Search relevance
RankReceipt
Every search result comes with its receipt.
The problem
Search results can look plausible and still put the wrong record first. Without the reason behind a score, changing the ranking becomes guesswork.
The solution
I built RankReceipt so each result shows the rules and points behind its position. Load a CSV, TSV, or JSON list, correct the column mapping, then tune field weights and synonyms against searches with known answers.
Read the receipt
- Exact ID / SKU
- 1,000
- Title equals the query
- 800
- Alias match
- 240
- Brand match
- 150
The result explains which rules contributed. Change a weight, then compare against a pinned baseline.
The workflow
Load a list and map ID, title, brand, category, aliases, and other text to the correct fields.
Run a query and read the rule-by-rule score. Pin the results as a baseline before changing the rules.
Adjust one weight or synonym, then run saved test searches. Set the expected record and the top-result threshold to see whether the change helped.
Export the rules, results, or test report. Files are read in the browser; this workspace does not replace a store’s live search engine.
A conversation is a good place to start.
Let’s talk about
the work.
Send the role, or your website and what needs to change.
Warren, Michigan