SEO and rankings
Five hundred keywords are not five hundred pages
It groups by meaning, not by shared words. Each group comes out with the page format it calls for and a priority, so the content plan writes itself.
- Grouping by meaning: running shoes and trainers for running end up together
- Adjustable similarity threshold: from broad groups to strict ones
- Recommended page format per group: guide, comparison or service page
The problem
- You have the list and stare at it weekly without knowing how to attack it.
- You wrote three articles on the same thing with different titles and none ranks.
- Grouping five hundred keywords by hand takes two days and comes out different every time.
One page per keyword is the most expensive way to compete with yourself.
How it works
Grouping by shared words joins things that are unalike and splits things that are identical. Grouping happens by meaning.
- 1
It compares by meaning
Two queries leading to the same intent go into the same group even with no word in common. That is the difference between a content plan and an alphabetically sorted list.
- 2
You adjust how much it groups
A loose threshold gives few large groups, useful for a small site. A strict one gives many fine groups, which is what a shop with a catalogue wants. There is no correct value: there is one correct for your case.
- 3
Each group calls for a format
A cluster of comparison queries is not solved by a guide, and a cluster of informational ones is not solved by a product page. The format comes out with the group; it is not decided afterwards.
What you get
The result is the list of pages to write, in order.
- The groups with their head keyword, the secondary ones and the volume they add up to.
- The recommended format and the priority of each group.
- The groups that already have a page of yours, so you do not write the same thing twice.
Measure before you pay anything
The scanner runs 10 real responses in about 60 seconds. No signup, no card. If the number it returns tells you nothing, do not buy.
Measure my brandWho it is for
- Content teams planning a quarter over a list they already have.
- Shops with a catalogue, where clustering decides the category structure.
- Anyone who has published a lot and suspects half of it overlaps.
It is not for you if...
It is not for you if you do not have the list yet. Clustering ten keywords adds nothing: first you have to discover what the market is made of.
An honest comparison
| By hand | With Citenza | |
|---|---|---|
| What it costs you | Two days dragging rows around a spreadsheet. | Five hundred keywords grouped in the time it takes to read this. |
| The method | Grouping by similar words, which is not the same as by meaning. | By meaning, with the threshold you decide. |
| Repeating it next month | With the next list, another two days and another criterion. | The same threshold gives the same result, always. |
Availability by plan
Available from Starter. The maximum list size grows with the plan, and cost depends on the number of keywords clustered, not on how many groups come out.
Honest questions
- What threshold should I use?
- Start at the middle value and look at two or three groups. If you see things together that you would write on separate pages, raise it. If you see twenty groups that would say almost the same, lower it. You tune it by looking at the result, not by choosing right first time.
- Can I move a keyword between groups by hand?
- Yes, and the change is kept if you cluster the same list again. You know your business better than a threshold, and there are cases where two similar queries serve different customers.
- Is a group always one page?
- Almost always, and that is the point. The exceptions are very large groups, where one page cannot answer everything well and it is better to split them into a main page and several linked supporting ones. The report flags groups above that size.
- Does it help organise AI prompts too?
- The clustering tool works on search keywords. For questions in a measurement set, what orders them is intent, and that is what the prompt intelligence tool does. Two similar problems with two different solutions.
From list to plan, in one afternoon
And while you plan the content, also measure whether the models already name you. It is free and takes a minute.