Skip to main content

Search

Concept: Configurable search indexing, algorithms, and interfaces.

The Concept​

Search in Touchpoint is not a single, monolithic feature. It's a configurable system controlled through search configs that define:

  • Indexing — What data to index and how
  • Algorithm — How to rank and filter results
  • Interface — How to display results in the UI

Multiple search configs can exist, allowing different search experiences for different product types, categories, or use cases.

Search Configs​

A search config defines a complete search experience:

Multiple Configs for Different Needs​

Example Scenario:

Search Config A: "Industrial Supplies"

  • Indexes technical specifications
  • Ranks by exact match priority
  • Shows detailed filters (material, size, rating)

Search Config B: "Office Products"

  • Indexes brand and category
  • Ranks by popularity
  • Shows simple filters (category, price range)

Search Config C: "Custom Products"

  • Indexes configuration options
  • Ranks by customization compatibility
  • Shows configuration-aware filters

Indexing Strategy​

The search config defines what gets indexed and how:

Product Type-Specific Indexing​

Different product types index different fields:

Configurable products:

  • Base product data
  • All configuration attributes
  • Option values and combinations
  • Compatibility rules

Standard products:

  • Product name, SKU, description
  • Category and brand
  • Specifications
  • Inventory status

Services:

  • Service description
  • Availability schedules
  • Pricing tiers
  • Location coverage

Category-Specific Indexing​

Refine indexing per category:

  • Plumbing supplies → Index dimensions, material, pressure ratings
  • Electronics → Index specs, compatibility, certifications
  • Apparel → Index sizes, colors, materials

Algorithm Configuration​

Each search config can use a different search algorithm:

Ranking Strategies​

Exact Match Priority:

  • Exact SKU matches rank highest
  • Exact product name matches next
  • Partial matches lower

Relevance Scoring:

  • TF-IDF or BM25 text matching
  • Boost important fields (name, SKU)
  • Decay by distance from search terms

Business Rules:

  • Boost in-stock items
  • Boost margin leaders
  • Boost featured products
  • Demote discontinued items

Custom Algorithms:

  • Tenant-specific ranking logic
  • Industry-specific relevance
  • Context-aware scoring

Filtering Logic​

Hard Filters (Must Match):

  • Category must be "Plumbing"
  • Must be in stock
  • Must ship to customer's zone

Soft Filters (Preference):

  • Prefer brand X
  • Prefer price range Y
  • Prefer recent additions

UI Interface Configuration​

The search config defines how results are displayed:

Filter Panels​

Configure which filters appear:

Product Search Config:
  Filters:
    - Category (multi-select)
    - Price Range (slider)
    - In Stock (toggle)
    - Brand (multi-select)
    - Material (multi-select, if category = "Industrial")

Result Display​

Configure data bindings for result cards:

Search Result Slot configured with:
  - product.name
  - product.sku
  - pricing.finalPrice
  - inventory.inStock
  - product.primaryImage

Sort Options​

Define available sort methods:

  • Relevance (default algorithm ranking)
  • Price: Low to High
  • Price: High to Low
  • Newest First
  • Name: A-Z

Multi-Config Example​

Scenario: A distributor sells both industrial supplies and office products.

Industrial Search Config:

Indexing: Technical specs, materials, dimensions
Algorithm: Exact match priority (customers know part numbers)
UI: Detailed filters (spec-heavy), compact result cards

Office Search Config:

Indexing: Brand, product type, features
Algorithm: Relevance scoring (customers browse)
UI: Simple filters (category, price), visual result cards

Same search infrastructure, completely different experiences based on config.

Search Flow​

Benefits​

Per-Category Optimization​

Refine search for each product category without affecting others.

Tenant Customization​

Each tenant can have custom search algorithms and interfaces.

A/B Testing​

Deploy different search configs and compare results.

Evolution​

Update indexing, algorithms, or UI independently without code changes.

Learn More​

For related concepts, see:

Search: One engine, infinite configurations.