Why Flow performance matters
Salesforce Flow is a powerful no-code tool, but poorly designed flows can cause higher CPU time, exceed SOQL/DML limits, slow user interfaces, and increase costs. Optimizing Flow performance leads to faster page loads, better scalability for data volume, and fewer platform limit exceptions.
1. Choose the right trigger and context
Prefer before-save (Fast Field Updates) record-triggered flows for simple field calculations and assignments because they run before DML and don’t consume extra DML operations. Use after-save flows only when you must: for sending emails, creating related records, or performing actions that require the record ID or committed data.
2. Minimize SOQL and DML — bulkify your logic
Avoid doing record queries and updates inside loops. Collect records in a collection variable, perform a single Get Records where possible, and use a single Update Records element on the collection. This reduces the number of SOQL/DML transactions and keeps Flows within governor limits.
3. Use filters, field selection and limits on Get Records
When using Get Records, filter precisely and only select the fields you need. If you only need one record, set the element to return only the first match and avoid bringing back large result sets into memory.
Example: Get Records (Accounts) — Filter: AnnualRevenue > 1000000, Return: First record only, Fields: Id, Name
4. Reduce and optimize looping
Loops are expensive. Try to convert logic into collection-based operations (Update Records using a collection) rather than per-record processing. If a loop is necessary, use map-style collection handling with assignment elements to avoid nested queries or updates inside the loop.
5. Reuse logic with Subflows and Invocable Apex
Move repeated or complex operations into subflows to centralize logic; however, keep subflows lean and avoid excessive subflow calls in tight loops. For CPU- or query-intensive operations, write bulkified invocable Apex — Apex is more efficient for complex data processing than Flow elements in high-volume scenarios.
6. Leverage Scheduled Flows and Batch Processing
For large data updates, use scheduled-triggered flows or batchable Apex instead of running record-triggered flows for each record change. Scheduled flows can process records in bulk during off-peak hours and reduce contention and UI latency.
7. Limit screen components and heavy elements
Screen flows should keep UI components minimal for faster rendering. Avoid unnecessary lookups and rich components on screens. Lazy-load choices where possible (populate dynamic picklists via a single Get Records element instead of repeated queries).
8. Monitor and debug — measure before and after
Use Flow Debugger, Debug Logs, and the Flow Interviews & paused flow monitoring to track execution time, element counts, and SOQL/DML usage. Note the number of interviews and average execution time to identify hotspots. For invocable Apex, use Apex debug logs and Limits methods to measure actual resource usage.
9. Optimize decision elements and formula fields
Keep decision logic simple and short-circuit checks in the most likely paths first. Complex formulas evaluated repeatedly can be moved to a formula resource that’s computed once, or to a before-save update so it executes only when needed.
10. Error handling and graceful failure
Design fault paths to capture and log failures (Custom Object for errors, Platform Events, or email alerts) instead of letting failed interviews retry or cause cascading issues. Proper error handling reduces repeated expensive retries and helps you spot expensive failing logic.
Quick checklist for Flow performance
– Use before-save flows for simple field updates
– Minimize element count and avoid SOQL/DML in loops
– Use collections and bulk Update Records
– Filter Get Records and select only required fields
– Consider invocable Apex for heavy processing
– Use scheduled flows for large batch jobs
– Monitor with Flow Debugger and platform logs
Example: Replace per-record update loop
BAD (conceptual): For each Lead record, perform a Get Records and Update Records inside the loop — causes N+1 queries and many DML statements.
BETTER: Use a single Get Records to collect related data, build collections with Assignment elements, then perform one Update Records on the entire collection. If logic is complex, call an invocable Apex method that accepts a list and handles processing server-side.
Summary
Optimizing Flow performance is about reducing unnecessary queries and DML, choosing the correct flow type, bulkifying processing, and using Apex or scheduled processing when flows are not the right tool. Measure, refactor, and monitor — small improvements in flow design can yield large gains in scalability and reliability.