📖 Semantic Search helps shoppers find the right products even when they phrase things differently (e.g., searching “beach clothes” after “swim”) by using natural language processing (NLP) to infer intent—not just match keywords.
How to use Semantic Search
Go to Product Discovery > Search > Ranking > Semantic Search tab
Use the slider to set how strongly Semantic Search influences results.
In Relevant settings section, pick which product attributes should be searched when a shopper types a query (Searchable fields). If an attribute shouldn’t affect results, set it to Non-Searchable.
For each searchable field, choose a Search weight to influence ranking: High, Medium, Low, or Non-Searchable.
Higher-priority attributes have more impact on order of results.
Example: If a shopper searches “pink” and color is High while product title is Medium, items matching color appear higher.
Click Save settings (or Discard to reset).
Best practices
Start at the default 50% impact and adjust gradually based on conversion and click-through.
Put the most important attributes (e.g., color, size, title) at High to guide ranking toward what matters most to your shoppers.
Use Non-Searchable for attributes that create noise.
FAQ
What does “search relevance” mean?
What does “search relevance” mean?
It’s how closely results match a user’s intent and needs, with the most relevant items ranked first.
What happens at 100% impact weight?
What happens at 100% impact weight?
Semantic Search fully controls the search engine, and queries can return a broader set of relevant products.
Can Semantic Search work without exact keyword matches?
Can Semantic Search work without exact keyword matches?
Yes—NLP techniques help match intent even when words differ (e.g., “beach clothes” vs. “swim”).
How do I stop a field from affecting results?
How do I stop a field from affecting results?
Set that attribute to Non-Searchable.
Feel free to reach out to our dedicated support team via chat if you have any questions or require additional assistance.


