Experimentation on AEM Edge Delivery Services with the Experimentation Plugin

When you think about testing experiences in Adobe’s ecosystem, Adobe Target is the full-featured solution for cross-channel experimentation and personalization—spanning any site or app, with advanced targeting and decisioning. The AEM Edge Delivery Services (EDS) experimentation framework is complementary: it focuses on lightweight, page- and component-level A/B tests that are tightly integrated with EDS’s authoring and delivery model, with minimal runtime overhead. For many content and UX changes on an EDS site, the built-in experimentation framework (and its plugin) is the fastest way to test; for more sophisticated audience strategies and omnichannel targeting, you’d typically pair EDS with Target and the Web SDK.

Experimentation on EDS is designed to fit the platform’s core strengths: simplicity, speed, and author-friendly workflows. Instead of wiring tests manually, teams can use the experimentation plugin in the AEM Sidekick to define, run, and analyze experiments directly from the same environment they use to author content.

What the Experimentation Plugin Does

The experimentation plugin connects three things:

From the Sidekick, authors and marketers can:

Under the hood, the experimentation framework routes traffic between control and challengers, and Operational Telemetry gathers data so you can see which variant performs better without adding a heavy client-side framework.

How It Fits Into Your EDS Authoring Flow

With the plugin in place, experiments become a natural extension of your existing EDS authoring model:

  1. Create your variants using your normal tools

    • For document-based authoring, duplicate the control document and make content or layout changes.
    • For Universal Editor, create a variant page or adjust components as needed.
  2. Open the experimentation plugin in Sidekick

    • Navigate to the control page in preview.
    • Launch the experimentation plugin from the Sidekick toolbar.
    • Select the control and map each challenger URL as a variant.
  3. Configure basic experiment settings

    • Give the experiment a clear name/ID.
    • Choose the traffic split (for example 50/50, 70/30).
    • Save and publish so the configuration is deployed together with your content.
  4. Run the test and review results

    • Let the experiment run until you have enough traffic and statistical confidence.
    • Use the experimentation reporting to compare variant performance.
    • Promote the winning variant to become the new control and retire the experiment.

Because the plugin is part of the EDS tooling, there is no separate login or external UI to manage, and your experiments are naturally versioned alongside your content and code.

Why This Approach Works Well on EDS

This plugin-based approach is designed to preserve the key properties of an EDS site: