Languages

Brand/Non-Brand Splitter

Free, no signup. Classify GSC queries or Ads search terms with exact, substring, typo-tolerant, and common spacing variants. Optionally import reviewed labels to audit the split and discover candidate aliases. Your CSV stays in this browser.

Built by Patrick Stox

Search Console export limit: the UI can return only about 1,000 representative rows, and privacy-anonymized queries are omitted. Treat every analysis as a view of the exported sample, not the complete query set.

Or paste CSV
Feedback
Report a bug

Found something broken in Brand Nonbrand Splitter? Let us know what happened — this goes straight to a private triage queue, not a public list.

What will be sent 
 No tool inputs, uploads, pasted source, complete results, query parameters, or URL fragments are attached automatically. You can edit or remove the selected passage above. Browser and anti-abuse metadata is processed for spam prevention. 
  
 
  
 
 
  
 
 
   
 
     
 Local data 
Saved targets, named lists, and recent check summaries remain only in this browser.
 
   
 
     
  Runs entirely in your browser — nothing you paste is uploaded or stored. Anonymous run-level outcome counters may be used for aggregate research; URLs, domains, IPs, and identifiers are never included, and no statistic is released below 100 runs. 
 
 
How to use it
 
     
  1. Enter one brand term per line. Include the official name and genuinely distinct names people use; spacing and simple typos are handled automatically.
    2.  
  2. Choose a GSC or Ads CSV, or paste the table. The file must contain a recognized query or search-term column. To audit classifications, add a Reviewed class column with Brand or Non-brand.
    3.  
  3. Select Split rows. Check the row totals and the exact, contains, and fuzzy match counts.
    4.  
  4. Resolve any reviewed-label disagreements. Verify candidate aliases before adding them to the brand terms, then rerun and export.
    5.  
 
Example output Example data
 
With brand terms Acme Widgets and Acme, these fictional rows demonstrate the actual matching order.
 
 
QueryClassWhy
 
acme widgetsBrandExact normalized match
 
acme widgets returnsBrandContains a brand variant
 
acne widgets supportBrandFuzzy token match; review this row
 
best blue widgetsNon-brandNo supplied brand term matched
 
 
What you get
 
Brand rows and non-brand rows are mutually exclusive and keep the original header and columns. The method summary counts brand classifications by the first successful method: exact, then contains, then fuzzy. A review disagreement appears only when an imported human label contradicts the matcher. A candidate brand variant is repeated evidence to investigate, not a confirmed alias.
 
How it works
 
Brand terms and queries are lowercased and normalized for punctuation and spacing. For each term, the matcher also creates a compact version and adjacent-character-swap variants. It checks full-query equality first, substring containment second, and bounded edit distance against individual query tokens last. Rows are written back with CSV-safe escaping.
 
Features
 
  • Reads common GSC, Ads, and keyword-tool query headers.
  • Preserves every original column in two downloadable CSV files.
  • Handles spacing variants, adjacent letter swaps, and limited misspellings.
  • Compares optional independent reviewed labels with the automatic split.
  • Suggests repeated terms from reviewed false negatives as aliases requiring approval.
 
Limitations
 
  • Fuzzy matching can produce false positives for short or dictionary-word brands.
  • Without an imported reviewed-class column, the tool cannot independently know that its own classification is wrong.
  • Alias suggestions require at least two reviewed false negatives and still need human verification; frequent generic words can be poor aliases.
  • The tool does not recognize brands semantically, translate names, or infer parent/sub-brand relationships.
  • Aggregate “other search terms” rows are treated like ordinary input rows because this UI does not infer platform suppression semantics.
  • Files larger than 10 MB are rejected; split them before analysis.
 
Frequently asked questions
 
What counts as a branded search query?
A row is brandedSegmenting organic search performance into branded queries (your brand name, its variants and misspellings, and products uniquely tied to you) versus non-branded queries (everything else). Splitting the two stops brand demand from masking the SEO-driven growth in your reporting. when its query or search term exactly matches, contains, or closely misspells one of the brand terms you provide. Everything else is placed in the non-brand export.
Which CSV columns can the splitter read?
It recognizes a query column named Query, Top queries, Search term, Search terms, or Keyword. All other columns are preserved in the exported CSV.
How can I check the automatic classifications?
Add an optional Reviewed class, Expected class, Ground truth, or Manual class column containing Brand or Non-brand. The tool reports only rows where that independent label disagrees with its automatic classification.
Are suggested brand variants added automatically?
No. A candidate is shown only when a term repeats across at least two independently reviewed brand rows that the matcher missed. You must verify it before adding it to the brand terms.
How does typo-tolerant matching work?
The matcher creates normalized spacing and adjacent-character-swap variants, then allows a small edit distance against individual query tokens: one edit for short terms and up to two for longer terms.
Can fuzzy matching misclassify generic words as brand?
Yes, especially when a brand term is also a common word or is very short. Review the split and use the exact, contains, and fuzzy match counts as a quality-control prompt.
Does the splitter upload my search data?
No. Files up to 10 MB are read, classified, and exported locally in your browser.
 
