A/B testing on the Tagalys dashboard allows merchants to simultaneously test multiple variants of the same collection with different merchandising strategies. Here are some use cases of A/B testing:
Use case: Compare the performance of a collection with Smart Sequencing enabled vs. disabled
Click here to understand the functionality of smart sequencing
How it works:
- Create an A/B test with 2 variants:
- Variant A: The current version of the collection page can be merchandised using sequences of products. The sequences will remain static until manually reordered.
- Variant A: The current version of the collection page can be merchandised using sequences of products. The sequences will remain static until manually reordered.
-
- Variant B: The secondary variant of the collection page, will also be merchandised using product sequences but with the 'smart sequence' feature enabled. This will enable the product sequences to be automatically re-ordered once a day based on the chosen performance metric. In the below example, units sold over the last 30 days is the metric based on which sequences will be reordered.
- Variant B: The secondary variant of the collection page, will also be merchandised using product sequences but with the 'smart sequence' feature enabled. This will enable the product sequences to be automatically re-ordered once a day based on the chosen performance metric. In the below example, units sold over the last 30 days is the metric based on which sequences will be reordered.
- Objective of the test: To determine whether activating smart sequencing (which automates the sorting of product sequences) enhances the overall effectiveness of the collection page.
Use case: Tagalys 'Trending' vs. Manual Product Pinning
- How it works:
- Create an A/B test with 2 variants.
- Variant A: The primary variant of the collection can be merchandised by manually pinning products in the desired positions.
- Variant B: The secondary variant can be merchandised dynamically using the Tagalys 'trending' sorting option where products are being reordered based on the trending score of the product.
- Variant A: The primary variant of the collection can be merchandised by manually pinning products in the desired positions.
- Create an A/B test with 2 variants.
- Objective of the test: To analyze if pinning products to static positions in a collection results in better performance compared to automating the product sorting based on the trending score.
- To take it a step further, users can also create a third variant which can be a combination of both merchandising strategies. For instance, pinning the first 20 products and then using 'Trending' to sort the remaining products to analyze if that would be the best approach.
Use case: Product Sequences vs 'Field'
- How it works:
- Create an A/B test with 2 variants:
- Variant A: The primary variant can be merchandised using product sequences to create a visually pleasing sorting order
- Variant B: The secondary variant can be set up using the 'Field' merchandising option to unidirectionally sort the entire collection page using a single performance metric. (For example, conversion %)
- Variant A: The primary variant can be merchandised using product sequences to create a visually pleasing sorting order
- Create an A/B test with 2 variants:
- Objective of the test: To understand which strategy yields better results, whether displaying products in a visual story by adding structure and showing variety or arranging them based on a performance metric.