NEW: E-commerce search benchmarks and insights. Click here for updated 2024 data.

How Muziker and Luigi’s Box Improved Product Recommendations

Ready for some inspiration? Take a look at our case study, and learn how Muziker and Luigi's Box have revolutionized online product recommendations together.

How Muziker and Luigi’s Box Improved Product Recommendations

The cooperation between Muziker and Luigi’s Box is an example of a successful implementation of Luigi’s Box Recommender innovations, resulting in increased sales and providing high added value for the client and their customers. Together, we managed to discover new possibilities of the Recommender and implement them successfully. Personalization of content is a topic for every e-shop today, and individual customization is an indispensable part of the development of our products.

What is Muziker?

Muziker is a leading European multisegment online shop specializing in selling products for all leisure activities aimed at bringing people joy. The entire portfolio is available to customers in more than 30 countries. Muziker, as we know it, is a stable partner constantly striving to innovate and provide something extra for its customers.

One of the first clients

Muziker was one of the first clients to start using Luigi’s Box Search services on their e-shop. Even in the early stages of cooperation, years ago, they realized the importance of quality search for their customers. With its help, they managed to create an effective and accurate product search system that allows customers to quickly and easily find exactly what they are looking for. It was logical, therefore, that when looking for personalized recommendations (Recommender), Muziker turned to us.

Why Recommender?

Today, personalization of content is an indispensable part of the development of online stores, and Muziker recognizes the importance of this strategy to acquire and retain customers. Luigi’s Box Recommender allows the client to explore the habits and preferences of their customers and, based on this data, offer products that will interest them and be relevant to them. This increases the chance that customers will discover and purchase products they are looking for precisely in their online store, which has a direct impact on their satisfaction and loyalty to the brand.

“We want the recommender, but the best one!”

Muziker, as a client, demonstrated from the outset its commitment to using innovative technologies and strategies to improve its e-shop and provide customers with a better, more personalized range of products. Therefore, it wanted to implement various recommenders in its e-shop, not only at the product detail or shopping cart level. However, the path was not entirely straightforward because Muziker wanted the best product and wanted to choose logically.

“Muziker” hackathon

Do you know what a hackathon is? It is an activity where people meet for a few hours or days and intensely collaborate to create, for example, software solutions to a current problem or challenge. This is exactly the kind of challenge Muziker decided on. “As part of the hackathon, we focused on personalizing website content. Before it started, we spent a few days finalizing topics and decided to use Luigi’s Box Recommender since we already had positive experiences with their search service,” says Muziker’s CTO, Ondrej Proksa, about the hackathon.

On Wednesday evening, they contacted Ondrej Kaššák, our data scientist, requesting a recommender that would work not only in the cart and on product pages but also on the homepage, segment, and campaign pages. By Friday morning, he had already sent a basic draft of the recommender.

The hackathon started on Friday, and during the discussions, the client decided that in addition to products, we could recommend categories on the homepage as they have many categories and wanted to guide customers not only to specific products but also to the relevant categories. On Friday evening, they again contacted Ondrej with this requirement, and a few hours later, by morning, the solution was ready. “I was amazed at the speed with which Ondrej prepared a solution that meets our expectations, especially because we were moving outside of regular products and working with categories,” recalls Ondrej Proksa from Muziker. “We also involved our product content team and came up with an adaptable system where we can create various control elements on the homepage or segment pages,” he continues. “On Saturday, we integrated the Luigi’s Box Recommender, programmed the control element system, and on Sunday, we could present it.”

Finding Solutions

The challenge for us was how to make a homepage recommender that would not recommend products, but categories. Categories have completely different data sets and input information. Therefore, such recommendations must be tailored for this specific purpose – this was the first time we encountered such a client requirement.

Imagine coming to your favorite e-shop homepage and seeing recommendations for specific categories rather than for products similar or complementary to those you’ve already purchased. Great, right? Can’t imagine it? See the picture below.

Of course, Muziker also required recommenders at the category and some segment levels, such as And not just one, but several right away. So, the recommendations need to load and display really quickly to minimize the time the user waits for the page to load. After all, we all struggle with website loading speeds in the online space. Since we fulfilled the assignment and added our invention, we won Muziker’s hackathon.

Multiple types of recommended items according to category

Implementation of the Recommender

After the successful hackathon, we were delighted with the convincing results. Even more so when we learned that Muziker wanted to incorporate our implementation into their roadmap – that is always a serious matter. It does not mean any attempt or test. Both parties realize that while they have sufficient room to prepare, there is also a final implementation deadline and everything needs to “work like clockwork.” Based on our successful past cooperation, we already had an idea of how Muziker works and what their demands are, which we certainly used to the fullest.

The client had a clear need to implement personalized product recommendations in their e-shop according to specific requirements and was also motivated to achieve this as quickly and efficiently as possible.

From implementation to deployment on the web, it took a few weeks. It was preceded by further testing and fine-tuning. It was necessary to “clean up” the code so that everything was done properly and, of course, it was sustainable and easily applicable to other places in the long run.

Strong support and quick assistance during development

At Luigi’s Box, we value our clients. Our role is to create an environment for each client that encourages them to innovate and improve the online customer experience continuously. Cooperation with the client involves fast and flexible deployment of our solutions, allowing us to meet their changing requirements. Strong support and quick response during development are key factors that allow us to build long-term, prosperous partnerships.

As Muziker’s CTO, Ondrej Proksa, added: “The Luigi’s Box team provided strong support and agile assistance during development, ensuring that personalization in recommendations would fully meet our needs as well as those of our customers.”

