Let Marketing Cloud Next pick the winning path in a campaign flow
Automated Path Selection tests messaging, timing and channel in Flow Builder for Marketers and picks the winner on engagement or custom data. Advanced.
Since Winter '26, a campaign flow in Marketing Cloud Next can test variations of messaging, timing and channel and then choose the winning path automatically. Salesforce's Winter '26 release highlights place Automated Path Selection in the Agentforce Marketing Advanced editions, and state that the selection runs on engagement metrics or on custom data such as purchases or webinar attendance.
- Product
- Marketing Cloud Next
- Edition
- Advanced
- Topic
- Campaigns
- Confidence
- High - documented by Salesforce
- Last verified
Who this applies to
- Marketers running campaign flows in Flow Builder for Marketers
- Marketing Cloud Next Advanced Edition. The source names Advanced explicitly; it does not list Growth
- Access to the campaign flow and to the data you want to judge the winner on (engagement, or a custom object such as purchases)
Why it matters
Most teams never run the second test because the first one took a person to read the numbers and rewire the flow. Automated selection removes that step, and because the winner can be judged on custom data rather than clicks alone, the test can be tied to what the campaign was for, not to the metric that was easiest to measure.
Try it
- Pick one campaign flow that already runs on a schedule and where the variable is genuinely in doubt: send time, channel, or a message variant.
- In Flow Builder for Marketers, set up the test with two or three paths that differ in exactly one thing. A path that changes copy and channel at once cannot tell you which mattered.
- Choose the metric the winner is judged on. Use an engagement metric for a top-of-funnel message, or a custom data point such as a purchase or a registration when the campaign has a real conversion.
- Let the experiment run long enough to collect enough events on each path before the selection kicks in. Check the volumes before you trust the result.
- Record the winning path and the metric in the campaign record so the next person knows why the flow looks the way it does.
Example use case
A member-renewal campaign sends a reminder 30 days before expiry. Two paths test the same email at 8am and at 6pm, judged on renewals recorded on the membership object rather than on opens. After the first cycle the flow routes everyone down the path that produced more renewals, and the team stops arguing about send time.
Prerequisites, limitations and risks
- Advanced Edition, as stated by Salesforce's Winter '26 release highlights. Growth is not listed
- The metric you judge on must exist in your org's data. Judging on purchases needs purchase records connected to the individual
- A test needs volume. On a small audience the selection may run on very few events; the source does not state a minimum, so treat early winners with caution
- Variations in channel need that channel provisioned (SMS and WhatsApp are add-ons with their own consent requirements)
Documented behaviour above comes from the sources listed below. Availability varies by edition, SKU and region - confirm entitlements in your own org before you rely on a feature.
Sources
- Winter '26 Release Highlights: see what's new with Service, Marketing and Commerce (Trailhead)
- Winter '26 release notes
Stuck on this in your own org?
Blueport is a Sydney-based Certified Salesforce Consulting Partner with 100+ implementations since 2018. If a step above does not match what you see in Marketing Cloud Next, tell us - it is either an edition difference or a change worth knowing about, and we will find out which.
Talk to our team