Yes, you can run a Salesforce Data Cloud sandbox. Data 360 (formerly Data Cloud) works in every standard Salesforce sandbox type once production has it, though what lands in the sandbox is your setup without the records.
If you are new to the platform, start with what Salesforce Data Cloud is. After that you will need test records, an identity resolution run and a DevOps data kit to move changes to production, and each one gets its own section below.
Can you use Data 360 in a Salesforce sandbox?
Salesforce made it official in January 2025, when its Developers blog announced that Data Cloud in sandboxes is generally available. According to the post, "All Data Cloud features are fully functional in sandboxes."
On the Salesforce Platform sandboxes page, Data Cloud sandboxes sit in the same list as Full, Partial Copy and Developer ones. Salesforce pitches them for testing AI tools and Agentforce actions powered by Data Cloud data before new Agentforce or Data Cloud features reach production.
Trailhead's setup unit puts Developer and Developer Pro sandboxes on first setup and basic tests, with Partial Copy for a slice of the data model. Performance tests and user acceptance testing (UAT) belong in a Full sandbox.
Your core sandbox environments still differ on things like storage limits, and our guide to Salesforce sandbox environment types has the full rundown. For Data 360, though, Trailhead says the sandboxes "behave exactly the same once it is provisioned."
Why is Data Cloud unavailable in my sandbox?
Usually it's because production doesn't have Data 360 yet. Salesforce Help sets out the requirement: "To use Data 360 in a sandbox, provision Data Cloud licenses in production and create the Data 360 tenant."
Older sandboxes trip people up too. The GA announcement says a sandbox created before August 1, 2024 needs a refresh before Data Cloud can provision, while newer ones can use the License Match tool.
If you're still stuck after that, the GA post's advice is to go to your account team: "Speak with your Salesforce AE to provision Data Cloud in sandboxes for your org."
Practicing in a Trailhead Playground won't work either. Trailhead notes that "Sandboxes aren't available in the Trailhead Playground," so a Playground can't stand in for a Data 360 sandbox.
How do you set up a Data 360 sandbox?
Setup breaks into four stages. Look for Data Cloud Setup in the menu, since Salesforce says you may still see references to Data Cloud in the app during the rename.
Create or refresh the sandbox
In production Setup, enter Sandboxes in Quick Find, select New Sandbox, enter the details and choose a type. If your sandbox was created before August 1, 2024, refresh it instead, per the GA post.
Turn on Data 360 in the sandbox
Log in to the sandbox, open Setup, select Data Cloud Setup and click Get Started. Trailhead says enablement usually takes 15 to 30 minutes, then another 30 to 60 minutes for basic settings to arrive from production.
Reconnect and clean up
Your connections come across, but their authorization data doesn't, so the GA post says they'll sit there marked Inactive. Expect to re-enter credentials for the CRM connector and for any external storage you use, such as Amazon S3 or Microsoft Azure.
If you run Data Cloud One (DC1) across several orgs, Trailhead wants you to confirm the sandbox is connected as your companion org. Then use the metadata cleanup Trailhead describes to clear out orphaned data streams and stale mappings you inherited from production.
Help's sandbox considerations warns that "Activations copied to a sandbox retain their production segment and activation IDs," so a Marketing Cloud Engagement target shares production's data extension key. To avoid clearing or overwriting production data, Help says to "keep sandbox activation targets inactive or ensure they use a completely isolated Enterprise ID (EID) or destination during the authentication process."
Create a data space and start small
Trailhead's advice here is modest: make a local data space and pull in one simple source to prove the connection works. Once that first stream is flowing, bring in the bigger ones.
What gets copied into a Data 360 sandbox?
A Data 360 sandbox copies your configuration but not your data. The table below shows what turns up, what turns up empty and what you'll have to rebuild.
| Item | In a new or refreshed Data 360 sandbox | What you do |
|---|---|---|
| Data streams | Copied, no data | Reauthorize connectors, then ingest test data |
| Data model objects and mappings | Copied, no data | Check mappings against test sources and remove stale ones |
| Calculated insights | Copied, no results | Rerun after test data is ingested |
| Segments | Copied, no members | Rerun after identity resolution |
| Identity resolution rulesets | Copied | Run manually to build unified profiles |
| Unified profiles | Not copied | Built from scratch in the sandbox |
| CRM and external connectors | Copied as inactive | Reauthorize in Data Cloud Setup |
| Data Cloud One companion connection | Breaks on refresh | Re-establish after the refresh |
Trailhead says data streams, data model objects, calculated insights and segments come over "but they don't contain any data." Unified profiles aren't copied and "must be built from scratch," and you won't see production's calculated insight results either.
The empty profiles affect segments most. Salesforce's Architects guide calls segmentation the process of turning unified customer profiles into audiences, and a fresh sandbox has no profiles until you run identity resolution.
Where should Data 360 test data come from?
A good chunk of Trailhead's refresh unit is about one setup: a CRM sandbox acting as the source for your data streams. Whatever records live in that sandbox end up as your Data 360 test data.
A Developer sandbox copies only your configuration, per Salesforce Help, so identity resolution has nothing to match. Partial Copy and Full sandboxes have the opposite issue, since they bring a sample or the whole of your production data, real people included.
