Keyword Explorer

Summary#

Keyword Explorer takes one seed keyword and returns the surrounding keyword universe: related terms with real monthly search volume, keyword difficulty, search intent and cost per click, plus the questions people ask, grouped into sub-topic clusters. Its internal tool id is keyword_magic.

Purpose#

Most content plans start with a single phrase somebody typed into a document. That phrase is rarely the best target, and on its own it never justifies a page — let alone the hub of pages that actually wins a topic.

This tool exists to answer one question: given this seed, what should I actually build? It replaces a guessed keyword list with a measured one, so you can see which parts of a topic have demand, which parts are winnable, and where the questions cluster. The decision it supports is scope: one page or a hub, and which term goes on which page.

Overview#

You type a seed keyword. Metric Vault queries Google keyword data for terms related to that seed, from three angles at once: phrase suggestions built around the seed, semantically related terms, and broader ideas from the same topic space. Every returned keyword carries its real search volume, difficulty score, competition level, intent and CPC.

The result screen then does the grouping for you. It buckets the keywords by volume and difficulty, splits them by intent, derives sub-topic clusters from the words that recur across the list, and pulls out the question keywords separately, because those map cleanly onto headings and FAQ blocks.

Nothing on the screen is invented. If the provider returns no keywords for your seed, you get an honest empty result rather than a plausible-looking list.

Benefits#

• Turns a single idea into a measured, prioritized keyword set in one run.
• Shows demand and difficulty together, so you can spot the winnable volume.
• Separates question keywords, which are the fastest route to headings, FAQ blocks and featured snippets.
• Costs 1 credit, which puts it below the monthly premium-report threshold — see How credits work.
• Feeds the rest of the workspace: keywords can be saved to Keyword Vault and carried into a brief.

Use Cases#

Planning a content hub. You know the topic but not the page structure. Run the seed, read the sub-topic clusters, and give each cluster one page instead of writing one page per keyword.

Finding a realistic first target. A new site cannot win the head term. Filter to lower volumes and read the difficulty distribution to find the terms a small site can actually rank for.

Building an FAQ section. The Related Questions block gives you real question phrasings to answer, rather than questions you imagined.

Sizing a topic before committing. Before pitching a content programme, you want evidence there is demand. The average volume and total keyword count give you that in one screen.

Refreshing a plan quarterly. New keywords appearing around a seed are early demand signals; re-running the same seed surfaces them.

Requirements#

• A signed-in account. Anonymous visitors see Please sign in to run this tool.
• A paid plan. This tool costs 1 credit, and any tool costing more than zero is blocked on Free.
• Available credits and headroom in the hourly fair-use limit.
• No integration and no verified domain are needed — the tool works on any keyword.

Permissions#

Any member of a workspace on Starter or above can run this tool; there is no role restriction on running it. A Free-plan account is refused before anything is fetched, with This tool needs a paid plan. Free includes the 10 technical SEO tools; upgrade to Pro to unlock the rest. The Get recommendations panel below the result is gated separately at Pro and above — see Get Recommendations.

Cost#

1 credit per run. Because the cost is below 3, this tool is never blocked by your monthly premium-report quota; it is limited only by the hourly fair-use cap of 100 light-tool calls per hour. A result served from the shared data cache still charges the credit. Full model in How credits work.

Navigation Path#

Dashboard → Keyword & Content Research → Keyword Explorer

Inputs#

Step-by-Step Guide#

1. Open Keyword & Content Research in the left rail and choose Keyword Explorer.
2. Type your seed into the field marked e.g. running shoes. Press Try sample if you want to see the shape of a result first.
3. Optionally set MIN VOLUME and INTENT.
4. Click Generate Keywords · 1 report, or press Enter in the field.
5. Wait while the screen shows Analyzing… this can take a few seconds for live data.
6. Read the result from the KPI row down.
7. Use Save Cluster, Add to Tracker or Create Content Hub to carry the keywords forward, or Export CSV to take them out.

Reading the Results#

The header. The seed keyword sits next to three chips — Verified, Live DataForSEO and Keyword universe analysis — and a verdict badge on the right that summarizes the opportunity in the cluster. Green means the list is rich in winnable volume; amber means mixed; red means this seed is either thin or dominated by hard terms.

The four KPI tiles.

Top Sub-Topic Clusters. A cloud of the recurring themes derived from the real keyword list. This is your page plan: one page per cluster, not one page per keyword. If a cluster has only one or two keywords behind it, fold it into a section of a bigger page.

Intent Mix. A donut showing how the related keywords split across informational, commercial, transactional and navigational intent. Match your page format to the dominant slice. A cluster that is 80% informational will not convert on a product page, and a transactional cluster will not rank with a guide.

Volume Distribution. Counts of keywords in five bands (0–100, 100–1K, 1K–10K, 10K–100K, 100K+). A pyramid weighted towards the lower bands is normal and healthy — it is where the achievable traffic lives.

Difficulty Distribution. The same idea for difficulty. If the mass sits at the hard end, do not start here; pick the low-difficulty terms and earn the rest.

