English
SkupSzop has 300,000 books, mostly unique used copies. Luigi's Box Search and Recommender made them findable and lifted basket value by 2,36 €.
SkupSzop is a Polish platform for buying and selling new and used books. With a catalog of 300,000 SKUs, most of them single-copy items in varying conditions, it operates at a scale and complexity that most standard search solutions aren’t built for. Customers buy and sell their books through the platform at competitive prices, with thousands of new items added every day.
Search quality improved significantly, and basket value grew. Our team no longer needs daily manual maintenance, and we estimate Luigi’s Box contributes an additional 10-15% to our revenue.
SkupSzop runs a fully custom backend, so the integration went through the API. Luigi’s Box CTO, Tomáš Kramár, was directly involved in the early stages to handle the technical complexity.
The catalog was the hard part. The same title can exist in multiple condition variants, attached to an author, a publisher, a series, a cycle, and a set of tags, all of which customers might search by. To handle this, Luigi’s Box built a multi-feed architecture all linked via relational IDs.
That structure allows search to treat authors, publishers, series, and condition variants as distinct attributes rather than undifferentiated text. When someone searches for a specific author or a book in a particular condition, the system knows what they mean.
SkupSzop’s previous search, built on Elasticsearch and PrestaShop, had no AI-driven relevance, no typo tolerance, and no synonym matching. Specific queries like ISBNs failed entirely, and the popularity scoring pushed the wrong books to the top, showing irrelevant titles instead of what customers actually wanted.
Luigi’s Box replaced the engine with behavior-driven search that understands what customers mean, not just what they type. ISBN lookups work, typo tolerance handles imprecise queries, and popularity signals now reflect actual customer behavior rather than arbitrary scoring. Search usage rate grew by 11.4% year-over-year, and not-clicked searches dropped by 7.34%.
Most of the books SkupSzop sells are one-of-a-kind used copies that sell quickly. A stale feed means customers see products that are already gone. Standard setups couldn’t handle this combination of format, size, and frequency.
Luigi’s Box built a custom API integration to process the heavy JSON data on schedule. If anything fails, the team gets an automatic email alert before customers notice.
SkupSzop adds more than 4,000 new books every day at exactly 5 p.m., but their setup had no way to schedule when products go live. On top of that, all-time popular titles dominated the product listing pages, so fresh arrivals were buried before customers had a chance to find them.
Luigi’s Box implemented feed scheduling where the new inventory pre-loads and goes live automatically at 5 p.m. A custom one-day trending window runs alongside the standard three-day window on category pages, giving fresh drops a real chance to surface before longer-running titles take over.
SkupSzop didn’t have automated recommendation capability. They had to set up every recommendation mechanism manually. Customers routinely bought single items with nothing prompting them to add more.
SkupSzop planned to use two or three recommendation modules. Once they saw results, they kept adding more and now run 15, placing recommenders across category pages, product detail pages, the shopping basket, and blog content, including article-to-product matching that surfaces relevant books within their author features and reviews. Average basket value increased by 10 PLN (~€2.36) directly from the recommendations.
The previous Elasticsearch and PrestaShop setup ate hardware resources, slowed response times, and required manual maintenance. Products out of stock for over 60 days piled up in search results with no way to catch or remove them. The team had no way to separate mobile from desktop traffic, there was no way to spot which searches were failing and no reliable basis for performance reviews.
Luigi’s Box removed the need for the daily maintenance. Analytics now break down pageviews, requests, and events by desktop, mobile app, and OS. The SkupSzop team can see long-term out-of-stock items so they don’t clutter recommendations.
Luigi’s Box reliably processes our heavy product feed, something other search vendors couldn’t handle. Their AI-driven search and extensive suite of recommendation modules enhanced our cross-selling and increased average basket value.
Results were measured year-over-year, December 2024 vs. December 2025. As search results got more relevant, more customers used search to find what they were looking for. Recommendations did the rest, nudging basket value up with every session.
Lucide chose Luigi’s Box and achieved an 8% B2B conversion rate uplift and other improvements. Find out more in our case study.
Read Case Study
Barron is a custom merchandise seller that chose Luigi’s Box and received a Product Listing conversion rate of almost 18% and other important improvements.
Read Case Study
Switching to a new search paid off and brought a 175% increase in search usage, and positive results for search conversions, clicks, and other important areas.
Read Case Study
GAP chose Luigi’s Box and improved their website with new search functionality that brings results across multiple domains.
Read Case Study
Chicco's switch to Luigi's Box brought them numerous improvements, including Autocomplete CTR and Search usage increases.
Read Case Study
