Ipsos Logo
  TACTIC - Classification: Brand Matching
Developed by Ipsos ACE

Step by Step Instructions

Labeling: Brand Matching attempts to identify which entity names from a user-provided list appear in each verbatim. For example, if the user supplied a list of tech companies, [Facebook, Apple, Amazon, Netflix, Google], and a list of survey open ends, TACTIC will attempt to identify which of the tech companies are mentioned in each open end. In the open end "I love goggle and nElfix", TACTIC would identify "Google" and "Netflix" as matches.

  1. Upload a CSV, XLSX, or SAV file containing a column of text using the "Choose File" button.
    1. If you uploaded a multi-sheet XLSX file, you will be given the option to select the sheet name which holds the text data.
    2. You will be given the option to select the column name for the text column.
    3. Hit the "Upload Text Data" button.
  2. After uploading the file, you can view the data in the DATA tab at the bottom of the screen.
  3. In the text box next to "Short description of brand type", provide a brief description of the category of labels being supplied. This description should be only a few words long at most and in the singular form (not plural). For example, if the brand list the user is supplying is a list of automobile models, the description could be "Car Model". Keep in mind that this description will be attached to every label in the list, as well as all entities in the open end when passed to the GenAI model to help it disambiguate between different types of entities. For instance, "Car Model: Odyssey" makes it clear that "Odyssey" is a car model and not the title of Homer's epic poem.
  4. Upload a CSV or XLSX file containing a list of the brands or entities you wish to target.
    • If you uploaded a multi-sheet XLSX file, you will be given the option to select the sheet name which holds the text data.
    • You will be given the option to select the column name for the text column.
    • You will be given the option to select the column name for the ID column, if one exists.
    • Hit the "Upload Entity List" button.
  5. After uploading the file, you can view the brands list in the Brand Classification Stats tab at the bottom of the screen.
  6. Before hitting the RUN button, evaluate the estimated cost and number of texts. In most cases, the actual cost will be far lower than the maximum possible estimate. Hit RUN.
  7. After a few moments, general statistics on the process will be displayed in the Brand Classification Stats tab at the bottom. To view results for each row, refer to the DATA tab.
    • As a note, any brands/entities in an open end which TACTIC is unable to match with a brand in the user-supplied list will be assigned the label "Other".
  8. If you are pleased with the results, you can download a XLSX of the results using the "Export Report" button in the "Data Control" section.

User Feedback & Bug Report Form


TACTIC supports CSV, XLSX, and SAV files of up to 99MB.
Files >5MB will take longer to load and process, please have patience.


Brand Classification Performance
Est. max cost/run: $0.00 Texts: 0 Entities: 0 Document Coverage: 0%
0%

Brand Classification Data


ID Text Input Order Brand Matches

.