Canvs analyzes open-ended text responses from surveys and feedback files.

Before uploading, your file needs to be structured in a way Canvs can read correctly. This article walks you through everything you need to know - from choosing the right format to fixing the most common issues before they cause upload errors.

The fastest way to get started is to use one of our pre-formatted templates. Each template is set up to match a supported data structure so you can drop your data in and upload with confidence.

Columns as Questions template.xlsx

Rows as Questions Template.xlsx

Tabs as Questions template.xlsx

Canvs supports three file structures. The right choice depends on how your data was originally exported from your survey tool — not on personal preference. Pick the format that matches your file as-is; restructuring your data to fit a different format is rarely necessary.

This is the most common format. Each column represents one survey question, and each row represents one respondent's full set of answers. For example, Column A might contain responses to "What did you enjoy most?" and Column B might contain responses to "What could be improved?" If your survey tool exports one row per respondent, this is almost certainly your format.

In this format, each row represents a single response to a single open-ended question. Rather than spreading questions across columns, questions and responses are stacked vertically — multiple rows per respondent, all within the same set of columns. This format is common in exports from tools that use a "long" or "transposed" data structure.

This is a less common format. Each tab in the workbook represents one survey question, and all responses to that question live on that tab. The same structure repeats across every tab: one question, one tab, throughout the entire file.

If you're unsure which format applies to your file, start with Columns as Questions — it's the easiest to work with and the most straightforward to upload.

Once you've identified your structure, make sure the file itself is set up to upload cleanly.

Label your open-ended questions clearly. Use descriptive, human-readable column headers — "What did you enjoy most?" rather than "Q3_Enjoy." Clear labels help Canvs' AI Assistant understand the actual question being asked, which improves the quality of your analysis.

Set up your filters. Filters are the demographic or quantitative fields — like age, region, or ratings — that let you segment your results inside Canvs. Label filter columns clearly and use consistent, readable values. Write "Male" and "Female" rather than "1" and "2." You can include up to 50 filter columns in a single file, but you'll select which ones to activate during the upload process, so it's fine to include more than you plan to use immediately.

Add waves or timestamps (optional). Wave or timestamp columns are only necessary if you're tracking responses over time or comparing data across multiple collection periods. If your data comes from a single survey at a single point in time and you don't need time-based analysis, you can skip this entirely.

Before uploading, confirm the following:

Headers are clean. Remove special characters, numeric codes, and extra spaces. Every column header should be simple and readable at a glance.

Multi-select questions are consolidated. Some survey tools export multi-select questions as separate columns, one per answer option. Combine these into a single column before uploading so Canvs treats them as one filter variable.

Dates are consistently formatted. If your file includes timestamps, make sure every date follows the same format (e.g., MM/DD/YYYY) throughout the entire file. Mixed formats will cause upload errors.

Your verbatim columns aren't mostly empty. Canvs automatically skips individual empty rows, but if more than 20% of rows in a verbatim column are empty, the upload will fail.

Date format not recognized. Reformat your date column to use one of Canvs' accepted date formats. See Date Format Requirements for the full list.
​

Inconsistent filter values. Standardize your filter values before uploading. If a filter uses "Male" in some rows and "M" or "male" in others, Canvs will treat these as separate values. Pick one format and apply it consistently across the entire column.

Upload fails due to empty verbatim rows. If more than 20% of rows in an open-ended column are blank, the upload will fail. Review the column and either populate the missing responses or remove those rows before uploading.

File includes responses in multiple languages. If you're using Canvs Translations, your entire file will be translated and will consume translation credits for every row — including English responses. To avoid unnecessary credit usage, we recommend separating English responses into their own file and uploading non-English responses separately.