Exploring the Discover feed

The Explorer is the core analysis surface — your live view of Google Discover. Open it from the Analysis group in the left sidebar and you land on Discover explorer, titled with a LIVE dot and an updated HH:MM stamp so you always know how fresh the view is. Everything on the page answers one question: what is Discover surfacing right now, in this market, and which stories are on the way up. This guide walks the page top to bottom — the filter bar, the Est Impressions chart, the publications table, and the right rail — and then teaches you how to read the feed to catch a rising story early.

#The filter bar

Every number on the page is scoped by the filter bar across the top. Set it once and the chart, the table, and the right rail all recompute together.

• Market — the country you are looking at, shown as a flag and name. Discover is polled per market, so the feed changes completely when you switch. Pick the market you care about before reading anything else.
• Language — narrows the feed to publications in one language within that market. Useful in multilingual markets where you only want, say, the local-language agenda.
• Time window — the buttons 24H, 7D, 14D, 30D, and Custom. This is the single most important control: impressions, deltas, and the recency buckets are all measured over the window you choose. A short window like 24H shows what is hot today; a longer window smooths out the noise.
• Compare — a toggle that lines the current window up against the prior one, so you read change rather than a raw level.
• Filters — the button that opens finer constraints on top of the basics above — narrow the feed down before you start reading.

#The Est Impressions chart

Below the filter bar, the Est Impressions chart shows how much estimated reach Discover is handing out, and to what. Impressions (url-reach) is D2TR's estimate of how much exposure a URL gets in the feed; rolled up here, it tells you where attention is going as a whole. The chart has two tabs:

• Groups — impressions split across the five format groups (below). This is the fastest read on the shape of the feed: is Discover leaning on hard news, or on service and lifestyle content right now?
• Types — the same reach broken down into the finer format types underneath those groups, for when a group-level move is too coarse to act on.

#The five format groups

A format is how an article is packaged editorially — not what it is about. Every publication rolls up into one of five groups, each shown with a value and a delta versus the prior window:

• News — straight reporting: breaking stories, briefs, developing events.
• Analysis — explainers, opinion, deep dives that interpret rather than report.
• Service — practical, how-to and utility content: guides, deals, listicles.
• Atmosphere — mood and lifestyle pieces: human interest, culture, the softer feed.
• Ritual — recurring, calendar-driven content: horoscopes, sports fixtures, daily columns.

Reading the deltas next to each group tells you what Discover is favoring today. A big positive delta on News usually means a live event is pulling the feed; a swing toward Service or Atmosphere often marks a quieter, evergreen day. Switch to the Types tab when you need to know exactly which kind of packaging is winning inside a group.

#The publications table

The heart of the Explorer is the publications table — one row per publication we indexed in the window. This is the raw feed, and learning to read its columns is most of what makes the Explorer useful.

#What each column means

• Title — the publication's headline. English-translated where the original is in another language, so a foreign market is still readable at a glance.
• Publisher — the domain the piece came from. Repeated appearances by one publisher on a single story is itself a signal.
• Main entity — the primary person, organization, place, or product the article is about, linked into the Knowledge Graph. This is the fastest way to see what a row is really about without reading the title.
• Impressions — estimated reach for that publication over the current window. This is the size number — sort or scan on it to find what is actually getting seen, not just what was published.
• Pulse — a mini trend sparkline of that publication's reach over time. Shape matters more than the absolute value: a line still climbing at the right edge is live; one that has already rolled over has peaked.
• Group — the format group (News, Analysis, Service, Atmosphere, Ritual) the row belongs to.
• Category — the Google category the article was classified into — the topical bucket (sport, finance, tech, and so on), independent of its format.
• Type — the finer format type beneath the group, for when the group label is too broad.
• Assigned trend — the trend cluster this publication was grouped into. When several rows share one assigned trend, you are watching a story spread across publishers in real time — click through to the trend to see the whole cluster.

#Recency buckets

Rows are grouped into recency buckets — Past 1 hour, Past 24 hours, and so on — rather than shown as one flat list. The buckets are the timeline of the feed: the top bucket is what Discover picked up most recently. When you are hunting for something rising, the top buckets are where it shows up first, before it has had time to accumulate impressions.

#Search, Columns and Export

Three controls sit above the table:

• Filter input — a search box over the current rows. Type an entity, publisher, or keyword to pull the feed down to just what you are tracking.
• Columns — choose which columns are visible, so you can focus the table on the two or three signals you actually read.
• Export — pull the current, filtered view out of the app for a report or downstream analysis. What you export matches exactly what the filter bar and search box have narrowed the table to.

#The right rail

The right rail is the leaderboard view of the same window — three ranked lists, each with values, deltas, and a View all link:

• Top entities — the people, organizations, places and products getting the most reach right now. A new name climbing here is often the first sign of a story before it is obvious in the table.
• Top publishers — the domains winning the most impressions in the window. Watch for publishers that don't usually rank — a new entrant is one of the strongest early signals in the feed.
• Top trends — the trend clusters with the most reach, ranked by size and change. This is your shortlist of what is actually moving; use View all to jump into the full trend view.

#Reading the feed to spot a rising story

Put the pieces together and the Explorer becomes an early-warning system. Here is a repeatable pass:

1. Set the market and a short window
   Pick your market, then set the window to 24H and turn on Compare. You want the freshest slice with change visible against yesterday.
2. Read the chart for the shape of the day
   Glance at Est Impressions on the Groups tab. A big positive delta on News means an event is pulling the feed; a quiet, service-heavy day means you are hunting for something smaller and earlier.
3. Scan the right rail for new names
   Look down Top entities and Top publishers for anything unfamiliar with a large positive delta. A new entity or an unusual publisher climbing fast is the earliest sign that something is starting.
4. Confirm in the top recency buckets
   Drop into the table and read the Past 1 hour and Past 24 hours buckets. Use the filter input to isolate the entity or publisher you spotted, and watch whether more rows keep arriving.
5. Check the Pulse and the Assigned trend
   On those rows, the Pulse sparkline should still be climbing at its right edge — that is a live story, not a spent one. If several rows share one Assigned trend, the story is spreading across publishers; click the trend to see the whole cluster and its momentum.
6. Export or track it
   When you have confirmed a story is rising, Export the filtered view for your report, or put its entity, publisher, or trend on your radar so the app watches it for you going forward.

#Where to go next