Bulk Keyword Clustering — up to 50,000 keywords on Agency

Handling messy exports

Every rank tracker exports differently: separators change, encodings change, headers are translated, numbers arrive with thousand separators or as ranges. The parser detects the structure of your file, including UTF-16 exports and files with byte-order marks, and separates numbers that are part of a keyword from numbers that are search volume.

That distinction matters more than it sounds. "iphone 15" contains a number that is part of the query; "iphone case 4400" usually contains a monthly volume glued to the term. Getting it wrong corrupts every downstream priority.

Performance at scale

Above a few hundred keywords, clustering moves into a background worker, so the page stays interactive and progress is visible while grouping runs. Nothing is uploaded for standard methods — the computation happens on your machine, which also means large proprietary keyword lists never leave it.

• Live progress with cluster counts as they form
• Interface stays responsive during long runs
• Standard clustering runs locally in the browser
• Projects can be saved and reopened later

Prioritising the output

A 50,000-keyword run can produce thousands of clusters, which is only useful if it is ordered. Clusters are ranked by total search demand across their members, so the first screen of the export is the first quarter of your content plan.

From there, the usual filters apply: strip branded terms, separate transactional from informational groups, and mark clusters that already have a matching URL on the site.

Exporting for the rest of the team

Different people need different shapes of the same data. Writers want one sheet per cluster; strategists want a flat table they can pivot; developers want JSON. All four formats come out of the same run, and the multi-sheet spreadsheet export puts each group on its own tab with volumes and totals intact.