May 14, 2025
Einstein in SFMC: Which AI Feature Solves Which Problem
SFMC's Einstein catalog has six distinct features and most teams only know two of them exist. Here's the mapping from common problems to the right Einstein capability.
Read MoreThree ways to produce a segmented audience in SFMC. Pick the wrong one and you either can't express the logic or rebuild the same segment four times a month. Here's the decision matrix.

Three SFMC tools that produce a segmented audience: Filtered Group, Filtered DE, and SQL Query Activity. They look similar on a wireframe - they all turn a larger audience into a smaller one - but the mechanics and refresh behavior differ enough that picking the wrong one costs rework.
Here's the side-by-side and when each applies.
Used when subscribers live in Lists (the older subscriber model). Today most builds use DEs, so Filtered Group shows up mostly in legacy accounts.
Used when subscribers live in DEs and the segmentation logic is simple: attribute comparisons from a single DE.
Point-and-click in Email Studio > Subscribers > Data Filters. Result is a new DE that can be used as the send audience.
Used when the logic needs:
SQL runs on SFMC's limited SQL dialect (SELECT / INNER / LEFT / basic aggregates, no stored procedures, no MERGE). Result writes to a target DE.
Filtered DE does not auto-refresh. Build a Filtered DE of "Gold tier subscribers," import a new file adding more Gold members to the master DE, and the Filtered DE still holds yesterday's list.
Refresh options:
Same pattern for Filtered Groups - they don't auto-refresh either.
SQL Query Activity also needs to run before the send; you schedule it in the same automation.
Rule: any automation that ends in Send Email and uses a Filtered DE or SQL-produced DE as audience should have the refresh step earlier in the same automation. Don't count on a filter built yesterday containing today's data.
We see this often - SQL Query for logic Data Filter would've covered. Two costs:
Before writing SQL, check if Data Filter (covered in the previous post) would've been enough. Often it is.
Signs you actually need SQL:
If any apply, SQL Query Activity is the right tool. Data Filter will either fail to express it or need workarounds.
On large DEs (millions of rows), SQL Query Activities with proper indexes outperform Filter Activities. Filters iterate the whole DE; SQL can use the underlying indexing.
For small/medium DEs the difference is negligible. For large ones, SQL wins for speed, though the complexity trade-off still matters.
For production sends, the automation chain looks like:
SQL Query refreshes the audience DE. Verification Activity confirms the row count is in the expected range. Send Email runs against the freshly built audience.
Three steps, no stale audience, and it halts if the data is wrong.
Filtered Group, Filtered DE, and SQL Query Activity aren't interchangeable. Filtered DE for simple single-source logic. SQL Query when joining, calculating, or reading tracking. Filtered Group only if you're in a legacy List-based account. Either way, refresh the filter in an automation step before the send - manually refreshed filters are the top cause of "we sent yesterday's segment" incidents.
Designing SFMC segmentation architecture? Our Salesforce team ships audience build pipelines with proper refresh automation and verification on production engagements. Get in touch ->
See our full platform services for the stack we cover.
We build trust by delivering what we promise – the first time and every time!
We'd love to hear your vision. Our IT experts will reach out to you during business hours to discuss making it happen.
"Collaborate, Elevate, Celebrate where Associates - Create Project Excellence"
SapotaCorp beyond the IT industry standard, we are