 
Next stepWhich Pages Should I Work On? — look up the exact spec and expected values.

People also explore
ToolSEO Opportunity FinderToolAI Search Volume EstimatorToolGoogle Search Console Hidden Query EstimatorArticleSEO Due Diligence for M&AArticleGoogle Search Console (GSC)ArticleRedirect Error (GSC Status)
 
 
 
Feature requests for Brand Nonbrand Splitter
 

Upvote what you want most. New ideas can be submitted from the floating Feedback menu;
      requests appear here once approved, and the most-wanted rise to the top.

   
Loading…
  
  ➕ Request a feature   
New requests are reviewed before they appear here.
 
     
  
  
 
 
 
 
     
   
 
Where this tool helps
Common use cases
Create repeatable brand and non-brand reporting files
Classify supplied query rows with an explicit brand-term list and preserve every original column in mutually exclusive local CSV exports.
Audit typo-tolerant matches before reporting
Separate exact, contains, and fuzzy match counts and review fuzzy-only rows that may be genuine misspellings or generic false positives.
Compare automation with independent reviewed labels
Surface only rows where an optional reviewed Brand or Non-brand label disagrees with the automatic split while leaving unlabeled rows unevaluated.
Discover candidate brand variants cautiously
Find repeated terms across reviewed brand false negatives as alias candidates that still require ownership and intent verification.
Build a reusable GSC brand filter
Generate an escaped RE2-compatible regex from the verified brand list, test it separately, and export both classified datasets.
Watch the full workflow
Brand/Non-Brand Splitter walkthrough
Read the transcript

      
      

        
Brand/Non-Brand Splitter

        

        
This beginner walkthrough defines brand and non-brand queries, classifies eight fictional rows, explains exact, contains, and fuzzy matches, uses independent review labels to find mistakes, tests one important setting, and downloads both finished files.

      

      

        
Step 1

        

        
Brand Non-Brand Splitter divides search-query rows into two local C-S-V files. Brand rows match a name you supplied. Non-brand rows do not. It supports common Search Console, Google Ads, and keyword-file headers, preserves the original columns, and runs entirely in your browser without connecting to an account or uploading the file.

      

      

        
Step 2

        

        
A query is the phrase a person searched. A brand term is an official name or a verified name people use for the same brand. A fuzzy match allows a small spelling difference. A reviewed class is an independent human label—Brand or Non-brand—already stored in the input file. The reviewed label helps audit automation, but it can also be wrong and should not be treated as unquestionable truth.

      

      

        
Step 3

        

        
Use the splitter to build repeatable reporting files, check typo-only matches, compare automatic results with a reviewed sample, find possible unlisted brand names, or create a reusable Search Console filter. Human review matters most when a brand is short, looks like a common word, has several names, or is often misspelled.

      

      

        
Step 4

        

        
Enter one verified brand term per line. Leave typo-tolerant matching on for the first run, then choose a C-S-V or T-S-V file. The tool recognizes Query, Top queries, Search term, Search terms, or Keyword as the phrase column. Every other column is preserved, files over ten megabytes are rejected, and you can paste a small table instead.

      

      

        
Step 5

        

        
To audit the result, add a column named Reviewed class, Expected class, Ground truth, or Manual class. Use explicit Brand or Non-brand values. Blank or unclear values remain unevaluated. A disagreement means investigate the matcher and the human label. It does not automatically prove which side is correct.

      

      

        
Step 6

        

        
Enter Acme Widgets and Acme, then select the fictional eight-row C-S-V. It includes obvious brand and non-brand queries, two typo-only matches, two reviewed brand rows that the term list misses, two automatic brand rows reviewed as non-brand, and a repeated unlisted name. No customer, account, or real company data is used.

      

      

        
Step 7

        

        
Select Split rows. The browser normalizes each query, checks exact matching first, contains matching second, and fuzzy matching last. Every row goes into exactly one output. The tool then compares the result with any reviewed labels and creates findings only when the supplied evidence supports a specific review question.

      

      

        
Step 8

        

        
The first run produces five brand rows and three non-brand rows, accounting for all eight inputs. The hidden method summary records one exact, two contains, and two fuzzy brand matches. Exact, contains, and fuzzy are mutually exclusive because only the first successful method counts. A large fuzzy share is a prompt for closer review.