Our post on sandbox security covers what can go wrong with real records in a sandbox. Even Salesforce's GA post tells you to use Data Mask to "continue masking your CRM data prior to ingestion into Data Cloud."
What tends to work is a realistic subset, with relationships kept intact, that gets masked before Data 360 ever reads it. If you haven't done that before, start with how to seed a Salesforce sandbox and then Salesforce data masking.
A test copy still counts: GDPR Article 4(2) defines processing to include the storage and use of personal data. Make sure your GDPR compliance controls reach the sandbox too.
This is the part Flosum Sandbox Management handles for the source environment, provisioning Developer sandboxes with "PII masked, production data seeded, access locked down." That gives the CRM sandbox realistic records with PII already masked.
What happens when you refresh a Data 360 sandbox?
A refresh pulls the latest metadata down from production and wipes out whatever you'd built in the sandbox. So before you refresh, the refresh unit has you save the sandbox metadata in a DevOps data kit.
The worse problem is refreshing the CRM sandbox behind Data 360, because its OrgId changes and its authentication permission sets disappear. Trailhead warns that "This permanently breaks the connection to Data 360."
Trailhead lays out the order to do it in:
- Save the CRM data stream configurations in a DevOps data kit.
- Delete every data stream that uses that sandbox as a source, along with things like mappings that depend on them.
- Go to Data Cloud Setup, open the Salesforce CRM Connector and disconnect the sandbox.
- Now run the refresh.
- Afterward, recreate the permission set, connect the sandbox again and install your data kit.
- Reconnect external systems such as Marketing Cloud Engagement, and run identity resolution yourself, since Trailhead says the first run has to be started manually.
Never refreshing has its own cost. Trailhead describes a segment that found 10,000 unified profiles in an out-of-date sandbox but only 8,500 in production, where the identity rules had been tightened.
In that example, 1,500 target customers would have dropped out of a loyalty program launch. If the rest of your sandboxes have the same problem, our post on keeping Salesforce sandboxes in sync walks through it.
Does a Data 360 sandbox consume credits?
It does, and Trailhead says so outright: "Using a Data 360 sandbox impacts your production org credits."
Look at Salesforce's Flex Credits rate card and you'll find a separate sandbox column for Data 360 usage. A footnote spells it out, "Data 360 in Sandbox environments does not utilize tiered multipliers," and the fine print adds that the sandbox multiplier covers scratch orgs as well.
Salesforce also warns that multipliers "may be updated from time to time," so check the newest card before budgeting. Keep test loads small, and agree on a spending cap before UAT starts in a Full sandbox.
How do you move Data 360 changes from a sandbox to production?
The vehicle is a DevOps data kit, which Trailhead says exists "to facilitate the controlled, repeatable promotion of configuration changes." One kit can hold data streams, data model objects, calculated insights, identity resolution rules and segments.
To ship the kit, Trailhead's deployment unit gives you change sets, the Salesforce CLI or Metadata API, or an outside DevOps tool. Change sets come with a warning ("not recommended for multi-component data kits"), which fits the usual change set limitations.
Back in January 2025, the GA post also named DevOps Center, since renamed Lifecycle Center. If you'd rather script it, Salesforce's developer guide shows how to deploy a data kit from the command line.
A few things in that unit are easy to miss:
- Data spaces have to line up between sandbox and production, or mappings break.
- Anything another component relies on has to go first, so a calculated insight ships before the segment built on it.
- Data streams show up as drafts, and CRM streams only deploy fully on a second pass, after you activate the connector.
- Run identity resolution before you go looking for your known test profiles.
We're planning a full walkthrough of deploying Data 360 changes to production. Until it's up, our guide to Salesforce deployment has the general release steps.
FAQs
Can you use Salesforce Data Cloud in a sandbox?
Yes, once Data 360 (the newer name for Data Cloud) is provisioned in your production org. From there you can switch it on in any standard sandbox, Developer or Full, and Trailhead says it behaves exactly the same in each.
Why is Data Cloud unavailable in my sandbox?
Usually production doesn't have Data 360 yet, or the sandbox is older than August 1, 2024 and needs a refresh first, per the Salesforce Developers blog. If Data Cloud Setup still won't appear, talk to your Salesforce account executive about sandbox provisioning.
What gets copied into a Data 360 sandbox?
Just the setup: data streams, data model objects, calculated insights and segments all show up empty, and unified profiles don't come across at all. You'll need to reconnect sources, load some test records and run identity resolution before segments return anything.
How do you move Data 360 changes from a sandbox to production?
Bundle the changes into a DevOps data kit, then deploy it with the Salesforce CLI, the Metadata API or a DevOps tool once data spaces and dependency order check out. Afterward, switch on the data streams, which land as drafts, and run identity resolution in production.
Does a Data 360 sandbox use credits?
Yes, sandbox work comes out of your production org's credits, although the Flex Credits rate card gives Data 360 its own sandbox rate without volume tiers. Keep test loads modest and look at the latest card before you run anything big in a Full sandbox.
Book a demo and we'll show you how to get realistic, masked CRM data into every sandbox before Data 360 ingests it.
Thank you for subscribing



