
Syncing CRM data for engagement in Braze works best when you choose a short list of fields that change a message or an audience. Match each record to the right identifier, and send only values that changed. Three decisions shape the result: which fields earn a place, how they arrive, and how often they sync.
Key Takeaways
- Sync a CRM field only if it changes a message, an audience or a trigger.
- Cloud Data Ingestion identifies users by external ID, user alias, Braze ID, email address or phone number, with one identifier type per row.
- Recurring syncs run from every 5 minutes to once per month, and the Trigger a sync endpoint starts one on demand.
- Braze tracks changes by the timestamp column, so write only new or changed values.
- Views built on the current time, shared timestamps and two updates to one attribute are the failure modes to design out.
- Connected sources let Braze query your warehouse to build segments without copying the data.
Which CRM data for engagement in Braze is worth syncing?
A CRM holds far more than a campaign can use. The useful test is whether a field changes what someone receives or who receives it.
We put three questions to each field, in order:
- Would a message read differently because of it?
- Would a segment or a Canvas entry condition use it?
- Does the value change often enough to be worth tracking?
A field that fails all three can stay in the CRM.
What does a CRM record need before Braze can use it?
Cloud Data Ingestion identifies users by external ID, user alias, Braze ID, email address or phone number. Each row in your source table should carry a value for only one identifier type at a time. The external ID should match the external_id value already used in Braze, because it identifies the profile to create or update.
New users cannot be created through a Braze ID, so new profiles need an external user ID or a user alias. If several profiles share an email address, the profile updated latest is prioritised for updates.
The CRM contact ID may not be the ID your app already sends to Braze. Decide which one is the external ID before the first sync, and write the mapping down.
If identifier mapping across systems is the sticking point, CustomerIK's customer data management service covers identity resolution and the pipelines that move customer data between systems.
How should CRM data reach Braze?
The pages we reviewed describe warehouse and file storage sources for Cloud Data Ingestion. They don't describe a direct CRM connection, so check current Braze documentation before assuming one.
Our default plan is for the CRM to land in your warehouse first, with Braze reading from there. The table sets out three patterns the documentation describes.
| Pattern | What happens | Cautions from the documentation |
|---|---|---|
| Scheduled sync | Braze reads new rows from a warehouse table on a schedule you set | Recurring syncs run from every 5 minutes to once per month. The default shortest interval in the dashboard is 15 minutes |
| On-demand sync | You call the Trigger a sync endpoint when your load or transformation job finishes | It doesn't affect the recurring schedule. Only one sync runs per integration at a time |
| Connected source | Braze queries your warehouse to build a segment without copying the data | You pay the warehouse query costs. No data points are logged. The maximum runtime is 60 minutes |
How often should CRM data sync?
Match the sync to the speed of the decision. A renewal date used in a monthly campaign doesn't need a five-minute sync. A lifecycle stage that starts a Canvas might.
Braze documents an on-demand option for the moment your warehouse job finishes. Calling the Trigger a sync endpoint at the end of that job starts a sync when the data is ready. Only one sync can run per integration at a time, so keep scheduled intervals longer than a normal run.
How do you stop a sync from resending unchanged data?
Braze tracks changes strictly by the timestamp column. It does this whether or not the row matches what's already on the profile.
Braze recommends syncing only new or updated data, which avoids unnecessary data point usage. It updates or adds fields, so you needn't send the full profile.
Treat the source table as a change log. Append a row when a value changes, with only the changed attributes in it. Cloud Data Ingestion counts toward the same rate limit as the Braze API, so a noisy sync uses capacity your other integrations need.
Data point billing for Cloud Data Ingestion is equivalent to billing for updates through the /users/track endpoint. Braze no longer bills new customers based on data points, so check your contract before treating this as a cost.
What goes wrong with timestamps and updates?
Four failure modes come from the timestamp column or from two updates meeting in one run. The documentation describes each one, and each has a plain fix.
| Failure mode | What happens | Fix |
|---|---|---|
| View built on the current time | The timestamp is the time the query runs, so all data syncs every run | Store a real change time in the source |
| Shared timestamps | New rows at the last synced timestamp trigger an inclusive boundary, and rows already processed sync again | Use unique, increasing timestamps and write rows in one transaction |
| Local time | Daylight saving shifts cause sync issues | Store the timestamp in Coordinated Universal Time |
| Two updates to one attribute | A race condition can let one update overwrite another | Keep one row per user holding the latest state |
We rank the first two as likeliest to reach production unnoticed, because both look like a working sync. Compare the row counts of the first few runs with what changed in the CRM.
