Keywords & clusters

The keywords table

The Keywords page lists everything you track for the client, with columns for Cluster, Volume (monthly searches), CPC, and Position (latest ranked position, with a dash until the first nightly fetch). Add keywords in bulk — one per line, up to 100 at a time — and remove any you no longer want tracked. Positions, volume, and CPC are refreshed nightly; see Analytics & rankings for where those numbers come from.

Clusters: grouped by meaning, not by string

Every night, an AI-assisted pipeline reads your tracked keywords and groups the ones that mean the same thing — even when they share no words. “Waterproof hiking boots” and “rain-proof trail footwear” land in the same cluster, while a plain text filter would call them strangers. Each cluster then receives a short generated name from its member keywords. Keywords with no clear group stay unclustered rather than being forced somewhere they don’t belong.

Cluster detail pages

Click any cluster to open its detail page: the full member-keyword list plus cluster-level aggregates — total search volume across members and the average CPC where CPC is known. Use it to judge whether a topic is worth a content investment before writing a single brief.

Share of Voice

The Share of Voice tab estimates how much of the available search demand your client actually captures. The model:

• For each tracked keyword, your position is converted into an expected share of clicks using a click-through-rate curve, weighted by the keyword’s search volume.
• The sum of those expected clicks, divided by the clicks a site ranking #1 on everything would get, is your Share of Voice percentage.
• Where competitor ranking data is available, the same arithmetic produces the SOV vs Competitors chart, and SOV by Cluster breaks the estimate down per topic.

The page is dated (“Snapshot as of …”) and states its own coverage: “N of M keywords have volume data”.

Good to know

• Clustering quality grows with list quality. Tracking your actual target queries — not every permutation a tool can generate — produces tighter clusters and a more honest Share of Voice.
• The demo client’s keywords and clusters are hand-authored sample data and cannot be edited (“The demo client is read-only.”).
• Share of Voice is an estimate built from positions, volumes, and a CTR curve — a directional instrument for comparing periods and competitors, not an audited traffic figure.

FAQ

Why is a keyword showing no cluster?

Clustering runs nightly and only groups keywords whose meaning clearly belongs together. Genuinely one-off keywords stay unclustered on purpose — forcing them into a group would make the clusters worse, not better.

Can I rename or manually edit a cluster?

Not currently. Cluster names are generated from the member keywords and refresh with the nightly run.

Share of Voice says only some keywords have volume data. Is it broken?

No — the page states it plainly: "N of M keywords have volume data". Keywords without a known search volume cannot contribute to a click-share estimate, so they are counted but not weighted.