Documentation

Naming Convention is a new product from Improvado that helps teams automatically audit, clean, and govern naming logic across data sources.

It enables consistent formatting across campaigns by splitting names into structured parts, validating allowed values, and offering AI-assisted corrections.

At its core, the system is built on two simple primitives:

1. Dimension Table
   This is your actual campaign data, where each campaign name is split into structured parts (like geo, platform, objective, etc.). It includes:
2. ~campaign_id
3. ~original_campaign_name (never changes)
4. ~new_campaign_name (updated after cleanup)
5. ~Parsed parts based on your delimiter
6. Allowed Values Table
   This is a dictionary of valid values for each part. It defines what’s allowed, and everything outside of it is flagged.

By comparing your campaign data against the dictionary, the system can instantly detect naming issues. You can then fix them manually or use AI suggestions.

How It Works

1. Select a Data Source and Dimension
   Start by selecting:
2. ~A data source (e.g. Facebook Ads)
3. ~A dimension to clean (e.g. campaign)
4. ~Optional filters (e.g. specific accounts)
5. ~Delimiter (optional)
6. Run Sync Data
   When triggered, Improvado automatically creates two tables in the database:
7. ~Dimension Table — a cleaned and structured view of your campaign names.
8. ~Allowed Values Table — a dictionary of valid parts for each segment.

Dimension Table

This contains parsed campaign name parts. The delimiter is auto-detected or user-defined. The table includes:

Allowed Values Table

This is a dictionary of valid values for each part. For example:

This table is tightly linked to the Dimension Table:

• Each part in the Dimension Table must match a value in Allowed Values.
• If not, it’s flagged as an error.

Error Highlighting & Workflow

As soon as a value in the Dimension Table doesn’t match its dictionary in the Allowed Values:

• That row is automatically highlighted in red.
• You get instant visibility into which campaigns have broken naming logic.

Typical cleanup flow:

1. Start with Allowed Values — remove outdated or invalid entries.
2. Errors appear in the Dimension Table.
3. Fix values via dropdown (auto-synced to Allowed Values).
4. Save and commit changes.

Fix with AI

Once errors are identified, click Fix with AI.

• The Agent reviews inconsistent parts.
• It suggests best-guess corrections based on your dictionary and historical structure.
• You can accept or reject suggestions — nothing is changed automatically.

Database Sync & Transformation Usage

Both tables are live in your warehouse and update automatically on each save. They can be joined to other tables using the Dimension ID and used in transformation logic, reporting, and dashboards.

Data Activation (Beta)

Once you’ve cleaned and validated your naming convention, you can push the cleaned names back into source platforms like Facebook or Google Ads.

This is done via Dimension ID, so there’s no guesswork or duplication. Your naming becomes consistent not only in the database but also directly in your ad accounts.

Conclusion

Naming Convention is currently in private beta.

It’s designed for companies that want to finally take control of naming at scale, not just for QA, but as part of a broader data governance system.

If your team struggles with inconsistent naming, misaligned dashboards, or painful campaign cleanup workflows, this is the system we built for you. It gives you live validation, AI assistance, clean joins, and even pushback into ad platforms.

To get access, reach out now — we’re rolling it out gradually and onboarding partners one-on-one.