Site Search and Site Network Search
I got an email from Sarah Shaw following my last post about the relaunch of the InterSites.co.uk hub site, asking for more information about the network search which I mentioned briefly. This post will talk in more detail about site search, and how it can help your users as well as adding a revenue stream.
Do You Need Site Search?
The first thing you should ask yourself, is “Do I really need site search?”. If you have a small site, perhaps 50 pages or so, or indeed a site comprised of pages with little textual content you probably don’t have need for site search. There is nothing worse than searching a small site with an internal search engine only to have no results returned for even common terms and phrases. One the other hand if you have several thousand pages or more, a well-implemented site search facility should probably work well.
Requirements and Features
If you’ve decided you want or need site search on your web site, the next step is to put down a list of requirements and features. These will vary depending on the size and scale of your site, as well as what you hope to get out of the functionality. Some of the requirements we have for site search on the individual Inter Sites are:
- Scalable (needs to perform quickly even with tens or hundreds of thousands of pages)
- Customisable layout (needs to fit in with site look and feel)
- Scripting language (php or cgi based)
- Able to replicate across new sites easily
- Ability to index and search multiple sites from one location (our network search)
- Cost effective
- Different match types (broad, exact, phrase, negative etc)
- Category based search (ability to search parts of site as well as whole site)
- Low overheads (in terms of upgrades, processing power, bandwidth requirements etc)
- Ability to implement a revenue stream (paid XML feed integration, selling sponsored listings etc)
Inter Sites Site and Network Search
After testing many different solutions, including the indexed_search solution native to Typo3, we software we ended up choosing was Zoom Search. It offers some excellent functionality, is very scalable (for example our network search currently indexes over 14 million words on 150 thousand pages across the network and returns the most complex queries in well under a second). It also allows us to embed results within a tailored search page to fit in with the look and feel of the site it’s implemented on, and embed paid results in a sponsored box above the internal search results.
Hopefully this has answered some initial site search questions. As always if you have questions please feel free to leave a comment or contact me via the contact page. I will add another post outlining the exact implementation we use at a later date.