Create User Loyalty by Improving Search Capability

Why do most B2B sites fail? Could part of the reason be that potential customers had trouble finding what they were looking for?

Forrester interviewed 50 e-commerce executives and found that 93 percent of their sites either didn’t offer a site search at all or had engines on their sites that failed basic tasks. The absence of critical content and the fact that the most relevant data was buried at the bottom of the page were among the most critical failure points reported. Although 68 percent of those interviewed described a search capability as “extremely important” to their web site functionality, only 52 percent had measured how successful users were in searching for information on their site engines.

Forrester found that many causes contribute to the performance failure of web site search functions:

• Weak technology: technology that’s unforgiving, too aggressive
• Inadequate background knowledge: no common language or defined relationships
• Poorly managed content: inconsistent tagging, redundant titles, and unusable descriptions
• Unfriendly interfaces: unhelpful input, inflexible output, and no way to take action

I’ve been asked by readers to suggest how e-commerce web sites can improve site search capabilities. I interviewed several of the industry’s leading search engine designers to gather information that might help you decide on an engine that best serves your web site’s needs.

AltaVista Search Engine 3.0 recognizes more than 225 different file formats and allows Internet and intranet sites to index information in databases and web, multimedia, news, and file servers. The software also allows site users to search in 30 languages across multiple geographies and customize e-commerce catalog searches. AltaVista Search Engine 3.0 was designed with a scalable, 64-bit multiprocessing architecture that integrates easily with back-end databases and business applications. AltaVista Business Solutions’ software and services are used by more than 1,000 customers worldwide, including e-commerce and Internet leaders such as Amazon.com, Reuters, Borders Online, Ariba, and BUY.COM.

Sameer Samat, chief technology officer and founder of Mohomine Inc., was kind enough to provide me with his top 10 suggestions for what to look for when evaluating knowledge management technology:

1. Advanced machine learning: “content-based” information retrieval and organization
2. Automation: scalable, machine-driven knowledge management procedures
3. Customization: personalization
4. Topic-specific knowledge mining: client-defined content
5. Information extraction and summarization: the ability to extract meaningful content
6. Handcrafted taxonomies that are automaintained
7. Good precision and recall: the selection and editing of documents from consideration as the search engine crawls
8. Parametric searching: retrieval organization of information by textual qualities and properties
9. Hosted ASP models: elimination of the need for companies to purchase, maintain, and upgrade knowledge management technology in-house
10. There is no number 10: The world is not always neat and simple even though we try to make it that way.

The name Mohomine was created by merging the geological term “Mohorovicic discontinuity,” the boundary separating the earth’s crust and mantle, with the word “mine.” These people are serious about their business, and it appears Mohomine can organize all your data, no matter what the document type or format: white papers, product specs, product catalogs, news, general information, and so on.

Twirlix offers hybrid search-and-directory solutions that are 99.9 percent automated. The company’s quick-to-market solutions can typically be implemented in two to four weeks. Working on an ASP model for vertical portals and corporate sites, they are fully customizable and capable of integrating any category and subcategory of information that the client defines.

Thomas Neubert, president of Twirlix Internet Technologies Inc., offered that “Twirlix provides both topic-specific and corporate sites an opportunity to expand the breadth of information offered to their end users.” The major benefit is greatly enhanced customer retention while keeping the end user in the familiar surroundings of your site.

EasyAsk, when added to an e-commerce site, accepts natural language or keyword input and then precisely matches the search requirements to the entire contents of the site’s product catalog. Typically, over 90 percent of searches can be accurately satisfied because EasyAsk uses all possible information: product categorization, attributes requested (price range, color, etc.), and descriptive information. David Harris, vice president of marketing at EasyAsk Inc., said, “For example, with EasyAsk, a search for ‘wine’ will find burgundy, chardonnay, and any other wine but will not show ‘wine racks.'”

“Analysts have estimated that more than $6 billion in online purchases was lost in 1999 because of badly designed search functionality and poor results,” said Rajiv Parikh, director of product marketing for AltaVista Business Solutions. Don’t fall into this trap. Your e-commerce site can turn one-time shoppers into loyal, repeat customers when you help online buyers find exactly what they’re looking for.

And this goes for all types of sites. I hope this research helps get you started on your quest for a quality search engine for your web site.