Ecommerce ops · Production

YourSurprise gains efficiency improvements and search functionality with Algolia

The problem

YourSurprise relied on an outdated, over-complicated custom Elasticsearch integration that only one person fully understood, creating a bottleneck where issues were routinely deferred and search optimisation required a ten-person team working country-by-country.

First attempt

The custom Elasticsearch system had become a single point of failure: only one person in the entire company fully understood it, and that individual's overload created a permanent backlog of unresolved search issues.

Workflow diagram · grounded in source
1
Customer search query submitted
trigger
“YourSurprise is using Algolia for its main search bar and on individual gift pages for search as well as for recommendations, to provide a federated global search experience on FAQ and category pages”
2
AI Dynamic Re-Ranking applied
ai_action
“We've tested Dynamic Re-ranking on and off, but also the number of gifts we should re-rank. We tested the sorting strategies based on sales amounts or based on margin”
3
Results and recommendations delivered
output
“The company's FAQ page is completely powered by Algolia, and it is using the platform in emails as well”
4
A/B testing refines strategy
feedback_loop
“By testing and Dynamic Re-ranking, Muys intends to determine the best product sorting strategy for each shopping peak period, but also between peaks”
Reported outcome

YourSurprise reduced its search management team from 10 to 2 people while covering 28 countries, achieved a 9% improvement in conversions, and resolved issues that previously took a day near-instantly.

Reported metrics
Search management headcountfrom 10 to 2 people
Conversion improvement9%
Implementation time<1 month
Countries managed by 2 employees28 countries
Reported stack
AlgoliaElasticsearchReact
Source
https://www.algolia.com/customers/YourSurprise
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

YourSurprise reduced its search management team from 10 to 2 people while covering 28 countries, achieved a 9% improvement in conversions, and resolved issues that previously took a day near-instantly.

What tools did this team use?

Algolia, Elasticsearch, React.

What results were reported?

Search management headcount: from 10 to 2 people; Conversion improvement: 9%; Implementation time: <1 month; Countries managed by 2 employees: 28 countries (source-reported, not independently verified).

What failed first in this deployment?

The custom Elasticsearch system had become a single point of failure: only one person in the entire company fully understood it, and that individual's overload created a permanent backlog of unresolved search issues.

How is this ecommerce ops AI workflow structured?

Customer search query submitted → AI Dynamic Re-Ranking applied → Results and recommendations delivered → A/B testing refines strategy.

Related ecommerce ops cases