Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Enhanced Search & Suggestions

Problem Statement

Currently, the search function only looks at restaurant names and sometimes returns irrelevant “Did you mean” suggestions when a typo or incomplete word is entered. It does not cover deeper restaurant metadata (e.g., menu names, package names, tags, description), making it harder for users to find what they’re looking for.

[

hungryhubgroup.slack.com

https://hungryhubgroup.slack.com/archives/C02DA311WVD/p1746625848939299?thread_ts=1746613981.353339&cid=C02DA311WVD

](https://hungryhubgroup.slack.com/archives/C02DA311WVD/p1746625848939299?thread_ts=1746613981.353339&cid=C02DA311WVD)

Old Design

Core Search

Searchable fields (need to arrange the priority for the search result)

  • Restaurant name (provided)
  • Restaurant tags (provided)
  • Restaurant description → (provided)
  • Package names → (provided)
  • Package group names → (provided)
  • Package tags/label → phase 2
  • Menu names → phase 2
  • Menu descriptions → phase 2

https://hungryhub.com/admin/custom_labels?locale=en

Survey Pop Up

Provide survey pop up to get the feedback from user after search and get the result → evaluate the re ranking Private (https://app.clickup.com/9003122396/docs/8ca1fpw-7922/8ca1fpw-50896)

Keyword

Related Keyword: "{keyword}" in {section} like google search results → Design file : https://www.figma.com/design/xI2931FaU8rYUuwMY4BzVz/Search-2.0--Desktop-?node-id=1772-16953&t=BVacIS37V7EZhbVm-1 Scope : Search Suggestion and Search Result (web and app) → pls check design file

  • No need related keyword for restaurant name

Indexing approach

  • Combine structured fields (e.g., name + tags) : copper international
  • Include extracted keywords/phrases from longer texts (e.g., extract clipped text from descriptions)
  • Consider adding synonyms for common food terms (e.g., “shabu” = “hot pot”)
  • Include AI/ML-generated alternative phrasings (e.g., “romantic dinner” = “candlelight dinner”, “birthday restaurant” = “birthday-friendly”)

“Did You Mean” Suggestions

  • Must be dynamically generated from actual searchable data (not a hardcoded suggestion list).
  • Match against:
    • Top searched keywords (query logs)
    • Similar restaurant/menu/package names
    • Common misspellings or fuzzy matches
  • Use typo tolerance (Levenshtein distance or similar) to surface meaningful corrections.
  • Prioritize suggestions that exist in the current database over generic words.

Suggestions

  • ✅ Mobile → Show up to 3 suggestions
  • ✅ Desktop → Show up to 5 suggestions
  • ✅ If the line overflows, wrap to a new line (no horizontal scrolling, no truncation ellipsis) This keeps the design:
  • Clear
  • Readable
  • Consistent across screen sizes
  • Accessible without fancy hover/tap-to-expand interactions

Search Performance

  • Search backend should support fuzzy search and multi-field queries efficiently (consider OpenSearch, Elasticsearch, or similar).
  • Results should return within ≤1 second for a good user experience.
  • Maintain a lightweight frontend implementation (search bar, result list, and suggestion display).

Success Metrics

✅ Increase search success rate (measured as % of queries that return at least one result). ✅ Decrease bounce rate on search results page. ✅ Improve click-through rate on “Did you mean” suggestions. ✅ Reduce customer support complaints about “No Result Found” errors.