eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Extending the boundaries for customers to create more category-refined and cost-friendly Web content search systems, CrownPeak Technology is readying software capable of learning from content and user behavior to improve its site search relevance.
CrownPeak unveiled its new hosted on-demand CrownPeak Search offering Tuesday at the Gilbane Conference on Content Technologies in Boston. The service is available now.
Written in C++ and operating system platform-independent, the technology is priced at $500 a month to search less than 500 Web pages, at $1,800 a month for 500 to 20,000 pages, and at $200 monthly per each additional site added to the hosted search service, said Jim Howard, CEO of Los Angeles-based CrownPeak.
In order to enable CrownPeak Search, a customer must set up a line of code for every page she wants a search box in.
From there, Howard said, CrownPeak remotely provides the appropriate search response through a protected firewall on the back end and then spiders, indexes and eventually customizes search results for the customers site.
CrownPeak Search features a component from CrownPeaks content management software called related linking. Using linguistic analysis, this engine allows customers to display areas of links of similar or related content. The new search technology also contains the ability to control search results in newest or any desired order, and provides progressive matching to adjust for misspellings and intent interpretation.
Some customers reliant upon designing Web content search systems for a diverse user base are discovering that large-scale search offerings built to sort through tens of millions or billions of documents are too difficult to maintain and too time-consuming to properly support without added costs.
“Our users cover a broad range of very different kinds of people. The problems we were encountering with [other] search engines, to create categories of information or category-specific searches didnt allow us to manage more complex relational [database] searches where we wanted to parse metadata,” said John Stubbs, director of Washington, D.C.-based Online Strategy and Operations for the Office of the U.S. Trade Representative.
“CrownPeaks search engine will be a great benefit to users of our Web site because now they will not be forced to think about trade-specific issues in one silo of information but actually in more natural environments,” he said.
Stubbs said he expects CrownPeak search to be implemented by February and fully operational within his IT environment by the end of Q1 2005.
Due to fixed yearly U.S. government budget restraints and massive volumes of shifting content published on his site, Stubbs said he favors CrownPeaks ASP (application service provider) model with built-in support.
“Theres really very little you can do to customize a search engine without necessary, ongoing, day-to-day of testing and retesting and tweaking,” he said. “Not only making sure you have appropriate ranking systems in place, but also in how to make sure youre publishing pages, and metacontent is in line and sync—it just requires way too much maintenance.”