What are the most common problems encountered during the implementation of Size Adviser? How can they be avoided?

Testing Environments

We strongly encourage testing in non-production environments. If you maintain a staging, development, or test environment, please provide us with the relevant URLs and credentials. We will whitelist them in our system to perform comprehensive end-to-end testing from our side before going live.

Primary Implementation Challenges

The vast majority of issues encountered during the onboarding phase relate directly to data quality. Specifically, we frequently encounter:

  • Sizing Discrepancies: Inconsistencies between the real physical fit of products and the information provided in the size guides.
  • Feed Inaccuracies: Missing fields, empty values, or inaccurate product information within the synchronized catalog feed.

The Impact of Missing Size Systems

A critical roadblock is the absence of explicit size system specifications. When a feed fails to provide the regional or demographic context for its sizing, identifying sizes accurately becomes impossible.

As illustrated in the shoe sizing example below, a single size number (such as 5) corresponds to entirely different foot measurements depending on whether it follows the UK, US Men’s, or US Women’s scale:

Size Systems

Without definitive data declaring the target size system, our engine cannot reliably map the product to its true physical measurements.

Content Mapping Errors

Minor errors in basic product attributes can severely degrade recommendation accuracy. Incorrect or missing values for the following fields have a significant impact on performance:

  • Gender (e.g., Men’s vs. Women’s cuts)
  • Age Group (e.g., Infant, Toddler, Kids vs. Adult)
  • Product Category (e.g., Slim-fit tops vs. outerwear)

Strategic Prevention

To avoid these implementation hurdles and drastically minimize errors, ensure that all required fields in your product feed are completely populated, accurate, and structurally consistent. Prioritizing the health of your catalog data guarantees the highest possible recommendation accuracy and a seamless experience for your customers.