Case study banner

Castorama Conversions Grow By 24% From Site Search

About Castorama Polska

Since 1997, Castorama has been a recognizable brand in the Polish market. Nowadays, they are the country’s largest company in the home improvement industry. As a chain with around 80 stores in 60 cities all across the country as of 2020, Castorama Poland employs almost 12,000 people. They support their extensive network of physical locations with a well-developed online store, which attracts almost 12 million unique visitors each month.

The company is an integral part of the Kingfisher group of companies, a community of over 75,000 people from 130 countries. They believe that everyone should have a home where they feel good. Together, they enable others to improve housing conditions and help people change their houses and apartments to the places they love.

Results From Site Search

Immediate Conversion Growth

    • Back in 2018, after effortless and fast implementation of SearchNode, Castorama Polska observed how their conversion rates from sessions with search skyrocketed by 24%, compared to their prior solution.

    • Since then, the results only kept growing, followed by the 30% growth of CTR from site search one year after the solution was launched live.

    • The company continues enjoying the fruits of advanced site search and continually sees the growing results, which is made possible due to the continuous search improvements process utilized by SearchNode.

Pandemic Related Growth In Results

    • In 2020 Castorama.pl saw even better results from their search. Due to the COVID-19 pandemic, they had to close brick-and-mortar locations, which led to +42% YOY growth in sessions, while sessions with site search grew by +25% YOY.

    • As SearchNode provides scalable solutions and reacts fast to any changes that occur, it was extremely helpful for the company to deal with the increased traffic load and ease visitors’ discovery process.  Sessions with site search brought the company +186% YOY conversion rates growth and +211% YOY revenue growth.

Challenges

Irrelevant Search Results
Castorama.pl struggled with providing accurate search results; correspondingly, search usage and conversions were low, and their clients were often complaining about the poor site search experience.

Lack of Automation
The prior solution of the company did not allow them to improve their search on the scalable algorithmic way, thus requiring tons of tedious manual work and hardcoding.

Failed Utilising Omnichannel
Castorama was challenged by the “Research Online, Purchase Offline” (ROPO) trend, which requires bringing and processing data across all offline and online channels in real-time.

Looking for solution

Castorama.pl was in the stage of rapid growth, accelerated by the overall expansion of the Polish eCommerce market, which made search problems more and more evident; so the company started to look for an effective cure for their pains.

Based on the recommendations of the KingFisher group, they were evaluating various different site search providers from huge enterprise search solutions and open source technologies to SaaS providers, and a hybrid between SaaS and consulting partner – SearchNode.

Eventually, advanced SearchNode search product, backed up with continuous consulting and exceptional experience in providing search for home improvement eCommerce stores, won over the competitors.

“We chose SearchNode because of their commitment to accuracy and ongoing improvements, as well as the fact that their solution is built for e-commerce platforms. They were able to work with us on properly optimizing for Polish language nuances, spelling and grammar specifics.”


Tomasz Stefanowski,
Web Development Manager at Castorama

The implementation of SearchNode

The implementation of SearchNode for Castorama.pl was made in such a way, what the company would benefit most, having to dedicate only minimum development effort.

The main task of the Castorama Polska search team was to communicate and discuss with SearchNode goals and features of the site search to make sure that it fits business objectives and priorities, and to provide feedback during the implementation process. From the technical point of view, they only needed to provide data export API according to our requirements. SearchNode did all the rest – from data analysis/optimization and tracking scripts to the creation of search user interface and custom AI algorithms.

The full cutting edge search solution was launched in seven weeks from the initial kick-off workshop. Castorama.pl conversion rates from users, who use site search instantly rose by 24%. Our search experts keep working together with Castorama Polska as an integral part of their search team, operating as if the company would have internal search engineers, to ensure even better results.

Relevant search results

Data Relevance Algorithms

SearchNode has a powerful AI-based and eCommerce optimized search engine at its core, which provides relevant results out of the box. However, to reach the next level of search relevance, which is essential when there are millions of search queries, we needed to go above one-size-fits-all technology. Our engineers analyzed Castorama product data and user behavior to create an additional layer of search algorithms for the company on top of the SearchNode AI search platform.

As we continue working together with Castorama.pl on search improvements, our search engineers develop new iterations of these algorithms specific to Castorama, thus improving their site search in a scalable way. This combination of AI-based search platform, fortified with custom algorithms, maximized Castorama.pl search relevance, and helped to achieve their business goals.

Optimization of Search Queries In Long-Tail

Before SearchNode Castorama.pl used to put their search improvement efforts to the most frequent queries, which accommodated only for a small number of unique search phrases (~20%), meaning that the majority of queries were hidden in the long-tail, and was not optimized at all. SearchNode implemented machine-learning algorithms, which automatically optimize all user queries, not only the most common ones. This allowed Castorama to maximize its revenue potential for all users’ queries, no matter how rare or complex they were.

Polish Language Nuances

As Castorama Polska is focused on the local market, its site search must comprehend Polish language perfectly, despite its complexity. By employing natural language processing (NLP), SearchNode built custom Polish language algorithms and rules, which helped to better understand the context of user queries. Also, custom stemming made sure that any forms of the word or use cases are understood, based on the language grammar. This helped to lift Castorama.pl search relevance significantly, resulting in improved conversion rates.

Contextual Filters

Contextual filters was an essential feature for Castorama, as selling homeware and garden goods requires a lot of context. We enhanced search results with dynamic facets and filters to improve user experience. What is more, we implemented advanced AI automation algorithms, which can analyze the context of each search query and select only relevant filters and facets. This provided Castorama.pl users with more flexibility, and notably boosted the probability of a sale.

