You give RankDots a domain or a keyword, a language and a country. This step returns every keyword real people search for around that topic, checked for relevance and demand, before anything is grouped, scored or put in order.
It is the first of the five steps you watch while your plan is being built. Everything that follows - pages, topic clusters, the recommended order - is built from the list this step produces, so it is worth knowing where that list comes from.
Why this step exists
No single keyword source is complete. Google Keyword Planner rounds volumes into ranges and folds similar phrases together, so long-tail keywords disappear. Autocomplete shows what people type but not how often. Trend data shows direction, not size.
A plan built from one source inherits that source’s blind spots. A plan built from a raw export of many sources has the opposite problem: thousands of phrases that are off-topic, duplicated, or searched by nobody.
This step solves both. It collects keywords from several sources, then removes the ones that do not belong.
How it works
What you give us
You start a project with three things: a domain or a keyword, a language and a country. Advanced location options lets you narrow the country to a state, region or city.
If you enter a keyword, that keyword is the seed. If you enter a domain, RankDots uses what your site already covers to work out your niche.
Before discovery runs, RankDots shows you groups of keywords it found and asks you to select the matching categories to continue. Untick a group that is not your business. If none of them fit, switch on Not seeing what you’re looking for? and describe your topic in your own words. The groups you keep tell RankDots what your topic is — and what it is not.
Where the keywords come from
RankDots combines search data from Google with data from other search engines and from AI assistants. The sources named in the app are:
| Source | What it adds |
| Google Keyword Planner | The baseline: phrase variants and monthly volumes from Google’s own advertising data. |
| Google Trends | Whether interest in a phrase is rising, flat or falling. |
| Autocomplete | The phrases people actually type, including long questions that keyword tools fold away. |
| Related searches | Adjacent phrases Google shows at the bottom of a results page. |
| AI assistants | Questions phrased the way people ask an assistant, not a search box. |
Each source is strongest at one thing and weak at others.
Why a mix beats one source
When two sources agree on a keyword, that is confirmation it is real. When they disagree, that is coverage: one source found what the other missed. A keyword only Autocomplete surfaces is usually a long-tail phrase Keyword Planner folded into a bigger bucket. A keyword only AI sources surface is usually a question, phrased the way people ask an assistant rather than a search box.
The overlap makes the list reliable. The differences make it complete.
Every keyword is checked
A merged list from many sources is large and noisy. While discovery runs, the app tells you what is happening to it: every keyword is checked by algorithm and AI for relevance and real search demand.
There are two checks:
- Relevance. Is this keyword about your topic? A phrase that shares words with your seed but belongs to a different business is removed.
- Demand. Do real people search for it? Phrases with no measurable searches in your country and language are removed.
Keywords that fail either check never reach your project. How we check a keyword is real [LINK PENDING: how-we-verify-keywords] explains both checks in detail.
Adding your own keywords
If you already have a list, Import on the Topic clusters, Pages or Keywords screen accepts pasted keywords or a .csv file. Imported keywords join the list and go through the steps that follow.
What this means for you
The list will surprise you in both directions. Expect keywords you had not thought of - that is the point of using many sources. Expect a few you were sure of to be missing: they failed a check, or nobody searches for them in your country and language.
Country and language decide what is found. Every source is queried for the country and language you chose when you created the project. The same seed keyword in a different country produces a different list.
Search volume is an estimate. The volumes attached to these keywords are modelled from the sources above. They are not measured on your site. Positions, impressions and clicks from Google Search Console are measured; discovery data is not.
This step decides the size of your plan. The number next to Keywords in the sidebar is the output of this step. Every page and topic cluster you see later was built from it.