, no search appearance. Discover is a feed pushed on interest rather than a ranked list pulled by a search, so none of the three has anything to measure.
- Pages is the only dimension that matters: Clicks, impressions and CTR by URL, country, device and date. Devices tells you almost nothing, because Discover is overwhelmingly mobile.
- Spikes are the normal shape: A story surfaces to a large audience and then stops. One URL falling out of the Pages table looks exactly like a penalty and almost never is.
- Half of SEO does nothing here: No keywords to target, no structured data that helps, no setting to switch on. The levers that exist are a large image, being indexed, timeliness, and titles that do not oversell.
Judge Discover on the Pages table over months, not on the trend line over days. The daily line is mostly the life cycle of individual articles.
CrawlRaven reads Search Console's Discover data alongside Search, GA4 and a crawl, so the last section is interested. Everything before it is how the report behaves on its own, including the parts where the honest answer is that nothing is wrong. Try CrawlRaven free: 1 site, no credit card →
Every other Search Console report answers a question about a search. The Discover report covers a surface where nobody searched at all, and almost every confusing thing about it follows from that one fact.
Here is what it holds, why three familiar fields are missing, and how to tell a story ending from a problem, which is the distinction that costs publishers the most wasted hours.
What the Discover report is
Google Discover is the personalised feed in the Google app and on the Chrome mobile home screen. Google added a performance report for it to Search Console in April 2019.
- What it reports: clicks, impressions and CTR for your content that surfaced in the feed, grouped by page, country, Discover appearance and date. There is no device breakdown and no average position.
- Where it lives: under Performance, as a separate report from Search results, alongside Google News. Each is its own search type with its own data.
- How far back: the same rolling 16 months as everything else, with no backfill, as covered in keeping Search Console data past 16 months.
It was the first Search Console report about something other than a results page, and it shows: several fields you expect are simply absent.
Where it is, and why you might not have it
The single most common question about this report is why it is not there, and the answer is unsatisfying in a useful way.
Google only shows the report once your property has meaningful Discover impressions in the trailing 16 months. There is nothing to enable, submit, verify or opt into.
Missing report, missing traffic: which is which
The important line in that table is the difference between the third and fourth rows. One URL leaving the Pages table is an article finishing its run. Every URL leaving at once, and staying gone, is the shape that deserves an investigation.
The fields Discover does not have
Open the report expecting the Search report and several things are missing. Each absence is a consequence of Discover not being a search.
The Discover report against the Search report
| Field | Search | Discover | Why |
|---|---|---|---|
| Clicks | Yes | Yes | Same definition on both surfaces |
| Impressions | Yes | Yes | A card shown in the feed counts, whether or not it was scrolled to |
| CTR | Yes | Yes | Derived from the two above |
| Average position | Yes | No | A feed has no ranked results, so there is no position to average |
| Queries | Yes | No | Nobody searched. Discover is pushed on interest, not pulled by a query |
| Pages | Yes | Yes | The most useful dimension in the Discover report by a distance |
| Countries | Yes | Yes | Same as Search |
| Devices | Yes | No | No device breakdown exists in Discover, which surprises people who expect the Search report |
| Appearance | Yes | Yes | Both have one, but different sets: Discover groups by AMP article, News Showcase, Web Story and video |
| Date range | Yes | Yes | Same rolling 16 months, same daily granularity |
Three absences and one substitution, in order of how much they cost you:
- No queries, because nobody typed one. Discover matches content to a person's interests and activity rather than to a search term, so there is no query dimension and no keyword to target.
- No average position, because a feed has no ranked results. There is no position five in Discover, so there is nothing to average, which is arguably an improvement on the metric's behaviour in Search.
- No devices, which is the one that surprises people. The Discover report has no device breakdown at all, so a question about phone against tablet performance simply has no answer here.
- A different appearance dimension. Search groups by rich-result type; Discover groups by AMP article, News Showcase, Web Story and video. Same tab, different vocabulary, and no markup earns you a place in it.
How to actually read it
With queries gone, the Pages table carries the whole report. Four habits make it useful rather than decorative.
