Keyword Universe

Insights

Keywords

• Universe
• Playground

Insights

Keywords

• the demand map behind your topics
• test any category, changes nothing (beta)

Overview#

The Keyword Universe is the map of search demand around your category: every keyword Seerly can find, grouped into clusters, with each cluster rated for how well it fits your brand.

Your first set of topics was derived from this map during setup, steered by the focus areas you entered. This screen is where you correct and extend that set afterwards, which makes it the one place where you can change what Seerly measures at all.

Panels and controls#

Country scopes the entire map. Volumes and competition figures describe that market and no other, so a universe built for the wrong country will hand you a coherent, confident and thoroughly useless picture of somewhere you do not sell.

Below that, a Cluster groups related keywords under a single name. The table gives you four columns: the Cluster itself, its Keywords count, its Monthly searches, and its Priority. A fifth, Matches, appears while you are searching and counts how many keywords in each cluster matched what you typed.

Competition is not a cluster-level figure and does not appear here. It belongs to individual keywords, shown as LOW, MEDIUM or HIGH once you open a cluster, which is the right grain for it: a cluster routinely contains both contested head terms and uncontested long-tail ones, and averaging them would describe neither.

Priority is the one to understand, because it is a recommendation rather than a score. A cluster is marked Target where searchers are squarely your audience, Consider where they are a likely customer, and Probably skip where they are unlikely to become one. That last label is hedged on purpose: it is a judgement about what to go after rather than a verdict, and those clusters stay on screen rather than being hidden, because knowing what adjacent demand exists is useful even when you have no intention of chasing it.

Topic Clusters

Cluster	Keywords	Monthly searches	Priority
project management software	240	18,100	Target
agency time tracking	86	2,400	Target
free gantt chart tools	154	9,900	Probably skip
client onboarding checklist	31	590	Consider

Sort by reorders the map, defaulting to Recommended and offering Highest priority, Most searches, Most keywords and their opposites, along with Cluster A–Z. Filter by cluster narrows to one, with All clusters to go back.

Minimum monthly searches hides the low-volume end, and this one deserves a light touch: buyer-intent demand is very often low volume, so filtering hard on searches tends to strip out precisely the clusters that convert. Hide shallow clusters drops clusters with only a handful of keywords in them, which helps when the map is noisy, and Shallow clusters hidden tells you when that is on. Clear filters and Clear all reset everything, which is worth knowing because filters persist between visits and No matches with your current filters turns up far more often than a genuinely empty universe. Clicking into a cluster lists its keywords, and Back to clusters returns you to the map.

The summary above the map counts Topic Clusters, Total Keywords and On-target keywords, and opening that last one lists the keywords Google already associates with your site, cluster by cluster. Anything badged Off-target there is drift: a keyword sitting in a cluster outside what your brand covers. That is a different thing from a cluster marked Probably skip, and worth keeping apart, since one is about a stray keyword and the other is about a whole area of demand.

Building or rebuilding the universe#

Generate Keyword Universe builds one, with Seed domain and Site URL (optional) setting what the run is built from and Country Code (optional) scoping the market. Re-run analysis rebuilds the whole thing, which is worth doing when your positioning or your market genuinely changes and not as routine maintenance. A rebuild alters the map your topics were drawn from, so treat it as a deliberate act rather than a refresh.

Promoting clusters into topics#

Promoting a cluster is how you change what gets tracked. Promoted clusters become topics and carry their keywords with them, and from that moment they determine which prompts get sent to the AI engines.

Choose for fit before volume when you do this. The largest cluster in a category is frequently not where its buyers come from, and a high-volume cluster sitting just adjacent to your business will quietly drag your tracked topics away from what you actually sell, with every downstream number following obediently behind. Read the keywords inside a cluster before you promote it, too, because cluster names are summaries and a name can look entirely right while the terms underneath describe something else. It is also worth deliberately including a few smaller, high-intent clusters, which rarely win on volume and quite often win on revenue.

Best practices#

• The country is right, checked before you read anything else on this screen.
• Clusters a colleague would recognise as a description of the business.
• Tracked topics matching what you sell, including the less obvious corners of it.
• Search volume treated as one input rather than as the ranking.

Next steps#

1. Check the country. It is the most consequential setting here and the easiest one to leave wrong.
2. Review the clusters that became your topics. They were chosen for you during setup, so this may well be the first time anyone has looked at them properly.
3. Promote anything genuinely missing. A service you sell with no cluster tracked against it is invisible to every other screen in the product.
4. Leave the Probably skip clusters visible instead of filtering them away, since knowing what adjacent demand exists costs you nothing.
5. Use Keyword Playground to test an idea before you commit it here, because playground runs never change what is tracked.

Limitations#

The universe describes search demand, which is a strong proxy for what people ask AI assistants without being the same thing. People phrase questions to an assistant far more conversationally than they type them into a search box.

Promoting clusters sets what gets tracked from that point forward. It does not retroactively change metrics already collected against a previous topic set, so a change here starts a fresh trend rather than rewriting the one you already have.