• Homepage
  • Blog
  • Query Suggestions - Synerise New Way of Searching

Query Suggestions - Synerise New Way of Searching

4 min read
Query Suggestions - New way of Searching

Companies continually produce data on their websites by adding new products, blog posts and content. The busier the site becomes, the harder is to find what we are looking for. To meet those expectations, we just need to implement one great tool—properly configured search and Query suggestions.

What are Query suggestions?

'Query suggestions' is a feature of a search engine that allows it to suggest search phrases to the user as they type in the search box. It works as an enhancement of the autocomplete functionality and allows faster discovery of customer intentions as they type.

Benefits from query suggestions

The main benefit of this feature is improving customer experience through easier product search. It is especially useful if you have many of them on your website. Query suggestions can also increase the effectiveness of searching on your blog or any other website which is based on text content.


First of all, if you want to start using queries you have to get content from all pages and subpages on which you would like to use this feature. You should gather all content from those places and pay special attention to every element of the content which can be used later in your search.   

Based on collected content, we are able to prepare catalogs in Synerise. You are able to basically add any information needed, the one obligatory piece of data is Item ID. Of course, your catalogs may look different depending on where the search is used and which elements you want to add. 

How does it work?

If you want to create a new query suggestion - the most important part will be to prepare a ranking of attributes collected in the catalogs. Query suggestions match the search phrase as best as possible with the set of suggestions we create in a few ways. In Synerise you can search based on search statistics, attributes and defined phrases and also use a blacklist. 

Query suggestions settings in synerise

Suggestions based on search statistics

In this setting the phrases suggested to the user are based on the popularity of search phrases. So, the index of the suggestions in this case is built out of popular search phrases. The suggestions will be scored on the similarity to the phrase searched for and returned upon request. The index can be set up to contain only popular searches above some threshold of popularity, for example the number of searches. Also, the time interval for which the statistics shall be taken into account can be set. So, if your search phrase is similar to something that is popular, you can add it in a hint. 

Suggestions based on attributes

It is possible to set up the suggestion index to be built on an attribute from the Items Catalog. It is also possible to add multiple attributes from multiple Items Catalogs. The suggestions will be returned based on the similarity of these attribute values to the search phrase being entered by the customer. 

Suggestions based on defined phrases

Another possibility is to define a list of phrases manually and assign them a score, also manually. This can co-exist with the previous two settings. 

Black list

You can also create a blacklist. It can be created containing specific queries or be based on expressions. The blacklist based on expressions will blacklist any suggestion containing the defined phrase. These phrases will never be returned as a query suggestion. 

Practical use case

Basic search based on existed categories 

In case of e-commerce companies, which have a lot of products and categories on their websites, the best case is to implement query suggestions based on attributes. This can help customers to more easily find products and categories. To do this, we have to start from gathering all data about products which made it possible to create a catalog with all characteristics important for us. Let’s assume that it will be brand and category. 

Adding attributes in query suggestions

In the attribute section you have to just choose the proper attributes. Based on this example a suggestion index will be created based on all “category” and “brand” values from a given item catalogue. 

If a user starts writing a phrase like “jeans”, he can get a list of all subpages with “jeans” in its name like; straight jeans, skinny jeans, loose jeans etc. Also, if he starts writing the name of a category with “new”, he will get brands with “new” which exist on this page like e.g. new look, new balance etc. 

If we mix two or more categories like in this case, you can start typing “skirt” and can get hints like “skirt mohito” or “skirt new look”. 

Advanced search based on popular searches

You do not have to base your query suggestions only on existing categories and product names. You can take into account popular search phrases. 

Stats engine suggestions which are based on the minimum number of queries and phrases

Based on this example, query suggestion indexes will include only those phrases searched by the user which have been searched at least 10 times within the last 10 days, have at least 3 results and at least 3 characters. Other phrases will not be included / returned while querying a suggestion index. 

In this way you are able to easily start to use autocomplete with the most popular phrases that are searched by your customers. 

We can also add manual suggestions and define lists of suggestions returned as enhancement of autocomplete option. And of course, you can add phrases which are not connected with your offer but are often typed by customers (or products which are no longer in your offer) and add them to a blacklist. 


Query suggestions can be extremely helpful in case of ecommerce, blogs or in any other sites where you constantly add new content. It is really important to be user friendly and do your best to help your customers find what they are looking for. If not, it’s very likely that they will leave the site unsatisfied. 

To learn more about query suggestions read our post in documentation >>> 

You will find technical requirements and more useful information. Watch also our video tutorial to learn more about query suggestions!