Smart Matching Suggestions: How EVA Recommends Matches
Smart Matching Suggestions are automatically generated recommendations that help you identify which Amazon products and Shopify variants should be matched.
They reduce the need for manual work by surfacing the most likely product connections.
The Big Picture
EVA analyzes your unmatched products and suggests connections based on similarity signals. These suggestions allow you to match products faster while still keeping full control over the final decision.
What Smart Matching Does
The system scans unmatched products and:
- compares products across channels
- calculates similarity between them
- generates match candidates
- ranks them by confidence
Only the most relevant suggestions are presented to the user.
Where Suggestions Appear
Smart Matching Suggestions are visible in two places:
- Match Suggestion popup when entering the Catalog page
- Matching interface inside the Unified Catalog
This ensures suggestions are accessible both at entry and during workflow.
Suggestion Components
Each suggestion includes:
- Amazon product details
- Shopify variant details
- Match score (0 to 100 percent)
- Similarity breakdown
The similarity breakdown typically includes:
- title similarity
- price similarity
- SKU similarity
Suggestion Logic
EVA evaluates multiple signals to determine whether two products are likely the same.
The process follows:
- Compare product attributes across channels
- Calculate similarity scores for each signal
- Combine signals into an overall match score
- Rank candidates based on confidence
Higher scores indicate stronger confidence in the match.
Suggestion Prioritization
Not all possible matches are shown.
EVA prioritizes:
- high-confidence matches
- strong SKU or price alignment
- clear one-to-one relationships
This prevents overwhelming the user with low-quality suggestions.
User Control
Even though suggestions are automated, matching is never forced.
Users can:
- apply the suggestion
- ignore it temporarily
- dismiss it permanently
This ensures that final decisions remain user-controlled.
Why This Matters
Without Smart Matching:
- users must manually scan and compare products
- matching becomes slow and error-prone
With Smart Matching:
- high-confidence matches are surfaced instantly
- decision-making becomes faster and more structured
Key Value Pillars
Reduced manual effort
Users do not need to compare products manually.
Faster matching workflow
High-confidence matches can be applied immediately.
Transparent recommendations
Similarity breakdown explains why a match is suggested.
Pro Insight
Even when product titles differ significantly, strong SKU and price alignment can still produce high-confidence matches.
This allows EVA to correctly identify products across different naming conventions.