Keyword Universe. A ranked opportunity list, ordered by volume. This is the fastest read on the screen: the top rows are the biggest prizes, and each row carries its own difficulty so you can see immediately whether the prize is reachable.

Related Questions. Real question keywords lifted from the list. Each one is a candidate H2. Answer them in one place and you have both a better page and a snippet target.

Top Keywords. The full table — volume, difficulty, intent and CPC per keyword. Difficulty is color-coded: green below 30, amber below 60, red above. CPC is your commercial-intent proxy: a term advertisers pay several dollars for is a term buyers search. If you set a MIN VOLUME or INTENT filter, rows outside it are hidden here.

Semantic Neighbors and Broader Cluster Discovery widen the net — the first with closely related terms, the second with keywords from the same topic space that do not contain your seed. Use them when the main list feels too narrow.

What to do with all of it. Pick one cluster. Take its highest-volume term with a difficulty you can live with as the page target, take the rest of the cluster as sections, take the questions as an FAQ, and send the target to a brief.

Filtering, and what the count means#

Every control above the keyword table filters the whole run, on the server. Narrowing to difficulty under 20 tells you how many of the thousand qualify, not how many of the hundred rows currently loaded do, and the number beside the table is that answer.

Export all matches writes every row the current filter matches, not the rows on screen. Filter to what you want first, then export, and the file and the count agree.

How keywords are grouped#

By the domains that actually rank for them. Two keywords join the same topic when at least three of the same sites appear in both results and the overlap is a quarter of the combined set or better.

This is deliberately not name similarity. Grouping by shared words puts apple watch with apple pie and separates cheap running shoes from running shoes cheap, and the result reads plausibly enough that nobody checks it. Grouping by SERP puts hoka shoes with brooks shoes, which share nothing as strings and everything as a search result.

The trade is honest coverage. The grouping only speaks for keywords the provider holds SERP data on, so a run typically groups somewhere over half of what it gathers and leaves the rest listed and ungrouped. An ungrouped keyword is not a worse keyword; it is one we will not guess about.

Examples#

Screenshots#

Tips#

• Seeds of two or three words cluster better than single words. running shoes beats shoes.
• Set the filters before you run. They are sent with the request as well as applied to the table.
• Use Export CSV when you want the list in a spreadsheet; use Save Cluster when you want it inside Metric Vault.
• Run a second seed from a promising cluster. The clusters of a cluster are usually where the uncontested terms are.

Best Practices#

• Build one page per cluster and link them together. A hub outranks scattered posts.
• Start with the lowest-difficulty terms that still have volume. Early wins are what make the hard terms reachable later.
• Check intent before choosing a format, not after writing.
• Re-run your core seeds on a schedule. New entries are demand appearing.
• Send the chosen target to Keyword Overview before committing — a single-keyword deep dive will tell you what the SERP actually looks like.

Common Mistakes#

• Treating every keyword as a page. The list is raw material for a handful of pages, not a publishing queue.
• Chasing the highest volume row. Volume without a reachable difficulty is a page that never ranks.
• Ignoring the intent split and writing a guide for a transactional cluster.
• Assuming the list is exhaustive. It is a scored sample of the universe, not every keyword that exists.

Limitations#

• A run gathers up to 1,000 keywords, merged and deduplicated across the three discovery endpoints. That is a depth, not a census: a broad seed like running shoes has about a million keywords behind it, and the header says both numbers so the two are never confused.
• Grouping covers the part of the run there is SERP evidence for. Measured on running shoes, 556 of 1,000 keywords had SERP data and were grouped into 26 topics; the other 444 are listed ungrouped rather than being placed by name similarity.
• Question keywords are limited to eight.
• Keyword data is pulled from Google's United States, English-language index. The country picker is recorded with the run but does not change which market the keyword metrics come from.
• Search volume is a monthly average, not a live counter.
• There is no historical trend on this screen. For a keyword's trend over time, use Keyword Overview.

Troubleshooting#

FAQs#

How is difficulty calculated? It is Google-derived keyword difficulty supplied by our data provider, on a 0–100 scale. Under 30 is a realistic quick win, 30–60 is winnable with better content, and above 60 usually needs backlinks and months of effort.

Why does the keyword count say thousands but I only see a few dozen rows? The count is the provider's total for that seed. The screen shows the strongest slice of it so the result stays readable and fast. Export or re-seed from a cluster to go deeper.

Does changing the country picker change the keyword data? No. The picker is recorded with the run, but the keyword metrics come from the United States, English-language index.

Does a cached result still cost a credit? Yes. The credit reflects the run, not the provider call. Reopening a saved run from the Library is free — see Saved Work.

Is this the same as Topic Map? No. Keyword Explorer starts from a keyword and returns keywords. Topic Map starts from a topic and returns clusters, trends, publishers and questions. Use Magic for the list, Topic Map for the plan.

Can I get recommendations on the result? Yes, from the panel under the result, on Pro and above. Get Recommendations explains the gate and its separate allowance.