Advice and answers from the Advanced Web Ranking Team
Search for articles
Keyword Clusters: How the Report Works
How AWR computes Keyword Clusters from your tracked SERPs, and how to read every part of the report - KPIs, statuses, topic groups, drawer and exports.
Keyword Clusters groups your tracked keywords by the Google results they share - keywords in one cluster overlap enough that a single page can rank for all of them.
In short
The report organizes every tracked keyword in your project into clusters (one page each), rolls clusters up into topic groups (one theme each), and tells you where each cluster stands: Top 10, 11–20, Below 20, or Uncovered. Clusters are recomputed after every ranking update. Use it to find pages to build (uncovered clusters), pages to strengthen (positions 11–20), and keywords where several of your own pages compete.
Where to find it & availability
Keyword Clusters lives in the left menu under SEO → SERP Analysis → Keyword Clusters, next to SERP Similarity. It’s available on all plans except Pro.
The report needs at least one eligible search engine in the project (see below). If the project has none, the report explains what to add instead of showing empty data.
How clustering works
AWR clusters keywords by SERP overlap: keywords are clustered together when their top Google results share URLs above a similarity threshold. Two keywords whose results barely overlap are different search needs; two whose results largely overlap are the same need worded differently, and one page can rank for both.
Computed per project, from one engine. Clusters are computed for the whole project from one Google engine in it - Google desktop or mobile, with any of the Universal, AI Overviews, Organic, or Organic & Places variations. The engine used is always shown in the scope pills at the top of the report. Every position, volume and SERP in the report comes from this engine.
Recomputed on every update. Clustering runs again after each ranking update of the project, so clusters follow Google rather than aging in place.
Cluster naming. Each cluster is named after its most connected keyword - the one sharing results with the most other keywords in it. In the drawer, that keyword carries the
primarybadge.Cluster page. Every cluster gets one cluster page, which is the page of your site best placed to rank for the whole cluster. An empty cluster page means the cluster is uncovered - no page of yours ranks for these keywords yet.
Topic groups. Several clusters on one theme form a topic group. Its most important page is the pillar page - the hub its cluster pages support. A group without one has demand and no page holding it together.
Unclustered keywords. Keywords whose results don’t overlap enough with any other keyword’s stay unclustered. They don’t share a page - treat them individually. They’re always visible in the left panel and the topic-groups bar.
No keyword ceiling. Clustering covers every tracked keyword in the project.
The Keyword Clusters report in Table view.
Report scope and dates
Two pills on the date-picker row state the report’s scope - they’re informational, not controls:
All keywords · N - clusters are computed from all tracked keywords in the project. This report always covers the whole project.
Engine pill (e.g. Google US Desktop) - the search engine the clusters were computed from.
The info mark next to the pills opens a popover with the computation summary: the overlap method, the engine, and the last computed date.
The date picker works like everywhere else in AWR: the report shows the clusters from the latest update inside your selected range, and the KPI changes compare it against the previous update. There is no “All time” option - clusters are point-in-time snapshots, not cumulative data.
The KPIs
Six tiles summarize the project. The last three filter the table when clicked; the first three are informational. The Show/Hide KPIs toggle collapses the row (Table view only).
The six KPIs, with changes against the previous update, and the Topic groups bar below them.
KPI | Definition | On click |
|---|---|---|
Clusters | Keywords whose Google results overlap enough that one page can rank for all of them. | — |
Keywords clustered | How many of your tracked keywords joined a cluster. The rest are unclustered - their results don’t overlap enough with any other keyword’s. | — |
Avg cluster-page position | The average Google position of each cluster’s page, across clusters where a page of yours ranks. | — |
Clusters in top 10 | Clusters whose page ranks in Google’s top 10. | Filters to Top 10 |
Clusters in 11–20 | Clusters ranking on page 2 - also called striking distance. Usually the fastest wins available: one push from page one. | Filters to 11–20 |
Uncovered clusters | Clusters where no page of yours ranks yet. Each one is a page worth building - the searches figure is the demand waiting. | Filters to Uncovered |
Topic groups
The Topic groups bar under the KPIs shows where your keywords concentrate: each segment is a group, sized by its share of tracked keywords, with unclustered keywords as the last segment. Hover a segment for its keyword count and share; click one to scope the table to that group.
The left panel, Topic groups & clusters, is the same navigation as a tree: pick a group to see only its clusters, or a single cluster to jump straight to it. All clusters resets the scope.
The Table view
One row per cluster. Clicking a row opens the cluster drawer; the checkbox selects the row for actions. By default the table is sorted by opportunity - uncovered and 11–20 clusters with the most searches first.
Column | Definition |
|---|---|
Cluster | Each cluster is named after its most connected keyword - the one sharing results with the most other keywords in it. |
Topic group | Several clusters on one theme. Use topic groups to see where your keywords, and your demand concentrate. |
Keywords | How many keywords are in the cluster - all of them targets for the cluster’s one page. |
Volume | Combined monthly searches of the cluster’s keywords. Overlapping keywords each count in full, so read it as the cluster’s ceiling, not a traffic forecast. |
Intent | The mix of search intents across the cluster’s keywords, as a stacked bar. Hover the bar for the breakdown; a spread across several colours means mixed intent: check that one page can serve all of it. |
Avg. position | The average position of your best-ranking page across the cluster’s keywords, on the search engine the clusters were computed from. |
Visibility % | The share of the cluster’s possible visibility your site captures, from your positions across its keywords. |
Est. traffic | Estimated monthly organic visits from the cluster’s keywords, from volume and CTR at your positions. |
Top 10 | How many of the cluster’s keywords your site ranks in Google’s top 10 for, as |
Cluster page | The page of your site best placed to rank for the whole cluster. Empty means uncovered - no page of yours ranks for these keywords yet. |
Status | Where the cluster stands: Top 10 · 11–20 · Below 20 · Uncovered |
Statuses
Top 10 - your cluster page ranks in Google’s top 10.
