Cloud Sync & Catalogs: sync your catalogs without the warehouse detour

Personalizing a CRM campaign rarely comes down to one piece of data, and that data is almost never in one place: customer information in a data warehouse, an order status in a production database, a product catalog in an ERP or PIM. Cloud Sync and Catalogs bring that data to your campaigns, whatever its source. This quarter, both features extend to cases where the data never touches a warehouse: CSV imports over SFTP for Cloud Sync, atomic replacement for large catalogs in Catalogs.

Key takeaways

• A personalized CRM campaign needs data spread across multiple systems: warehouse, production database, ERP, PIM, flat files.

• Cloud Sync now pulls CSV files dropped on an SFTP server, alongside its existing warehouse and database connectors (Snowflake, BigQuery, Databricks, ClickHouse, MySQL, PostgreSQL).

• Catalogs separates what describes the customer (the profile) from what you can offer them (the catalog): products, content, stores, offers, destinations.

• A large catalog can now be replaced in one shot, with no in-between state where a campaign might pull incomplete data.

• The principle isn’t e-commerce specific. It applies to any reference data that describes what you’re offering, not who the customer is.

How do Cloud Sync and Catalogs bring together data that never lives in one place?

Sending the right message means combining what you know about a customer (behavior, preferences) with data that keeps changing: price, stock, offers, content, operational details. Those two families of data rarely sit in the same place.

Two Batch features close that gap. Cloud Sync automates syncing from wherever the data already lives, instead of relying on manual imports or custom-built pipelines. Catalogs stores reference data inside Batch, the kind that doesn’t describe a customer but what you can offer them: products, content, stores, offers, destinations. A campaign pulls the right information from that catalog the moment it personalizes a message.

What happens when data doesn’t live in a data warehouse?

Cloud Sync already connects Batch to warehouses and databases like Snowflake, BigQuery, Databricks, ClickHouse, MySQL, or PostgreSQL. A data team might compute a propensity score in their warehouse, and Cloud Sync syncs it automatically to the matching profile in Batch, ready to target or personalize a campaign.

But few data stacks are that tidy. Even when a company centralizes customer data in a warehouse, some reference data stays in an ERP, a PIM, or an internal tool, the product catalog being the usual suspect. And for a lot of companies, the simplest way to pull that data out regularly is still a CSV file.

How does Cloud Sync now pull CSV files over SFTP?

Cloud Sync now supports a new source: SFTP plus CSV. The setup is straightforward. A company generates CSV exports and drops them on an SFTP server. Cloud Sync picks them up automatically and uses them to populate a Catalog in Batch. The catalog stays in sync with whatever system counts as the source of truth internally, with no integration to build or maintain against the Catalogs API. Once the flow is set up, updates run on their own.

The goal isn’t to replace warehouse connections. It’s to stop making a warehouse a mandatory stop for every piece of data your CRM needs.

How does Catalogs avoid recommending a product that just went out of stock?

Take an abandoned cart reminder. Batch knows a customer browsed a pair of shoes and added them to their cart, that’s part of their profile. But since that visit, the price might have changed, a new image might be live, a size might have sold out, the product might be out of stock entirely.

None of that describes the customer, it describes the product. Copying it onto every profile would be both wasteful and hard to maintain. That’s what Catalogs is for: the profile simply holds the product ID the customer cares about, and the Catalog holds the current information on that product (name, price, image, description, availability). When it’s time to personalize the message, Batch uses that ID to pull the right details from the catalog.

The profile says what the person is interested in. The catalog describes what you can offer them. And since that catalog can now be fed automatically by Cloud Sync from CSV exports, it stays in sync with the company’s own product data.

How do you update a catalog of millions of products without interrupting live campaigns?

Replacing a product catalog item by item creates a window where part of the old catalog is already gone and the new version isn’t complete yet, which is a real problem if a campaign happens to fire during that exact window.

Catalogs now supports atomic replacement. Batch builds the new version of the catalog in the background first. Throughout that process, campaigns keep using the old, complete version. Once the new version is ready, Batch switches from one to the other in a single move.

Product Catalog Sync

To make importing large reference datasets through Cloud Sync easier, data can also be split across multiple CSV files. The same full-replacement logic is available through the Catalogs API for companies that prefer to run their own syncs directly.

Does Catalogs only apply to e-commerce use cases?

A product catalog is the most obvious example, but Catalogs isn’t limited to retail. A media company can maintain a catalog of articles or shows, a marketplace a catalog of listings, a travel company a catalog of destinations or trips, a retail chain the details of its stores.

The principle holds across all of them: keep what’s specific to each customer separate from what you’re offering them. Cloud Sync then brings the right data into Batch from whichever source fits: a warehouse or database for customer and behavioral data, a CSV export over SFTP for product or content reference data.

In short

Batch builds its platform on one simple principle: data should be able to reach a campaign regardless of where it lives, without forcing a data warehouse as the only way in. Cloud Sync and Catalogs are two concrete applications of that principle, built to shorten the distance between the data a company already has and the campaigns that can put it to use.