A/B Testing – Implementation (Mapp vs Competitor)

  • 1 minute(s) read
Prev Next

Goal

Compare the performance of Mapp Fashion recommendations with a competitor solution using A/B testing.

Prerequisites

  • Mapp Intelligence tracking is implemented

  • Recommendation tracking is configured

  • A method for assigning variants is available (for example, A/B testing tool or custom logic)

Procedure

Step 1 – Assign Variants

Assign each user to a variant.

Use one of the following approaches:

  • An A/B testing tool (for example, Optimizely, AB Tasty)

  • Server-side logic in the storefront or backend

  • Client-side logic in the browser

Ensure that:

  • Each user remains in the same variant during the session

  • The assigned variant is available for tracking

Step 2 – Pass the Variant to Tracking

Include the assigned variant in tracking data.

Use a consistent parameter value, for example:

  • similarItems_mapp

  • similarItems_competitor

You can pass the variant using:

  • Recommendation tracking context (recommended for recommendation comparisons), or

  • Custom parameters (for example, ABTest and ABTestVariant)

Ensure that:

  • The same value is used across all events

  • The variant is included in page views, product interactions, and orders

Step 3 – Track Interactions

Ensure that all relevant interactions are tracked via Mapp Intelligence.

This includes:

  • Page views

  • Product views

  • Add-to-basket actions

  • Orders

If recommendation tracking is used:

  • Interactions with recommendation elements (such as clicks) are tracked

  • These interactions can be linked to downstream product and order events

Variant A – Mapp Fashion

  • Recommendation content is delivered by Mapp Fashion

  • Interactions are tracked via Mapp Intelligence recommendation tracking

Variant B – Competitor

  • Recommendation content is delivered by a third-party solution

  • Tracking is still performed via Mapp Intelligence

Limitations may apply:

  • Recommendation interactions may not be available

  • This depends on whether the third-party solution exposes interaction data (for example, not restricted by an iframe)

Step 4 – Ensure Consistent Tracking

Verify that tracking is consistent across all variants.

Ensure that:

  • Variant naming is consistent across all events

  • The variant is included in all relevant tracking requests

  • The variant is stored in a central location (for example, data layer)

  • Variants are not mixed within a session

Also ensure that:

  • Page views are tracked before product, basket, or order events

  • All events are correctly linked within the same session

Result

  • Each user is consistently assigned to a variant

  • All interactions and conversions are tracked with variant context

  • Recommendation performance can be compared across variants in Mapp Intelligence

Common Pitfalls

  • Inconsistent variant assignment

    Users switch between variants within a session

  • Missing variant in tracking

    Not all events include the variant

  • Limited tracking for competitor solutions

    Recommendation interactions cannot be captured

Next Steps