EcomAtlas

Module

Search results and recovery

For
Lean founder-operator · also Head of ecommerce / growth, Ecommerce marketer becoming an operator
Not for
Stores that intentionally provide no storefront search

Published · v1.0.0

Locked

Lessons

3 lessons, in order.

0/3
  1. Lesson 1: Give every empty search a next move

    Explain the empty result, preserve the query, and offer recovery paths for products and store information.

    Locked
  2. Lesson 2: Auto-redirect search only for an exact category match

    Send a product-type query to a category page only when the intent and destination are genuinely equivalent.

    Locked
  3. Lesson 3: Show shoppers why each search result matched

    Keep the query visible and expose the attribute or text that makes a result relevant.

    Locked

Search results must help shoppers understand what matched, assess the products, and recover when the answer is empty or wrong. The business outcome is a useful continuation path for both product discovery and store-information tasks.

This module covers result context, relevance, filtering and sorting within results, no-result recovery, and non-product queries such as shipping, returns, and order tasks. These decisions belong together because a technically non-empty page can still be a dead end when the results are irrelevant. Baymard supports specific recovery routes for unsuccessful searches and documents shopper expectations that ecommerce search can reach non-product content.12

Start with query intent. Verify whether the shopper seeks a product, category, policy, or task. Then inspect result quality, list controls, empty-state handling, and the next action. Do not treat unrelated recommendations as recovery.

Module diagnostic

Run a known product, broad product type, unavailable product, difficult spelling, empty filter combination, return-policy query, shipping query, and account-task query. On desktop and mobile, record the retained query, result explanation, relevant products or content, active filters, alternative routes, and ability to search again. Compare predictive suggestions with submitted results.

Small assignment

Choose the most important failed query. Map its likely intent, correct destination, current failure, smallest data or interface fix, owner, regression query, and acceptance criteria. Use search reports to prioritize repeated failures, but inspect the query before assigning intent.

Evidence and further reading

  1. “Search UX: 5 Proven Strategies for Improving No Results Pages,” Baymard Institute, https://baymard.com/blog/no-results-page. Supports offering specific related queries, categories, and other recovery paths. Accessed August 23, 2026.
  2. “E-Commerce Search Needs to Support Users’ Non-Product Search Queries,” Baymard Institute, https://baymard.com/blog/support-non-product-search. Supports surfacing relevant store content and tasks for non-product queries. Accessed August 23, 2026.

Was this helpful?