Custom integrations and personalization on websites

Sometimes, it is not easy at all. However, communication is always focused on maximum cooperation in identifying and resolving any challenges that may arise during the development and implementation of our tools.

In addition to technical support and development assistance, we provide customers, as was the case with Muziker, with expert advice and recommendations regarding best practices for implementing our products.

Category recommender is an interesting specialty

As we wrote above, one of the special recommenders that is not very common is the Category Recommender, which is unique in that what it recommends are not products or accessories but categories. You might think there is not much we can do on the homepage, but you would be surprised.

The more and more frequently a customer shops, the more precise the predictions tailored directly for each customer become, and this applies not only to products. Looking at the results, we find that the recommender brings a lot of interactions. For example, there were hundreds of thousands of impressions, tens of thousands of clicks into recommendations, and many conversions in recent months. When switching to assisted conversions, there are nearly a thousand times more. Simply multiply by the average order value, and it is clear that it would be a mistake not to use such a recommender.

What about A/B testing?

A/B testing is a simple effective tool that allows the client to compare two websites, text, application or image variants to determine which is more suitable for users. It involves testing two variants (A and B) and deciding which one is more productive based on evaluating the data. In our case, it is about two variants of recommenders.

Let’s take a look. During use, Muziker runs several concurrent A/B tests. It is important to continuously test because, as a client, you have an overview of the current condition of the recommenders and can subsequently improve it. Let’s look at one summer test after deployment this year.

The testing took place on the MUZIKER.CZ domain. In this particular case, two recommenders were tested and evaluated: Basket Popup and Item Detail. We compared the original versions of the recommenders with the Luigi’s Box versions.

Visually, these are the two recommenders:

Luigi's Box popup basket

Comparison with the performance of existing recommenders

The Basket Popup saw an increase in all metrics:

  • The number of recommender impressions was higher by almost 62%.
  • The number of clicks into the recommender was higher by more than 100%.
  • The number of users who clicked into the recommender was higher by 93%.
  • The number of conversions from the recommender was higher by 50%.

The Item Detail recommender saw similar results:

  • The number of recommender impressions was higher by 60%.
  • The number of clicks into the recommender was 1.47 times higher (or 147%if expressed as a percentage).
  • The number of users who clicked into the recommender was 1.2 times higher(or 120% if expressed as a percentage).
  • The number of conversions from the recommender was 1.1 times higher (or 110% if expressed as a percentage).
Basket Popup (%)
Item Detail (%)
Metric Impressions
Basket Popup (%) 62
Item Detail (%) 60
Metric Clicks
Basket Popup (%) 100
Item Detail (%) 147
Metric Users Clicking
Basket Popup (%) 93
Item Detail (%) 120
Metric Conversions
Basket Popup (%) 50
Item Detail (%) 110

The A/B testing results for our recommenders were truly remarkable and confirmed the effectiveness of the Luigi’s Box Recommender. In both cases, the difference was significant, giving us confidence that we were on the right track to optimizing the customer experience in online stores and increasing sales.

Long-term cooperation

We also use Muziker data in our demo store

Thanks to our cooperation with Muziker, we gained access to their product catalogs, which we also use in our demo store by agreement. Thus, we do not have to create sample data from an imagined structure but can work with existing structured data, which simplifies both the testing and presentation process.

Muziker data in Luigi's Box demoshop

Development of recommender features usable for LB clients

One of our priorities is the continuous development and improvement of our recommender functions. In cooperation with Muziker, we have developed and implemented various recommender features that we could then offer to other clients. These features include personalized recommendations, complementary recommendations, cross-selling recommendations, and more. Thanks to working with Muziker, we had the opportunity to test and verify the effectiveness of these features on their platform before implementing them for other Luigi’s Box customers. In this way, we can provide our clients with a high-performing recommender.


However we look at cooperation, we see our clients primarily as partners. And partnership brings benefits to both parties. As clearly evidenced by this case study, personalization of content and product recommendations to customers are essential tools for online stores. Muziker recognized their importance and collaborated with Luigi’s Box to create the ideal recommender. The data shows that such recommendations bring significant interactions and conversions, increasing customer satisfaction and the store’s turnover. As mentioned, A/B testing is then a key tool to verify the effectiveness of these recommendations.

The importance of personalized recommendations is clear: strengthening customer loyalty, increasing brand fidelity, and ultimately boosting the store’s profits. Muziker and Luigi’s Box have set a new standard for relationships that could inspire many other businesses.

What about you? Are you already using our Luigi’s Box Recommender?

How our clients benefit from Luigi's Box

Real Dutch Food: 48% Decrease in “No Search Results”

Real Dutch Food lowered the “no search results” by 48%, so the customers won’t be hungry for relevant results anymore.

Read Case Study

Tailoring Autocomplete for Zoot Brought a Fashionable Conversion Rate Increase of 2.5%

Zoot is an online fashion store and a part of the biggest Czech fashion house, Digital People. Along with its sub-brand, Bibloo it is present in 11 European countries, offering more than 1 million products from 300 fashion brands.

Read Case Study

Gandalf: +9.40% Uplift in AOV From Search is a Polish e-shop with over 300,000 books, audiobooks, e-books, music, movies, games, and more.

Read Case Study

Pilulka: Average Cart Value Increased by 4.21%

Pilulka is one of the largest e-shops in CEE in the field of health and medicines, focusing on high-quality and fast delivery of goods with a turnover of approximately 3 billion CZK (118 035 883 EUR) in 2023.

Read Case Study

The “Add to Cart” Rate for Grew by 5.25% In Just One Month is a Polish flooring and furniture seller with a long presence in the Polish market that decided to get rid of missed opportunities caused by ineffective website search.

Read Case Study