      

      

        
Step 9

        

        
The supplied terms generate this R-E-two-compatible Search Console pattern: Acme, whitespace plus Widgets, or Acme. You can copy it or send it to the dedicated regex tester. The pattern contains only the verified terms you entered. It does not automatically include fuzzy matches, review findings, or candidate aliases.

      

      

        
Step 10

        

        
The fuzzy finding lists acne support and acne pricing. Each is close enough to the short term Acme to match through typo tolerance. Acne support has a reviewed Brand label, while acne pricing is reviewed Non-brand. This is exactly why fuzzy means review needed—not confirmed typo, brand ownership, or branded intent.

      

      

        
Step 11

        

        
Widgetco support and Widgetco pricing are reviewed Brand, but neither matches the supplied Acme terms. Verify whether Widgetco is really an owned brand name and whether both reviewed labels are appropriate. If the evidence checks out, add a verified term and rerun. Do not change the list from this finding alone.

      

      

        
Step 12

        

        
Acme jobs and acne pricing are automatically Brand but reviewed Non-brand. Acme jobs is a literal contains match, while acne pricing is fuzzy. Depending on the reporting purpose, the next step might be narrowing intent, changing a term, disabling fuzzy matching, or correcting a human label. The tool does not choose among those options.

      

      

        
Step 13

        

        
Widgetco appears in two reviewed Brand rows the matcher missed, so it becomes a candidate variant. The suggestion requires repeated evidence and excludes supplied terms, stopwords, numbers, and one-off tokens. It is still only a candidate. Confirm ownership, real usage, and intent before adding it to the official term list.

      

      

        
Step 14

        

        
Turn off typo-tolerant matching and select Split rows again. The two fuzzy-only rows move to non-brand. The totals change from five brand and three non-brand to three brand and five non-brand. When one setting moves a quarter of this small sample, those fuzzy rows clearly deserve deliberate review before reporting.

      

      

        
Step 15

        

        
Without fuzzy matching, the method summary becomes one exact, two contains, and zero fuzzy. Acne support becomes a third reviewed Brand miss. Acne pricing stops contradicting its reviewed Non-brand label. Widgetco remains an alias candidate because its evidence does not depend on fuzzy matching. This comparison shows which findings are sensitive to the setting.

      

      

        
Step 16

        

        
After resolving the review queue, download Brand C-S-V and Non-brand C-S-V. This verified run creates a header plus five brand rows and a header plus three non-brand rows. Both files preserve every original column, including Reviewed class. Save the exact term list and fuzzy setting beside them so another reviewer can reproduce the split.

      

      

        
Step 17

        

        
Start with verified official names, then import a representative reviewed sample. Run the split and inspect the totals, method mix, fuzzy-only rows, and both directions of disagreement. Verify repeated alias candidates, rerun after deliberate changes, test the regex separately, and preserve the source file and review evidence with both exports.

      

      

        
Step 18

        

        
The main features are common query-header support, local C-S-V or T-S-V processing, original-column preservation, exact, contains, spacing, transposition, and bounded typo rules, optional reviewed-label auditing, repeated alias candidates that require approval, a Search Console regex handoff, and two mutually exclusive downloadable files.

      

      

        
Step 19

        

        
Short or dictionary-word brands can create fuzzy false positives. Without reviewed labels, the tool cannot independently know its own mistakes. Human labels and alias candidates also require verification. The matcher does not understand meaning, ownership, translations, parent or sub-brand relationships, account data, or platform-suppressed query groups. Files over ten megabytes must be split before analysis.

      

      

        
A useful split is reviewed, documented, and reproducible.

        

        
Start with verified official terms, investigate every uncertain or contradictory row, confirm ownership before adding an alias, and rerun the split. Save the source data, term list, fuzzy setting, review labels, and two exports together so future reporting can be audited.

      

    
 
         
 
Common issues & how to fix them
 
     
  •  Warning Fuzzy brand matches need review Fix: Review every fuzzy-only row and add or remove brand terms before using the split in reporting. 
  •  Warning Reviewed brand query was classified as non-brand Fix: Verify the imported review label, then add a confirmed brand term that matches the query and rerun the split. 
  •  Warning Reviewed non-brand query was classified as brand Fix: Verify the imported review label, then remove or narrow the brand term that matched this query and rerun the split. 
  •  Information Candidate brand variant needs review Fix: Confirm the repeated term is an owned brand, product, or alias before adding it to the brand-term list and rerunning the split. 
    •