How would a hypothetical retailer decide what to sync?
Take a hypothetical retailer whose CRM holds loyalty tier, lifecycle stage, renewal date, account owner, open-text service notes and an internal credit rating. The loyalty tier passes all three questions. It changes a message and defines an audience.
The lifecycle stage passes too, and the retailer syncs it as soon as its warehouse job finishes. The renewal date changes rarely, so a scheduled sync is enough. The account owner doesn't change a customer message, so it stays in the CRM.
The service notes and credit rating stay out as well. Open text can hold personal data, and whether any field may be sent to Braze depends on consent and contract terms, which this post doesn't settle.
For a one-off question, such as how many loyalty members have a renewal approaching, the retailer uses a connected source. That builds the segment from the warehouse without copying the data into Braze.
How does this connect to the rest of the programme?
CRM data is only useful once it reaches an audience or a message. Our guide to integrating CRM data into Braze marketing automation covers the automation side, and segmenting your audience for better engagement in Braze shows how synced fields become segments.
For message content, using Braze Connected Content for personalised campaigns covers pulling data at send time. Optimizing the customer journey with Braze data and personalization strategy from planning to execution in Braze cover what to do with the data once it's there. If you plan to use decisioning, is your data ready for BrazeAI Decisioning Studio sets out the requirements.
For identity and CRM work in practice, a case study on unifying fragmented guest profiles for a hospitality brand covers consolidating booking, tier, language and consent data into one profile.
What should you check before the first sync?
Work through these steps in order, in a sandbox workspace first.
- List the CRM fields that passed the three questions, and drop the rest.
- Confirm which identifier is the external ID, and test it on a small batch of known users.
- Consider one source table or view per data type, which the documentation suggests for clear separation, since each sync updates one data type at a time.
- Store timestamps in Coordinated Universal Time, with unique values for each batch.
- For syncs that use a payload column, keep each row to a single user ID and up to 250 attributes, with a payload of up to 1 MB. A single run can sync up to 500 million rows by default.
- Compare the rows synced in the first runs with the rows that actually changed.
We recommend starting with the smallest set of fields that changes a message or an audience, proving the identifier match on a small batch, and widening only when a campaign needs more.
Frequently Asked Questions
1. Does Braze connect directly to a CRM?
The pages we reviewed describe warehouse and file storage sources, not a direct CRM connection. Releases change, so check current Braze documentation before you plan around either route.
2. Can I sync everything my CRM holds?
You can, but for syncs that use a payload column each row supports up to 250 attributes and up to 1 MB. The better question is which fields change a message.
3. How often can a sync run?
Recurring syncs run from every 5 minutes to once per month. The default shortest interval in the dashboard is 15 minutes, and an on-demand sync is available through the Trigger a sync endpoint.
4. What happens if I send the same values again?
Braze tracks changes by the timestamp column, not by whether the content differs. Resent rows can add unnecessary data point usage on contracts that bill that way.
5. Does this post settle consent or data protection questions?
No. It describes documented mechanics only. Any field that touches personal data, cross-border transfer or subprocessors needs your own security and privacy review.
Sources
- Braze, Cloud Data Ingestion. Retrieved 07 Nov 2026.
- Braze, Cloud Data Ingestion best practices. Retrieved 07 Nov 2026.
- Braze, Cloud Data Ingestion table setup. Retrieved 07 Nov 2026.
- Braze, Connected sources. Retrieved 07 Nov 2026.
- Braze, Data points (Legacy). Retrieved 07 Nov 2026.
- This post rests on five Braze documentation pages retrieved on 07 Nov 2026. Limits, defaults and release status can change, so check the current pages before you build.
- Written by CustomerIK, a Braze implementation partner.
CustomerIK is a Braze implementation partner working across onboarding, technical integration, marketing operations and customer data management.
If you're deciding which CRM fields belong in Braze, let's talk.