- Start in Pages, not in the trend line. The daily chart is the sum of individual article life cycles, so it moves for reasons that have nothing to do with your site.
- Compare months, never days. A week is one story. A quarter is a pattern, and the pattern is what tells you whether Discover is a real channel for you.
- Read CTR against the image, not the copy. A card is a picture and a headline. If CTR is weak on a page that surfaced widely, the image is the first suspect.
- Cross-reference the winners with Search. A page that does well in Discover and nothing in Search is a content type worth understanding, and often a title problem rather than a topic one.
Export it monthly. Because the window rolls, a viral article that carried a quarter will quietly expire out of your year-on-year comparison, and nothing warns you when it does.
Why the traffic arrives in spikes
Search traffic is a plateau you build. Discover traffic is a series of events, and treating the second like the first produces a lot of unnecessary alarm.
- A story surfaces, then stops. Discover shows an article to a large audience over a few days and then moves on. The end of that run is not a ranking loss, because there was no ranking.
- The cliff looks identical to a penalty. Same shape on the chart, entirely different cause. The Pages table is what distinguishes them, and it takes about a minute to check.
- You cannot forecast it. No query volume exists to model against, so anyone forecasting Discover traffic is extrapolating from their own history and calling it a projection.
- You cannot slice it by device. Search Console offers no device breakdown for Discover at all, so a question about phone against tablet performance has no answer here.
I would not put Discover on a client dashboard as a headline number. Not because it does not matter, but because a chart that swings by multiples for reasons nobody controls invites a monthly conversation about a thing neither party can act on.
Report it as a page-level list instead: here are the five pieces Discover picked up this quarter, here is what they had in common. That is a finding. A line going up and down is not.
What affects whether you appear at all
Discover has no submission process, so the question is not how to get in but what makes content eligible. Google is reasonably specific about this, and a lot of standard advice does not apply.
What actually affects whether you appear in Discover
The image directive is the one people miss most often. Without max-image-preview:large in your robots meta tag, your card renders as a small thumbnail, and a thumbnail in a full-bleed feed is a card nobody taps.
The titles line is worth taking seriously too. Google's Discover guidance calls out clickbait and misleading headlines directly, and unlike most such guidance it is about a surface where the headline is most of the listing.
Discover in 2026: AI features and Search profiles
Two 2026 changes touch Discover specifically, and both are easy to miss because neither arrived in this report.
- The generative AI features report has a Discover view. When Google shipped it on 3 June 2026 it included a separate report for AI features in Discover, with the same impressions-only limitation covered in the AI performance reports breakdown.
- Search profiles feed Discover. A claimed profile lets audiences follow a creator, and Google makes followed content more likely to appear in their feed, which is a distribution route that does not depend on a query at all. The mechanics are in Google Search profiles.
- Discover got its own core update in 2026. February's was the first Discover-only core update Google has run, which makes "dating the drop" matter here as much as it does in Search. The rollout is covered in the February 2026 Discover core update.
Every one of those changes widened what Discover can do to your traffic without widening what you can do about it, which is the same pattern running through this year's Search Console changes.
A monthly read that takes ten minutes
This report does not reward daily attention and actively punishes it. Once a month is right.
- Export the Pages table. It is the only history you will have once the window rolls, and it is the whole report in practice.
- List what Discover picked up. Five to ten URLs. Look for the shared trait: format, topic, recency, image treatment.
- Check the image directive on the misses. A page in your Search-winning set that never surfaces in Discover is worth one look at its robots meta tag. The free meta tag checker reads what the page actually serves.
- Compare the quarter, not the week. If the quarter is flat and the weeks are wild, the channel is behaving normally.
- Only investigate the site-wide, sustained drop. Everything else is a story ending, and chasing it costs you a day.
Key Takeaways
- →What it is: A performance report for Google Discover, added to Search Console in April 2019. Clicks, impressions and CTR by page, country, Discover appearance and date, over the same rolling 16 months.
- →It appears only once you have traffic: Google shows the report when a property has meaningful Discover impressions in the trailing 16 months, and hides it again when they roll off. Nothing to enable or submit.
