Testing Your Configuration

Server-side implementation of Discover requires the integration on Category/Browse listing pages and Search listing pages to provide the desired response.

Testing lets retailers refine the assigned weights to find the optimal settings for their site.

Setting up tests should be done with a clear idea of what is being tested and a statistical understanding of significance.

You optimize sort for your site and your shoppers by running a series of MVTs comparing different weights for the personalized and global sort algorithms.

Multivariate Testing (MVT)

For full instructions on how to set up MVT testing, see MVT Overview. 

A/B Tests

Often retailers want to run A/B tests comparing Omnichannel Personalization sort results to their in-house sort solutions. The retailer manages the A/B test, splitting traffic between Search/Browse and their default sort.

• Retailer A/B tests should take place after optimization MVTs have completed. Running an A/B test while optimization MVTs are still underway will inherently include less-optimal performance in Omnichannel Personalization's results.

• We recommend having equal-sized test groups in the A/B.

• The retailer instruments a treatment ID so {rr} can measure performance during testing.