Overview
Provalytics can ingest custom marketing data even when it does not come from a native ad-platform connector. This is especially useful for channels such as:- SMS
- organic social
- billboard / OOH
- PR / earned media
The general rule
For most custom marketing sources, Provalytics typically needs:date- an exposure or delivery metric
channel namecampaign name(optional)creative name(optional)clicks(optional, when meaningful for the source)
Best practice: backfill historical data
When historical data is available, best practice is to backfill it. That matters because if the channel is going to appear in the ongoing dataset, it is better for the model to already have historical exposure to that source instead of encountering it only after the fact. In practical terms, backfilling helps:- train the model on the channel earlier
- improve continuity in the dataset
- reduce the chance that a newly added source appears too late to influence interpretation well
Best practice: keep the same cadence as current data
Custom marketing data should be delivered on the same cadence as the rest of the active dataset whenever possible. That keeps the dashboard current and helps the platform continue updating the relevant views as new activity comes in. In most cases, the best operating model is:- consistent recurring delivery
- stable file structure
- same update cadence as the rest of the connected data
Source-specific requirements
Use the pages below for the most common custom marketing sources:- Email Marketing Data
- SMS Marketing Data
- Organic Social Data
- Billboard and OOH Data
- PR and Earned Media Data
How data can be delivered
Once the data is structured, Provalytics can receive it through several supported delivery methods, including:- Recurring Email Delivery
- SharePoint Delivery
- Google Sheets Delivery
- S3 Bucket Delivery
- BigQuery Delivery
- Snowflake Delivery
- Manual File Uploads