Improve Performance with Path Experiments
Learning Objectives
After completing this unit, you’ll be able to:
- Explain A/B testing and its role in effective marketing techniques.
- Explain how to configure a path experiment.
- Apply experiment results to optimize future campaigns
Explore A/B Testing Fundamentals
Most marketers are familiar with A/B testing. You send different versions of an email or offer to sections of your audience and test what lands. Rather than guessing whether a discount or a free gift drives more conversions, you let customer behavior decide.
The challenge with traditional marketing platforms is that testing usually happens outside the journey. You split a list, send separate campaigns, and piece together the results manually. In Marketing Cloud Next, experimentation is built right into the flow. You run the test on the canvas, track results in the analytics panel, and the winning variation can automatically go out to the rest of your audience.
Configure a Path Experiment
You’ve already set up rule-based branching for the Super Kicks campaign. High-value customers, recent purchasers, and standard customers now follow different email paths. Now the marketing team wants to test two versions of the early access offer.
- Path A: Early access with a 15% loyalty discount
- Path B: Early access with a free gift on purchase of Super Kicks
To run this test, use the Path Experiment flow element. Path Experiment lets you compare different offers or messages directly within a campaign flow by combining two capabilities:
- Random Split: Distributes contacts randomly across paths
- Path Optimizer: Evaluates which path performs best
Start by configuring the experiment settings. Choose the metrics used to measure each path’s success, such as email link clicks. Then decide what percentage of the audience to include in the experiment and how long the experiment should run.
When contacts reach the Path Experiment element, the system randomly distributes them across the defined paths based on the percentages you assign to each one. Next, explore how to select a winning path.
Improve Campaign Performance
After the experiment runs, the next step is selecting a winner and applying the results.
Select a Winner
You can choose between automated or manual selection.
-
Automatic selection: The system tracks your selected metric and declares a winner when a path reaches 95% confidence against all other paths. It then automatically sends the remaining contacts, known as the delayed group, down the winning path. If no path reaches 95% confidence before the test ends, the system applies the fallback behavior defined in your experiment settings.
-
Manual selection: You can also review results in the Path Experiment element’s Analytics tab and choose the winning path yourself. You can even declare a winner before the test period ends. Manual selection works well when you want to include data outside the flow, such as CRM revenue data or feedback from the sales team.
In both cases, after identifying a winner, the delayed group follows the winning path and the experiment closes. A History tab records all changes made during the test. This creates a clear audit trail of the experiment.
Apply Insights to Future Campaigns
Selecting a winner ends the test, but the insights can improve future campaigns.
Update the campaign record: Campaign records store performance data for each initiative. After the test concludes, you can record the winning offer type, discount or gift, in campaign notes or custom fields. This makes it easier to apply that insight to future launches.
Refine segments: Results can also reveal differences between customer groups. For example, the free gift might convert better for high-value customers, while the discount performs better for recent purchasers. In future campaigns, you can apply this learning using a Decision element so each segment automatically receives the offer most likely to resonate.
Design better experiments: A narrow result can be just as useful as a clear winner. It may indicate that other variables, such as subject line, send timing, or visual design, are worth testing next.
Review campaign analytics: Marketing Cloud Next includes a Conversion Analytics dashboard. It tracks how email and SMS messages contribute to outcomes such as order completion or form submission within a 30-day conversion window. Reviewing this dashboard after each experiment helps build a broader understanding of what works across campaigns, not just within a single test.
Wrap Up
The Super Kicks campaign is live. Cloud Kicks can now reach every customer with a message that reflects who they are, whether that's an exclusive early access offer for their most loyal spenders, a product recommendation for new customers, or a standard launch message for everyone else. The campaign is no longer one-size-fits-all.
In this badge, you learned how Data 360 unifies customer data through identity resolution and data graphs, and how campaign records organize audiences, content, and flows in one place. You explored how decision logic routes each contact to the right path and how path experiments take the guesswork out of optimization by testing variations directly within the flow.
Now it's your turn. Take what you've learned and build your own dynamic campaigns in Marketing Cloud Next. Connect your data, design your logic, and let the flow do the work.