What is federated search ?

Federated search enables a user to search several different data sources at once by making a single query. The federator gathers results from one or more search engines and then presents all of the results in a single user interface.

Different types of federated search solutions have their own advantages and disadvantages. But from the user’s perspective, it appears as a single search that pulls from all available sources.

Why is federated search strategically important to your company?

Federated search makes it faster and easier for your customers and business users to find what they’re looking for. It gives users the ability to search from one single location and pull results from all available sources. That makes your content and data more browsable and useful.

It helps customers find products or services with fewer clicks or page views. That improves click through and conversion rates and makes them more likely to engage with your brand.

Federated search also helps business users find the information they need with less time and effort. That improves the user search experience, increases discoverability, and boosts efficiency.

Federated search is especially valuable for complex organizations with multiple data sources, whether they are on-site or in the cloud.

What are examples of federated search?

Federated search solutions are powerful tools for ecommerce. Users can quickly search through thousands of products, guides, reviews, and other online information to find answers. That leads to higher conversion rates and increased sales.

Likewise, federated search solutions add tremendous value at the enterprise level. Corporations typically have separate websites for customers, recruitment, investor relations, brand awareness, social responsibility, and so on. Without federated search, any user who lands on the wrong website may not find the information they’re looking for. They are more likely to leave, and that represents a lost opportunity.

Federated search allows users to search through all available content simultaneously. It helps them find exactly what they’re looking for, no matter where it resides. That provides them with a better experience and increases their engagement with the brand.

What are the different kinds of federated search?

There are two distinct types of federated search: search-time merging and index-time merging. These two approaches are very different, both technically and conceptually. That’s why each approach has its own set of advantages and disadvantages.

What is search-time merging?

Search-time merging is sometimes called query-time merging. In this type of federated search, each individual data source uses a separate search engine. The federator gets the query from the user and sends it out to all of the search engines at once.

Then the federator waits to hear back from those individual search engines. Once it has gathered all of the results, it aggregates them into a single list and presents it to the user.

What are the advantages of search-time merging?

Search-time merging is more straightforward to implement. Since it searches each index separately, it can handle data in different formats. It doesn’t require any standardization in the data.

Also, there’s no need to build a unified index to aggregate all of your data. That’s because the query federator taps into the existing search indexing systems of each separate data source.

What are the disadvantages of search-time merging?

From the user’s perspective, the biggest drawback to search-time merging is that the response time is often slower. The federator can’t deliver the final results until all of the individual search engines respond. That means its speed is limited to that of the slowest search engine. Results may not come fast enough to meet the real-time expectations of today’s users.

Ranking the aggregated results also presents a challenge. That’s because each individual search engine will score relevancy differently.

One solution is to present the results of each search engine separately to the user, such as under separate tabs. Another solution is to sort them by deterministic data such as price, date, or location. But that may not provide the best user experience.

Search-time merging doesn’t require the creation of a unified index. But you must still operate and maintain search tools for each individual data source. That rules out the use of any data that lacks its own search tool.

What is index-time merging?

Index-time merging is a completely different approach to federated search. It doesn’t use separate search tools for each data source. Instead, it utilizes a single, unified index of all searchable data.

This type of federated search system requires building a large central index. For that reason, there is a higher initial investment time. The efficiencies come after you have acquired all the content into your central index. From that point on, users can search the data faster and get more relevant results.

Index-time merging also allows you to tap into all of your data sources, even those that aren’t supported by local search tools.

What are the advantages of index-time merging?

Because this type of federated search relies on one central index, it doesn’t need to wait on local search tools to respond. It usually returns results faster.

Index-time merging also opens up the possibility of using any data or content that lacks its own search engine. That allows organizations to utilize a much wider range of data. Practically anything can be brought into the central index.

Perhaps the biggest advantage is that it provides a better user experience. The central index allows the use of sophisticated query enhancement and relevancy algorithms. Users get excellent search results ranked by relevancy.

Index-time merging also allows for the use of filtering, auto-complete, and other useful features.

What are the disadvantages of index-time merging?

The only real drawback to index-time merging is the higher up-front effort to create the index. That can lead to a longer implementation time, since there is more initial complexity.

Creating the index doesn’t have to involve moving or changing the content of different repositories. But the index can be more complicated to maintain over time, since it must reread each item every time a change occurs.