What Is Federated Search?

Have you ever wondered what exactly a "federated search" is and how it can help you? Find out what it means and more with Yext.

min read
Graphic of a Search bar

A federated search looks through the databases and search indexes of multiple websites to come up with a myriad of results.

If you've ever searched for a product on a particular website, only to have a bunch of similar products appear, some of them on the company's partner sites, then you've used a federated search. You just may not have known what the term meant.

Why do websites set up these searches? How can they make finding what you want easier? We'll answer both of those questions and more here.

Why Set Up Federated Search Capabilities?

Although federated searches occur on websites with many stores linked together, they appear more often on private company servers.

Why? Suppose an employee needs to find some specific information about a product that the company sells or even locate it amongst their many stores. In that case, a federated search makes things easier for everyone involved.

An Example of Federated Search

Let's use an example to show how this federated search works. Suppose someone is shopping at a certain store that has many chains located all over the country. Home Depot is a good example.

The shopper wants a specific item that the store they're currently in doesn't have in stock. In that case, an employee can do a federated search on their database to see which local stores do have the item in stock.

They can then order the item for the customer, have the other store set it aside, or, if there are plenty in stock, just send the customer on their way to that other store. Without federated search capabilities, the employee would search all of those databases individually. As you can imagine, this takes up a lot of time and energy and can lead to unhappy customers, as they need to wait while the searches take place.

Thanks to federated searches, this information can appear at the employee's fingertips in seconds.

The same is true on the internet, without the employee getting involved. A customer can look at a certain item, choose the store they want to shop at, and then see if the product is in stock there.

Federated Search Example #2

However, moving away from the store-based example, plenty of other companies and businesses have federated search capabilities built into their employee-only accessed databases.

For example, a company that prepares taxes for clients has a database that allows employees to pull up past tax returns, even if those specific clients went to a different company location the previous year.

If they had their taxes prepared at Location A the year before and Location B the year before and are now at Location C, a federated search can pull up all of their tax returns. It can even pull up those from the previous years that were done at the other locations.

Having this information on hand makes the tax preparation process easier for the current tax advisor or accountant.

Different Types of Federated Searches

There are currently two different types of federated searches.

Index-Time Merging

This type of federated search requires an extensive database that's updated regularly. Setting up a system to handle index-time merging isn't easy and can be quite time-consuming, but the results are worth it.

Instead of sending a search query through several search indexes, only one search index exists, containing every possible bit of information.

This type of setup makes searching very simple and rather quick, but it does take some time to set up and maintain.

Query-Time Merging

On the other hand, query-time merging only kicks in when a query is made to set up databases and search indexes. The query prompts the system to start looking through several different search indexes, which might be housed on various servers belonging to the company. While the search results are fairly accurate, the search process itself does take some time, as all of those indexes must be gone through.

This type of federated search system is less time-consuming and easier to keep up with. Still, as it consists of many different pieces, there's a reliance on individual employees to ensure that the systems are fully updated and ready for searching.

Which is Best?

Some companies prefer an index-time merging setup, while others want a query-time one. It comes down to two things: ease and time of searching and how much time is spent keeping the search indexes updated.

The index-time merging comes out ahead on one of these criteria (search time), while the other, query-time merging, is the winner when it comes to index update times.

As a result, neither of these methods is better than the other. Both are equal. This is why some companies prefer a hybrid approach.

Hybrid Federated Searches

A hybrid federated search is a kind of unofficial category that is on its own. This type of federated search uses a main central index that contains most of the information while also looking through some additional databases that are kept separated.

When a query is made, it goes through all indexes, just as in a query-time merging search. However, since fewer databases are to go through, the search is done a bit faster.

The hybrid approach works well if the main index is where all of the information is stored, with the others containing only a small bit of information. When the result appears, they look like those of a query-time search; however, those results appear quicker than usual.

It sounds like federated searches are the solution to many different problems on paper. However, there are some pros and cons to them, no matter the exact type of federated search that has been set up. Let's go over some of them quickly:

Pros of Federated Search

There are several good things about federated search capabilities, such as:

  1. Better Security

Not only do federated searches pull up information from various indexes (or a single index, depending on the type), but they also track the credentials of the person doing the searching.

This helps information technology professionals see who is looking up which information. It also allows them to stop that person's search if necessary.

  1. Fast and Accurate Results

Since you're looking in a certain computer database for the information you require, you'll receive more accurate results than what a regular Google search returns.

In addition, as the search indexes are regularly updated, the results will be somewhat accurate, as long as those indexes are indeed kept up to date.

Remember that you'll receive better results with an index-time merging system.

Some of the problems with federated search include:

Relevance of Results

Although it sounds as though the results that appear every single time are exactly what the user (searcher) is looking for, this isn't always the case.

Those results might be all over the map with query-time merging, making it hard to discern which of them are relevant.

The odds of getting the results you expect will increase with index-time merging, although this depends on how well-kept the search index is.

Search Index Upkeep Requirements

Search indexes, especially those related to index-time merging functions, are quite time-consuming. These indexes need to be updated regularly since they contain information on several things.

With different databases classifying things in various manners, you have to figure out how to make them uniform enough to search.

Final Thoughts

Although federated searches can be quite useful, especially since they save time by looking through several different databases, they aren't perfect.

The time involved in creating and updating a central search index might be too much to handle, while the imperfect results brought about by query-time searches might make the entire search a waste of time.

With that said, federated searches are good for several things, such as helping employees search various store databases or pull up specific inventory lists when a customer needs a specific item. It all depends on the quality of the databases or search indexes.

Click here for more information on using search to improve your customer's experience.


0 0 0 0 0 0

Share this article

Related Posts

Explore our Content Library