Skip to main content

What is federated search?

 

If you run a site or business, you know how quickly new databases, storage locations, and product catalogs can build up. With this constant accumulation, how can you ensure your customers always find what they’re looking for on your site? How can you also ensure this solution is simple for your business?

Federated search is one way to address both of these needs. By allowing users to search many data sources, and therefore multiple types of content, at the same time, federated search improves user experience and engagement. Meanwhile, it can make it easier for your business to manage data and search tools.

 

What is federated search?

Federated search is a technique used to search multiple data sources at once. With federated search, you can retrieve information from many different content locations with just one query and one search interface.

You may have encountered federated search without realizing it. When you type a search query into MacOS Spotlight or Windows Search, the search engine returns results of all kinds such as apps, webpages, contacts, and documents which are drawn from different sources.  The search engine uses different layouts to best present each type of content returned.. This is federated search in action.

Federated search can be used in many contexts. For example, if your company maintains separate documentation databases for different products in addition to your product catalog, a federated search tool would allow your customers to search from a single location and obtain results from all of the documentation and the product catalog simultaneously. This type of enhancement goes a long way to improving the customer experience on your site.

 

The Importance of Federated Search

Although users may not access your content through search alone, implementing federated search on your website or application has several key benefits:

 

Improves Customer Experience 

Federated search is an efficient option for mid-to-low funnel users who know exactly what they need. They can search through a large body of data from one location with one query, thus reaching their goal with fewer clicks. Even users who are still new to the topic benefit from federated search—by searching for one keyword or phrase, they can get a wide range of content from your product pages, documentation, multimedia assets, and more.

The fewer clicks or pageviews required to find a product or service, the more likely you are to convert the user. When users find key information faster as a result of federated search, you see improved click-through and conversion rates

 

Makes Website Expansion Manageable

Whenever you add new content or data locations, you can easily integrate them into your existing federated search tool, instead of having to set up an independent search tool along with each new type of content. 

 

Supports Browsability

With a centralized, federated search solution, you can modify, add, or restructure data easily while still keeping that data searchable.

You can adapt each part of your federated search user interface to perfectly showcase the content it returns, to help users interpret and navigate different categories of information more easily. This improves the browsing experience on your site, promotes discoverability, and increases user engagement

Birchbox federated site search

 

Improves Reliability and Security

Federated search requires you to manage only one search engine, which simplifies reliability and security. Monitoring, maintaining and securing a single search engine is easier to manage and troubleshoot than doing so for different search tools for each data set.

 

Increases Relevance of Search Results

Federated search allows you to optimize relevance for each type of content you surface to your users. You can take into account different parameters to rank your different types of content, rather than a one-size-fits-all ranking approach.

With a larger volume of information to draw from, search results are often more accurate and more relevant. And when users interact with these improved results, they generate valuable user intent data. Once you understand what users are searching for (and how), you can guide users to more helpful content and improve the searcher’s experience. 

 

Approaches to federated search

All federated search solutions rely on two fundamental components:

  • Index: a compilation of the data that you want to search, structured in a way that facilitates efficient searches. 
  • Search function: the part that parses the index to find relevant information within it in response to a given query.

The index and the search components can interact in different ways to achieve a federated search. The main approaches include search-time merging, index-time merging and using a federated search interface. 

 

Search-time merging

With this approach, your federated search solution runs separate searches on each data location that you want to include in search results, using multiple indices. Then, it aggregates the results from each of these searches into a final list, which is presented to the user. 

Search time merging schema

 

Merging the results of the various data sources at search time is typically the simplest type of federated search to implement because it does not require you to aggregate all data into a single index. However, it does require you to run and maintain specific search tools for each data source that you want to include in your searches or to use a single tool that can handle all types of content, but ingest each data source in a different index. Search results may also be slower to arrive because the central search engine has to wait until all of the local search engines have responded before it can deliver the final results. Finally, fine-tuning the relevance for the aggregated results list can be very challenging, as the search engine may have trouble ranking highly different forms of data.

 

Index-time merging

Index-time merging involves building a central index of all of the data that you want to include in search results, then searching that index to perform a federated search.

Index-time merging schema

 

This approach requires only one search engine and one index, and it is compatible with data sources that don’t have local search tools available to support them. It also typically generates results more quickly because there is no need to wait on local search tools to respond to a query.

On the other hand, index-time merging is more complicated to set up and maintain. You must find a way to aggregate data from multiple locations into a single index, which is particularly difficult if not all data sources exist in the same format (for example, some might be PDF files, while others are HTML pages). Once data is fed into a single index, index-time merging still requires you to decide on a unique relevance strategy for all your different types of content, which is a very complex, if not impossible problem to solve.

 

The Federated Search Interface

This approach to federated search is an extension of the search-time merging method. However, instead of aggregating the results in one combined result list, it presents one result list for each type of content the search is performed on in a unified interface.

The Federated Search Interface schema

 

This method requires a robust search solution equipped with the ability to index different types of content in different indices and create the unified federated search interface. It also requires site owners to give forethought to the final experience they want for users so content can be indexed and delivered in the most friendly way.

A federated search interface allows you to fine-tune the relevance for each type of content independently, thus providing a superior experience to your users.

 

Federated Search Examples

Implementing federated search well delivers real results for businesses across a range of industries by improving user experience, increasing user engagement and boosting conversion rates. 

 

Retail

A successful e-commerce site relies on advanced site search design to direct visitors to what they need. A typical site might list thousands of products, each sorted into different categories, accompanied by their guides and reviews. If a visitor doesn’t know which category a certain product falls within, they may become overwhelmed and leave the site. 

Take the simple case of a customer searching for a bath mat. Where should they start? Should the customer look in the bath section, the linen section, or look for education material to buy the best one for them? The categorization could easily differ from site to site.

With federated search, finding the right place to perform the search doesn’t matter. A simple query for “bath mat” in the search bar will search all product categories, leading the customer quickly to the item and its related content thus maximizing the likelihood of a successful sale.

 

Enterprise

Large corporations often have numerous websites serving various purposes and stakeholders. Separate sites may be maintained for investors relations, hiring, brand awareness, and corporate social responsibility, to name a few.

Without federated search, visitors may land on the main corporate website and have trouble targeted resource or information on the relevant site. It can represent a large array of lost opportunities for the company.

With federated search, however, the visitor can simply enter a keyword and search through all contents at once. This approach is likely to lead the visitor to what they need faster, providing a more positive experience and increasing the chances that the visitor will keep interacting with the company.

 

Software vendors

When customers are purchasing software, they already have many crucial factors to consider. Customers expect elegant, fast solutions, and a clunky search experience can turn them away. That’s why integrating a seamless federated search experience into websites is critical for software vendors.

Consider this federated search feature. Not only does the tool allow users to search the entire website from products to blogs to documentation with a single query, but search results also appear in real time, as users type. Results are categorized into different groups in order to make them easy to interpret. Pictures and short descriptions to illustrate the different types of content are added, when applicable, to make the content discovery and navigation more accessible to users. All of these features help enable a positive end-user experience.

Example of a realtime autocomplete

 

Getting Started with Federated Search

From a technical standpoint, implementing federated search can be complicated if you attempt to build a search engine yourself—especially if you want to maximize performance by using an index-time merging architecture. 

With Algolia, integrating federated search into your website or mobile app is fast and seamless. Algolia delivers lightning fast search results, no matter how many data sources you use, giving your customers what they want faster than ever. 

Watch our demo to see how Algolia’s federated search enabled solution is helping businesses around the world satisfy users.