Printer Friendly

Custom search engines: tools & tips.

Few of us have the resources to build a Google or Yahoo! from scratch. Yet anyone can build a search engine based on a subset of the large search engines' databases. Use Google Custom Search Engine or Yahoo! Search Builder or any of the other similar programs to create a vertical search engine targeting sites of interest to your users.

The basic steps to build one are fairly simple:

* Sign up

* Pick a search engine name

* Choose a list of sites

* Add the sites

* Publish

That quickly, a search engine can be created to search a specific portion of the web, such as local government sites, childcare resources, or historical archives. The process is relatively simple, and the technology is fairly easy to use. With the current range of custom search engine builders, librarians can create one based on Google, Live Search, or Yahoo!. So let's take a quick look at some of the capabilities and limitations of the contenders. Find comparisons, more examples, and details at


1. Google Custom Search Engines: Of all the options, the Google Custom Search Engines (CSEs) are some of the easiest to create and share with others. Setting up a CSE requires a free Google account. It is easy to create a simple customized vertical search engine as well as support much more advanced capabilities (see the Google AJAX search API article). Fill out the form, add the list of sites, and in a few more clicks, a CSE will be up and running. It can be hosted on Google's site, or it can be embedded in a blog or on a webpage. Unfortunately, at this point, the CSEs only work for the web database and not for news, scholar, books, or other Google databases.

2. Yahoo! Search Builder: To build a search engine using a different underlying database, try the Yahoo! Search Builder. It is just as simple to build, but it must be embedded on a website to use. Log in, fill out the form using the "Search the Web" option, and then click "Customize the experience for your users" to add a list of sites. It also includes an option to create a custom search engine based on Yahoo! News. The news version can be restricted to specific Yahoo! News categories as well as to specific news sources.

3. Rollyo: This search engine builder is also based on the Yahoo! database. Roll your own search engine (thus the name, Rollyo) can be hosted directly on the Rollyo site without needing to embed code in another website. Rollyo allows building an engine with up to 25 different sites specified.

4. Swicki: From Eurekster comes Swicki, which, like Rollyo, is also based on the Yahoo! database. Swickis can be hosted by Eurekster or embedded in a webpage. ASwicki can handle up to 50 sites, and it allows the creator to preseed a tag cloud of "hot searches." While a Swicki does give search statistics, the entries in the tag cloud are determined by the creator, whether or not the entries have ever been searched. ASwicki also gives searchers the opportunity to vote for or against a specific result; it then keeps track of the vote. Future searchers will see the votes displayed.


5. Live Search Macros: Microsoft takes a somewhat different approach to using its web database for creating a customized search engine. Called "Search Macros," these search engines must be hosted on the Live Search site and are not easily embedded on other webpages. Like Yahoo! Search Builder, either the web or news databases can be used. Live also gives a Basic as well as an Advanced form for building a custom engine.


Here are several tips to help you create a successful search engine once you have chosen your search engine builder.

1. Selection: While this is not the high-tech aspect of building a specialized search engine, it is the most important. It may also be the most difficult. Choosing a collection of appropriate sites that include some of the most information-rich sites on a particular topic is an exercise in good collection development. To test its effectiveness, be sure to compare the results from your custom engine to results from a general web search.

2. Updating: Once the initial identification identification of appropriate sites is finished, maintaining the search engine is also important. Some sites cease to update their content, change ownership, or just disappear. New sites appear with relevant content. Plan to periodically update the list of sites and re-evaluate the old ones.


3. Balance: One potential usability issue with a customized vertical search engine is that one site may constantly dominate results. If you mix an extremely large site with several small ones, the results may be unbalanced or they may never show any hits from the smaller sites.

4. The Google CSE Conundrum: The idea of limiting results to a specific group of sites and finding everything within those sites that Google has indexed sounds compelling. It should enable deeper searching in those sites. Unfortunately, use of a Google CSE will sometimes find fewer results from a specific site than if you searched directly using Google web search. This can happen in CSEs that have more than three sites due to the exclusion of some of what Google used to call "supplemental results." Test the CSE with some phrase searches for pages buried deep on a site and compare the results to a regular Google search.

5. Other CSE Letdowns: Google's CSE has a few other limitations, which is perhaps why it is still in beta. Unlike a regular Google web search, a CSE does not give an estimated number of total results. It just displays the first 10 results with links to more pages. Nor does a CSE always include cached copies of pages, links to "similar pages," or connections to other Google database links found on a regular Google search.

Try these tools and tips and build a customized search engine or two for your own users to help them find more targeted and relevant web information.

Greg R. Notess (; is reference team leader at Montana State University and founder of

Links Google Custom Search Engine

Live Search Macros



Yahoo! Search Builder
COPYRIGHT 2008 Information Today, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2008 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Notess, Greg R.
Publication:Computers in Libraries
Geographic Code:1USA
Date:May 1, 2008
Previous Article:Curating the Web: building a Google custom search engine for the arts: Google Custom Search Engine puts you in control of your own search engine.
Next Article:Simple search aids: with a little help from my friends.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters