

We've already described the benefits of autosuggestions which are invisible and indispensable assistants on any website. Today we would like to share analytics from a large online store which implemented a new Multisearch feature, namely smart automatic tips, in May this year. We will compare conversion rates before and after the implementation so that you can see on the concrete example how effective this feature is.
This case is related to the retail chain EVA. It is a leader on the Ukrainian market of household cleaning, perfumery, and home care products with the emphasis on active turnover and low markup. EVA is a five-time “Retail Awards” winner and the top chain in the drogerie segment by revenue, number of stores, and turnover.
Online store EVA.UA offers a huge range of goods. With such a large assortment, popularity of the store, and high expectations from users, search on the site should work 100%. Before Multisearch installation, customers of e-shop had typical problems with search.
In addition to basic functionality EVA.UA web store was provided with additional functions. They include dynamic filters, ID's, direct clicks, recommended products, popularity of products, search by specifications, and autosuggestions. Each Multisearch feature extensively increases quality search quality and boosts sales. That's why as soon as our developers rolled out new functionality, namely smart Autosuggestions, the team of EVA.UA immediately implemented it in their online store. There is no need to use basic autosuggestions when you can get the smart ones.
Basic And Smart Autotips: What Is The Difference?
Autosuggestions are the main assistants in any online store. They save time, guess our desires from the first letter, and immediately direct to the desired product. Hints are formed instantly and one doesn't even have to press enter. When a user starts typing in a phrase the search engine immediately suggests further search terms, categories, brands, and products.
Autosuggestions are not identical with search results. While hints help to quickly get to the required positions, search results offer all products that match exact information provided by a user.
The illustration below allows us to see the difference. The left part demonstrates search autosuggestions and the right one search results.
Multisearch autotips are an indispensable basic functionality for any online store. Hints are formed taking into account products names, categories, catalog structure. They perfectly help a user to navigate in the assortment and not just guess search endings. The store has the ability to influence the tips and to set them taking into account seasonality, supplier requirements, promotions, and sales. If your store has autosuggestions, that's already great!
The advantages of smart tips:
-
they are formed with behavioral factor, product popularity, and query conversion rates in mind;
-
they offer the most relevant suggestions for each consumer request;
-
they constantly learn from conversions through machine-learning algorithms;
-
they take into account seasonal factors and not only standard "spring-summer-fall-winter" ones. The algorithms take into account products/requests/trends that gain popularity over short periods of time as well as those that disappear from search as demand fades.
Smart autosuggestions make the service as personalized as possible and literally guess customers' desires. This is what all the world's top ecommerce companies strive for.
After the upgrade, EVA.UA online store was updated with request trackers, transitions, and transactions, so the prompts became more relevant. The table below allows to compare the autosuggestions before and after the upgrade.

The hints on the right are more precise and they match the customers' needs as much as possible. They are formed taking into account information on the popularity of products and their conversion rate.
Thanks to Multisearch internal analytics, you can analyze the most converted requests, products, and categories.

According to our statistics, autosuggestions in online stores are used by about 15% of clients who go to the search bar. As smart autosuggestions on EVA.UA site are more relevant, they are used by 25% of customers.
The most popular are query hints which continue a query with keywords. It's a convenient type of tip which saves a lot of time because one doesn't have to fill in the queries to the end. In EVA.UA online store, 75% of clients use query tips.

It's easy to explain why they are so popular. Usually during a search we use general queries such as "soap" and expect to see an extension of the key phrase "liquid soap" or "laundry soap" in the suggestions. This helps to navigate through the assortment and move quicker to the right position. We are used to a similar way of searching for products on the world's top ecommerce sites such as Amazon, Walmart, and Alibaba.
The second most popular type of tip are category suggestions. They are used by 16.6% of clients. From the analytics report we see, for example, that in one month more than 17,000 people went to the "hair care products" section thanks to the category tips.

Comparison Of Basis And Smart Autosuggestions
To compare, we looked at two periods and the top 50 popular requests from autosuggestions. The first period is a timeframe when basic autosuggestions were used (in orange). The second period is the time when smart autosuggestions were applied (in blue).
We excluded direct clicks from the advertising traffic, which is the responsibility of the contextual advertising specialist, and only considered requests from the search bar of the site.

We see a slight decrease in the number of search queries. But the number of clicks on the top 50 smart autosuggestions increased by 35%. Income also increased by 38%, and conversion rate grew by 1.3% .

Integration of Multisearch and smart autosuggestions had a positive impact on EVA.UA sales figures.
Conclusions
For large stores with a large assortment, it's critical to have a tool which allows to find products quickly and accurately. Smart autosuggestions are based on machine learning and take into account behavioral factors, popularity of products, and are trained on conversions. This greatly increases the relevance of query extensions. Smart tips in EVA.UA store improved user experience, increased effectiveness of basic autosuggestions by 38%, and increased sales by 7%.
Multisearch has many more features to boost sales. Test Multisearch for free and check out how efficient it is. Just click the "Get Started" button at the bottom of the page and the 14-day trial is yours.