Full-text search optimization for the magistra.cz e-shop

News release date: 3. 6. 2021
The last year was a record one from the point of view of e-commerce. After the movement of people was restricted due to government regulations, most sales moved online. In Europe alone, the year-on-year increase in online sales was almost 50%.
“If full text search is set incorrectly, up to half of search queries may end up without a result, which leads to visitors leaving the e-shop.”

Intuitive use and a positive user experience are important for every e-shop. However, it’s not just about well-known things such as a suitable category structure, a well-arranged product detail or a shopping cart journey. One of the widely used functionalities is full-text search. With the right settings, it will help users quickly find the product they are looking for or direct them to the appropriate category. If set incorrectly, up to half of all search queries may end up without a result, which most often leads to visitors leaving the e-shop.

In the spring of 2021, we optimized the possibilities of full-text search in the e-shop of MAGISTRA.cz pharmacies using fuzzy search, i.e. approximate string matching. It is a search algorithm that searches a website or e-shop for pages that are likely to be relevant to the search query, even if the query does not return any exact match. Exact and highly relevant matches are displayed at the top of the list. When optimizing the search, we focused on solving the most common causes of unsuccessful search, such as typos, spelling mistakes, Czechization of foreign names or missing diacritics. For example, by optimizing full text search, we solved the problem of finding fewer than 3 characters. These are specific cases, especially in the pharmaceutical assortment, where the names of vitamins often contain fewer than three characters (vitamin B2). Now the search works from one character and autocomplete from two. 

We performed A/B testing in the e-shop to measure the effect of modifications on search results and autocomplete predictions before and after optimizations. The results of the A/B test showed that the number of journeys from autocomplete increased by 55.23% and the target conversion rate by 13.04%. Autocomplete predictions were displayed to search users 38.64% more often than before optimization. The overall conversion rate of the e-shop increased by 21.13%.

We will guide you through the process of creating a successful e-shop.

Do you have any questions? contact us.
The file was successfully uploaded