
Search on a website can be quite the magician. Otherwise, how can you explain the result of “washing machines” when you search for the word “boilers”?
The average shopper doesn’t understand this kind of magic, and they’re unlikely to browse through the catalog to continue searching for the boiler under the term “water heater.”
The key is to “de-charm” the system — teach the search engine to think like a user.

A smart search engine quickly understands that under the synonym “boiler,” users might be looking for a water heater. So it will display a list of water heaters right from the start. This way, the user can find what they need without wasting a second.
Below the product list, the service for installation is also suggested — immediately tackling the issue head-on. This allows the shopper to address their household needs comprehensively, while the store can implement cross-selling.
We were curious how the smart search from Multisearch helped achieve this, so we asked Citrus about it.
Search Bugs Before Multisearch Integration
Citrus employees tried to configure a search system that would find products by matching the first 3 characters. However, the search results were incorrect even for basic queries. It became clear that the search algorithms required fundamental improvements.
Visitors to ctrs.com.ua could not find products because of several issues:
1. No Results Displayed: The search didn’t show results. Users could not see a product on the site not only if it was out of stock but also when it was available for purchase.
2. Irrelevant Results: Relevant queries were returning irrelevant results. For example, a search for household appliances (dryers, vacuum cleaners, irons) could display digital products (laptops, tablets, smartphones).
3. Search Failure for Modifications: The search failed when users copied product names from Google or other price aggregators but couldn’t find the product by its full name.

When copying a product name from Google and pasting it into the search bar on ctrs.com.ua, the user didn’t see the specific model (the part with numbers and letters) in the search results.
There was not enough time to improve the algorithms in-house. So, the task of setting up the search was delegated to the Multisearch specialists.
With an omnichannel business model, it's crucial for the search to work without delays, find products in the catalog, and support cross-selling. Additionally, it’s always beneficial to have technologies that improve conversion rates and NPS.
Features Updated in the Search
Every feature in the smart search helps streamline the customer's journey to a product and assists the online store in making sales more effectively.
The basic functionality allowed site visitors to find products that previously showed up as "not found." The search by modification was also fixed — it now shows complete product names rather than fragments.
Smart Auto-suggestions helped save time in forming queries and finding products. Now the search bar understands the user from the first letter and formulates words on its own. The system offers the user potential search options.
These auto-suggestions have taught the search system to analyze conversions. It now not only shows product names but also suggests the names of the most popular items right during the query creation stage.
The Product Popularity is shown to the user at the top of the conversion assortment, not at the end of the search results. This boosts trust in the product: people are more likely to buy a product that stays at the top. Especially when this "top" is formed by real customers, not the store. People literally vote with their wallets for the top spots when purchasing products.
Direct transitions guide the customer not to the results page but directly to the catalog. This way, the user gets a list of items that are closest to their initial query. When the customer can’t find a specific product, they are guaranteed to get alternatives to choose from.
Dynamic Filters allow users to narrow down the results from 4,000 items to 10 or fewer. Sorting products by basic filters such as category, brand, and price simplifies navigation. However, the results never reduce to zero. Users maintain their purchasing interest if they aren’t drained by a long and exhaustive search.

Feedback After 2 Years
The Citrus team confirms: there are no bugs. And that means the testers are getting bored (smiley) Specialists regularly check key nodes and applications. They also manually review all possible combinations of user queries.
For example, let's consider the cold season. Imagine a customer needs generators and they come to the site. It's important that they can find the product simply by typing the word "generator" in the search bar.

When the user receives a selection of items with key parameters in response to their query, the search is functioning correctly.
If any issues arise, Multisearch technical support is always available to assist with search adjustments or explain the workings of a new feature.
Collaboration with Multisearch
Citrus company is actively working on organic growth, improving both the infrastructure and software components, and placing significant emphasis on UI/UX. To maintain a high service quality, search configuration continues to be managed by the Multisearch team.
We asked how Citrus evaluates this collaboration and what their initial expectations were.
How did Citrus specialists view the work?
Our company has managed to become a reliable support system for the Citrus web development team, and together we’ve made the shopping process easier for users on ctrs.com.ua.
Advice for companies to improve search functionality
❗ Citrus experts emphasize: search is a tool that is used and shouldn’t be neglected.
What should a business do to improve search algorithms and increase sales efficiency?
*Popular products — this is a feature for training on conversion data. It takes into account not just sales, views, and clicks, but also seasonality and novelty.
How does this feature work?
Popular products influence the user almost like native advertising, which significantly boosts the mood of marketers.
For example, the Iphone 15. It was released at the end of September 2023, but by mid-October 2023, one of the models was already leading the search results on ctrs.com.ua for the query "Iphone."
It holds a respectable 5th place among similar results on the online store.
This example with the iPhone illustrates the birth of a trend, and the main result will be visible in November-December, when the Iphone 15 model will reach the top position in search results.
Main Insight on Smart Search
With smart search, Citrus specialists no longer get distracted by fixing errors. The store team now spends more time and effort on the creative development of the mobile app and enhancing the product cart.
Some ideas are still being implemented together with Multisearch:
We regularly discuss the superpowers of search algorithms in our blog.
Want to experience the power of smart search from Multisearch? Test its capabilities by signing up for a free 14-day trial.