Skip to main content
Custom data lets you upload your own CSV or JSON files and use them alongside data from connected data sources. It’s useful for:
  • Offline sales, lead lists, or CRM exports
  • Revenue forecasts and targets
  • Any KPI that lives outside the data sources we support natively

​
Where it lives

Custom data isn’t a standalone page in the sidebar. You reach it from the Data Query widget while editing a project:
1

Add a Data Query widget

In the project editor, add (or select) a Data Query widget to open the query builder in the left panel.
2

Pick Custom Data as the source

In the widget’s data-source dropdown, choose Custom Data. The picker also lists your connected data sources, Static Value, Calculated Metrics, and Goals.
3

Choose or upload a dataset

Pick an existing dataset from the Dataset dropdown, or use the Upload New File drop-zone below the OR divider. You can drag a file in or click to open the file picker.
4

Name and save

After Oviond parses the file, enter a Dataset Name and click Save Dataset. Oviond parses the headers into columns and stores the rows.
You can also start from a blank grid with Start from Blank, or use the Manage button next to the Dataset label to open the Manage Custom Data slide-over, where you can edit, rename, duplicate, and delete datasets across your account.

​
Supported files

Uploads accept .csv and .json files, up to 5 MB each.

​
How a dataset is used

Once you select a dataset, its custom_data_id is stored on the widget query, and you map the dataset’s columns to the widget’s metrics and dimensions. The widget then queries the dataset alongside your connected data sources.
Custom data is account-level — any client’s project can reference any dataset you’ve uploaded. A dataset can hold at most 10,000 rows.

​
Custom data vs a data source

  • Data sources (Google Ads, Facebook, etc.) — live data pulled from a connected marketing platform.
  • Custom data — static or periodically refreshed data you upload yourself.

​
API

POST /v1/custom-data creates a dataset. Send name, file_name, raw_data (an array of row objects, max 10,000), columns, and file_size. The response returns the new dataset id.