- →Two fields are missing, for one reason: No queries and no average position. Discover is pushed on interest rather than pulled by a search, so neither has anything to measure. There is no device split either.
- →Spikes are normal, cliffs usually are not a penalty: One URL leaving the Pages table is a story finishing its run. Every URL dropping at once and staying down is the only shape worth investigating.
- →Most SEO levers do nothing here: No keywords, no structured data, no settings. What Google names is a large image with max-image-preview:large, being indexed, timeliness, and titles that do not oversell.
- →Report pages, not the trend line: A chart that swings by multiples for reasons nobody controls makes a bad dashboard tile. The list of what Discover picked up this quarter is the finding.
- →Export monthly: The window rolls, so a viral article that carried a quarter expires silently out of your year-on-year comparison. Your export is the only history that survives.
Primary sources
Frequently asked questions
What is the Discover report in Google Search Console?
It is a performance report covering Google Discover, the personalised feed on the Google app and Chrome mobile home screen. Google added it to Search Console in April 2019. It reports clicks, impressions and CTR for content that surfaced in Discover, grouped by page, country, Discover appearance and date, over the same rolling 16-month window as the Search report.
When did Google launch the Discover report in Search Console?
April 2019, alongside Google's wider push to explain Discover to publishers. It was the first Search Console report covering a surface that is not a search results page, which is why it drops several fields the Search report has: there is no query and no ranked position in a feed.
Why don't I have a Discover report in Search Console?
Because your site has not had enough Discover impressions in the trailing 16 months. Google only surfaces the report for properties with meaningful Discover traffic, so its absence is a fact about your traffic rather than a configuration problem. There is nothing to enable, submit or opt into: the report appears on its own once the traffic exists.
Why does the Discover report have no queries?
Because nobody searched. Discover pushes content to people based on their interests and activity rather than matching it to a query they typed, so there is no search term to report. The same reasoning removes average position: a feed has no ranked list of results, so there is no position to average across.
Why did my Google Discover traffic suddenly drop?
Usually because one story finished its run. Discover surfaces an article to a large audience over a few days and then moves on, so a cliff in the trend line is the normal shape rather than a penalty. Open the Pages table: if a single URL fell out, that is a story ending. If every page dropped at once and stayed down, that is worth investigating.
How do I optimise for Google Discover?
There is no query to target, so most on-page SEO advice does not apply. The levers Google actually names are a large, high-quality image with max-image-preview:large, being indexed in the first place, content with genuine timeliness or topical interest, and titles that describe the page rather than overselling it. Clickbait is called out explicitly in Google's Discover guidance.
Does structured data help with Google Discover?
Not for eligibility. The Discover report does group by appearance, but those types are formats rather than schema: AMP article, Web Story, News Showcase and video. No markup earns you a place in Discover, so structured data does nothing here however useful it is elsewhere. The image preview directive matters far more: without max-image-preview:large your card renders as a small thumbnail.
Is Discover traffic worth chasing?
It depends what your site is for. Discover can send very large volumes to timely content and sends almost nothing to evergreen reference pages, and it is volatile enough that you cannot plan a quarter around it. Treat it as upside on content you were publishing anyway, not as a channel to build a business on.
Does the Discover report cover Google News too?
No, they are separate. Search Console reports Search, Discover and Google News as distinct search types, each with its own data. A property can have any combination of the three, and the News report follows the same rule as Discover: it appears only once there is traffic on that surface to report.
How far back does Discover data go?
The same rolling 16 months as the rest of the Performance report, with the oldest day dropping off every morning and no backfill. That matters more for Discover than for Search, because a single viral article is a large share of your history and it will silently expire out of your comparisons after 16 months.
15+ years of growing SaaS websites through SEO | Author, 200-Point Audit Checklist
Aditi has spent 15+ years helping SaaS companies scale organic traffic through technical SEO and content strategy. She is the author of the CrawlRaven 200-Point Audit checklist used by agencies and in-house teams to systematically improve search performance.