# Typo-tolerant site search: +11.7% revenue per user

- Outcome: Won
- Page type: Site-wide (Ecommerce)
- Elements: Filters, sorting & site search
- Test run by: [AB Tasty](https://liftarchive.com/practitioners/ab-tasty/)
- Commentary by: LiftArchive (our analysis, not theirs)
- Original source: https://wingify.com/success-stories/search-success-how-la-fee-maraboutee-boosted-revenue-with-ab-tasty-search/
- Page: https://liftarchive.com/test/la-fee-maraboutee-typo-tolerant-semantic-search/

Swapping native Shopify search for a typo-tolerant, semantic search engine lifted revenue per user by 11.7% for La Fée Maraboutée. Add-to-cart rate rose 9.4%. AB Tasty ran the test with its own search product, and it was published as a VWO case study.

## Result

**+11.7% revenue per user**

- Revenue per user: +11.7%.
- Add-to-cart rate: +9.4%.
- Both figures come from the first two weeks of the campaign.

## Control

The native Shopify search preview, which matched the brand's look but had no typo tolerance or semantic matching. Small spelling errors could return poor results or none.

## Variant

A search preview that looked the same but ran on an engine that handled typos and understood related terms.

## Why it likely won (our reading)

- Searchers show strong buying intent, so fixing 'no results' pages helps the most valuable visitors.
- Typo handling and synonym matching return relevant products for queries that used to fail.
- Keeping the same design meant the gain came from better results, not a new look.

## Caveats

The vendor tested its own search product, and the results cover only two weeks during a peak shopping period. Sample size and confidence were not shared.

## How to run this test

- Hypothesis: If we make site search tolerate typos and match related terms, then more searchers will find and buy products because fewer queries will end in poor or empty results.
- Primary metric: Revenue per user
- Guardrail: Search exit rate
- Guardrail: Zero-result search rate
- Guardrail: Add-to-cart rate
- Who should test it: Stores with large catalogs where many visitors use search. It matters less for shops with a handful of products that people browse rather than search.

---

Credit: AB Tasty ran this test and owns the result. Summary and commentary by LiftArchive (https://liftarchive.com/test/la-fee-maraboutee-typo-tolerant-semantic-search/).
Usage: Free to use, quote and build on, with credit and a link: name the practitioner or organization that ran the test and link to its LiftArchive page. The results belong to the practitioners who ran and shared them; the summaries and commentary are LiftArchive's.