Ultrafast Search Experience

Castorama search UI was designed and constructed by experienced SearchNode front-end engineers, including search result pages, filters navigation, and autocomplete, having mobile-first approach in mind. We applied best practices for a high-speed search experience; thus, it ensures that the search query is analyzed and results are displayed in a split second. The same applies to filtering – results are refreshed based on filters instantaneously. This allows Castorama.pl users to enjoy fast discovery effortlessly.

Resourceful Autocomplete

Smart queries and categories suggestions

Optimized autocomplete suggestions are the great source of revenue that Castorama was missing out. SearchNode smart query suggestions generate keywords from Castorama.pl product data and make conversion-optimized product suggestions. Also, it displays phrases within product categories, which ease of discovery for the users. This created a highly-converting search experience for the company.

Discovering products through custom content

Castorama.pl is actively using content marketing – they have over 7,000 articles related to DIY projects, ideas, and inspirations. The content they create has purely educational purposes; thus, it is an excellent resource in navigating the shopping and buying processes. SearchNode provides intelligent suggestions in autocomplete, helping to reach users who are not ready to buy yet through this non-product content. This way, Castorama manages to attract, engage, and capture customers early in their user journey.

AI-Powered Smart Merchandising

Custom ranking algorithms

Initial workshop between SearchNode and Castorama teams allowed setting up the custom merchandising solution to serve in the best interest of the company. Based on this, SearchNode created intelligent merchandising algorithms that automatically adjust results to achieve optimal conversion rates. This ensured that Castorama Polska site search is automatically focused on delivering the best results for the company.

Merging online and offline

The omnichannel approach is vital for home improvement industry players, like Castorama, as people often want to assess the real look and feel of the item. Knowing this, we implemented a resourceful solution, which automatically adjusts the search results page, to match the assortment, information, and availability of the goods in the physical store, based on the searchers’ location. Smart autocomplete suggestions also display information about the nearest Castorama brick-and-mortar store. This greatly increased the chance of sale, no matter whether it’s happening online or offline.

Automatically matching seasonality

For Castorama, the change of the seasons influences what people look for in terms of product selection. SearchNode implemented AI ranking algorithms, which continuously tracks users’ behavior and catches trending products early to adjust search ranking and display only the relevant products on top. This significantly reduced manual labor efforts for the Castorama.pl merchandising team.

Autonomy of merchandising team

Besides all automated features, driven by Artificial Intelligence, the Castorama merchandising team still has autonomy over the algorithms – they can manage synonyms and redirects if desired. Merchandisers can also build landing pages on SearchNode search engine and manage it in their CMS dashboard. However, the primary goal of the team is to analyze top, trending, and underperforming search queries from SearchNode analytics.

“SearchNode is a vital part of Castorama.pl user experience and business growth. As a home improvement company, we sell numerous products consisting of complex data, which we were struggling with before. With SearchNode, we were able to switch from a manual and time-consuming solution to a comfortably automated, faster, and conversion-optimized site search engine. This allows us to focus more on other aspects of the business, rather than tirelessly configuring site search. Working with SearchNode is easy, and we don’t need full-time managers to handle search.”


Marcin Jabłoński
Digital Director

Continuous Search Improvements

SearchNode – proactive member of Castorama.pl search team

SearchNode site search improvement process is very similar to having an in-house search development team. We work together with the Castorama Polska team on identifying underperforming search areas and generating ideas for new features and search enhancements. Castorama.pl product owner decides on task priorities and search improvement roadmap. After this, SearchNode engineers implement changes to Castorama.pl search algorithms and develop new features. All of this is added to the tailor-made solution for Castorama.pl (not to the core product of SearchNode); thus, Castorama has full ownership of the process.

Quick results

Since SearchNode search engineers, dedicated for Castorama are independent from the company development team, or global product roadmap, the company can always be sure that any problem, which appeared unexpectedly, will be tackled with speed and ease. This approach allows Castorama to keep on improving their site search for the best possible results, no matter how complex and changing the surface of their e-store is.

Scalable search improvements

SearchNode analytics and reporting system, based on AI, automatically detects potentially problematic queries and displays it to Castorama. Merchandisers from the company review these reports and decide which of the identified issues are essential for their business, and provide feedback to their success manager from SearchNode. After this, our engineers find the root cause of these problems and implement improvements on the algorithmic level.

The described process ensures that a pattern of the issue is fixed, not just an individual case, meaning that the fix is scaled automatically for all analogical problems. The site search solutions implemented by SearchNode allows Castorama to enjoy fruits of well-performing search with minimal effort from their side.

Other case studies

How JOTT doubled conversions by improving its product discovery journey on Shopify
How JOTT doubled conversions by improving its product discovery journey on Shopify

Read more
PELAGIC’s Journey to Increased Sales & Engagement through Nosto’s CX Optimization Services
PELAGIC’s Journey to Increased Sales & Engagement through Nosto’s CX Optimization Services

Read more
Colorful Conversions: How Show Me Your Mumu Boosted AOV by 20% with Nosto’s Product Recommendations
Colorful Conversions: How Show Me Your Mumu Boosted AOV by 20% with Nosto’s Product Recommendations

Read more
Chelsea Peers doubles category page conversions by personalizing to curve and maternity shoppers
Chelsea Peers doubles category page conversions by personalizing to curve and maternity shoppers

Read more