11–20 - ranks on page 2, striking distance. One push from page one.
Below 20 - your best page for this cluster ranks below position 20.
Uncovered - no page of yours ranks for this cluster’s keywords yet.
The clusters table, sorted by Avg position, where each row is a cluster with its page and status.
Filters and search
The search field matches cluster names and every keyword inside them; searching a keyword finds the cluster it belongs to.
While search or filters are active, the toolbar shows the match count (e.g. 8 of 80 clusters).
Add filters offers six cluster-level filters:
The cluster-level filters, grouped under SEO and Clusters categories.
Filter | What it does |
|---|---|
Status | Where each cluster stands today; same values as the Status column. |
Topic Group | Show only clusters from the selected topic groups. |
Volume | Combined monthly searches of the cluster’s keywords - min/max range. |
Avg. Position | The average position of your best-ranking page across the cluster’s keywords. Presets: Top 10 · 11–20 · Ranked · Not ranked. |
Cluster Page | Match against the cluster page URL - e.g. |
Search Intent | The intent of each keyword in the cluster, detected from its live Google results. Any includes clusters with at least one keyword of the selected intents; Mixed finds clusters where no single intent dominates - one page may struggle to serve all of them. |
The Cards view & topical map
The Cards tab is the scannable version of the same data, opening with the topical map: every topic group drawn as a bubble sized by search volume, its clusters inside. By default the map shows the top 10 topic groups by volume; clicking on + N more expands to all of them. The map answers one question fast: where is the demand?
Below it, cards summarize either Topic groups (pillar page, total volume, clusters and their statuses) or individual Clusters; switch between them using the toggle options. The pillar page shown on each group card is the most important page of the group - the hub its cluster pages support.
The Cards view showing the topical map of the top 10 topic groups by volume, with topic group cards and their pillar pages below.
The cluster drawer
Click any row (or card) to open the cluster’s drawer. The header carries the cluster name, its status and intent chips, its topic group, and a date chip naming the update you’re looking at - the latest update inside your selected range. Everything in the drawer is read from that one snapshot.
The cluster drawer - KPIs, your ranking pages with the cluster page badged, the cluster’s keywords with the primary one marked, and the Common URLs ranking across its SERPs.
Four KPIs repeat the row’s numbers - Keywords, Volume, Avg position, Top 10 - with the same definitions as the table columns. Below them, three sections:
Your pages in this cluster
Every page of yours ranking for at least one keyword in this cluster, with how many of the cluster’s keywords it ranks for and its best and average positions. The cluster page is badged. Several pages ranking here means they compete for the same searches - consider consolidating. If no page of yours ranks yet, this section shows an empty state and points you to Common URLs - the shape a new page could take.
Keywords in this cluster
Every keyword in the cluster, with intent, volume, CPC and your position. The keyword marked primary is the cluster’s most connected one - it gives the cluster its name.
Common URLs
Pages that appear in Google’s results across this cluster’s keywords — yours badged You, tracked competitors badged Competitor. SERP coverage counts how many of the cluster’s keywords each page ranks for - a page covering most of them is proof one page can rank for the whole cluster. Under the list, chips summarize the SERP features present across the cluster’s results (Organic, AI Overviews, Videos, Reviews, People also ask, Sitelinks, Images…), each with the number of SERPs it appears in.
Keywords in the cluster and their Common URLs; pages covering 2/2 of the cluster's SERPs prove one page can rank for the whole cluster.
Saving clusters & exporting
Save as keyword group
Select one or more clusters with the row checkboxes and choose Save as keyword group in the bar that appears. Each selected cluster becomes its own keyword group, named after the cluster - groups are never merged. The group is a static copy: it won’t change when clusters recompute, so segments you report on stay put. If a group with that name already exists, you can add the keywords to it or save under a suffixed name.
Selection actions - save the selected clusters as keyword groups, or export their keywords.
Export cluster keywords
The second action in the selection bar exports one row per keyword of the selected clusters: keyword, cluster, topic group, volume, CPC, intent, your position, your ranking URL, cluster status and cluster page. This is the file you hand to whoever writes or optimizes the page.
Export the entire cluster list
The toolbar’s Export menu allows to download the cluster list as CSV or XLSX, or upload it to a Google Sheet. It preserves your active filters and column settings, and the file header records the project, engine, last computed date, range and applied filters - so the spreadsheet still explains itself weeks later.
If you'd rather keep the data live instead of exporting a snapshot, use Share to generate a link others can open directly, without needing an AWR account. Choose Interactive to let recipients explore the data themselves, or View Only for a fixed, read-only version.
Export & Share actions - download the cluster list or generate a shareable link
Changes between updates
Clusters have no fixed identity across updates - Google’s results move, and the clustering follows. So change is shown at project level, not per cluster: the KPI tiles compare against the previous update (new cluster counts, coverage shifts, position movement), and the drawer always shows one update’s snapshot, named by its date chip. Cluster snapshots are kept for 12 months.
If you need a stable segment to track over time, that’s exactly what Save as keyword group is for.
Frequently Asked Questions
I don't see any clusters. Why?
Why is a keyword unclustered?
Why did my clusters change since last week?
Can I edit a cluster - merge, split, or rename it?
Which keywords are included?
Does the report work for client reporting?
stay in the loop