E-Commerce Web Queries: Excite and Ask Jeeves Study

Web queries are a key process in e-commerce. Web queries are a primary means for translating people's business product, service and information needs for e-commerce. This paper reports a study of business related queries submitted to the Excite and Ask Jeeves Web search services. We sampled a log of 10,000 Excite queries and 10,000 Ask Jeeves question format queries from 19 December 1999 to examine the business queries. Findings include: (1) business queries often include more search terms, are less modified, lead to fewer Web pages viewed, and include less advanced search features, than non-business queries; (2) company or product queries were the most common form of business query; and, (3) Ask Jeeves business queries in question form were largely limited to the format "Where can I buy ..." or the request "I want to buy ...". The study provides insights into the beginnings of e-commerce Web searching.

E-Commerce on the Web

E-commerce is a growing force in the world economy and the Web is becoming a major source of business products, services, and information for many people worldwide (Pavia, 1997; Wigand, 1997; Zwass, 2000). Choi, Stahl, and Whinston (1997) refer to e-commerce as the use of electronic means and technology to conduct commerce, within-business, business-to-business and business-to-consumer. E-commerce involves sharing information, developing, and sustaining business ties and monetary transactions across telecommunications networks (Hoffman and Novak, 1997). Studies estimate that e-commerce increased from $US135 million in 1995 (Martin, 1996) to $US1.6 billion in 1998 (Guglielmo, 1997). However, in many respects, e-commerce is in its formative stages. Major digital retail sites are also developing and building brands, and creating market spaces for information, goods, and services. Recent studies show that most Web content is business-related (Lawrence and Giles, 1999) as 83% of Web servers contain commercial content.

Web Searching

Buying and selling products and services over the Web is becoming a part of everyday life for many who search via many different Web search engines. Many are spending increasing amounts of time working with electronic information and engaging in e-commerce. Web searching services are now everyday tools for information seeking and e-commerce. However, many of these Web interactions are frustrating and constrained. Large-scale, quantitative, or qualitative studies have explored some of the problems experienced in effectively using Web search engines (Lawrence and Giles, 1998) as well as investigated how users' search the Web (Silverstein, Henzinger, Marais, and Moricz, 1999; Spink, Wolfram, Jansen, and Saracevic, 2001). To support human information behaviors we are seeing the development of a new generation of Web tools, such as Web meta-search engines, to help users persist in electronic information seeking and resolve their information problems.

Most Web queries are short, without much modification, and are simple in structure (Spink, Wolfram, Jansen, and Saracevic, 2001). Few queries incorporate advance search techniques, and when such techniques are used many mistakes result. However, relevance feedback and some advanced search features are growing in use. Frequently, searching results in a large number of Web sites, but most tend not to browse beyond the first or second pages of results. Overall, a small number of search terms are used with high frequency and many terms are used once. Web queries are very rich in subject diversity and some are unique.

The more effective use of Web search engines is crucial for the development of e-commerce, as search engines are routinely used to find business related information. Many shopping and business Web sites also use search engines as internal navigation tools for their users. Apart from entering a specific business related URL or using a Web classification tool such as Yahoo, most Web users have little choice but to interact with Web search engines, such as Excite, Go.com or Google. In this paper we explore the extent and nature of business related searching on the Web.

Business Web Sites

Many researchers have investigated the factors affecting Internet sales (Phau and Poon, 2000; Vijayasarathy and Jones, 2000), the role of technology on Internet sales (Aberg and Shahmehri, 2000; Rowley, 2000) and the role of trust in the e-commerce relationship (Ratnasingam, 2000). Many studies have examined aspects of business Web sites (Greaves, Kipling, and Wilson, 1999; Ng, Pan, and Wilson, 1998), but few have investigated business-related Web searching. A number of studies have examined how businesses use the Web, including the proliferation and usability of business Web sites. Due to the increase in e-commerce based commercial activity, thousands of companies are developing Web pages and starting to do business over the Web. By August 1998, the Yahoo Business Directory included over 364,000 company Web sites (Greaves, Kipling, and Wilson, 1998). Both business-to-consumer and business-to-business transactions are increasing rapidly over the Web. Ng, Pan, and Wilson (1998) found an increasing sophistication and diversity in business Web sites. Greaves, Kipling, and Wilson (1999) found little difference between U.S. and U.K. company Web sites.

Web Search Engines

Buying and selling products and services over the Web is becoming a part of everyday life for many who search with a variety of Web search engines. Spiteri (2000) compared the effectiveness of six Internet search engines. She found ambiguous and sometimes misleading categories in e-commerce sites, moderate consistency in e-commerce Web site organization, and few opportunities for comparison-shopping. Spiteri (2000) identified two types of consumer Web behavior: (1) goal-directed - to find a product or information; and, (2) experiential - non-directed exploratory browsing or surfing. However, using Web search engines to find information and conduct e-commerce transactions is challenging for many users (Jansen, Spink, and Saracevic, 2000; Spink, Wolfram, Jansen, and Saracevic, 2001). Many Web users find it difficult to conduct e-business via an information system more akin to an automated library catalog than an effective sales transaction system. Studies of business Web queries are important for understanding Web usage and for the development of Web tools and systems to facilitate more effective e-commerce.

Recent studies show that business related queries are increasing as a proportion of all Web queries (Wolfram, Spink, Jansen, and Saracevic, 2001). Jansen, Spink, and Saracevic (2000) found that in 1997 business-related terms constituted 8.3% of the 63 top terms and 13.5% of queries entered into the Excite search engine. By 1999, Wolfram, Spink, Jansen, and Saracevic (2001) found an increase to 24.4% for business-related Web queries. In a recent study of Excite users during December 1999, Spink, Milchak, Sollenberger, and Hurson (2000) found that business-related queries were the second largest category of queries behind people and places. Spink and Ozmutlu (in press) showed that many Ask Jeeves business queries were in request, not question, format. The most common format for e-commerce request queries were "I want to buy ..." and "Get me ...".

To further investigate business related Web queries, we analyzed a sample of 10,000 queries from the Excite Web search engine and 10,000 queries from the Ask Jeeves search engine in question and request format from 19 December 1999. The study is part of an ongoing series of studies on large-scale Excite and Ask Jeeves query logs (Jansen, Spink, and Saracevic, 2000; Spink, Wolfram, Jansen, and Saracevic, 2001).

This study investigates the nature of business related Web queries, by examining the proportion of Web queries that are business related as well as the types of requests for business related information. This paper also compares business related to non-business related queries.

Excite Data Set

Excite, Inc. is a major Internet media public company that offers free Web searching and a variety of other services. Excite searches are based on the exact terms that a user enters in the query. Capitalization is disregarded, with the exception of logical commands AND, OR, and AND NOT. Search results are provided in a ranked relevance order. A number of advanced search features are available. The total data set contained a transaction log of 1.7 million queries from 19 December 1999 contained three fields:

With these three fields, we were able to locate a user's initial query and recreate the chronological series of actions by each user in a session.

Ask Jeeves Data Set

The data set analyzed consisted of a random subset of 10,000 queries in question and request format selected from a transaction log of 800,000 queries submitted to Ask Jeeves on 19 December 1999. Our statistical and qualitative analysis focused on the 10,000 randomly selected queries, or 1.75% of the data set, including queries and terms. Each query is a set of one or more terms entered into the Web IR system during a single search. A term is any string of characters bounded by space. We created a computer program to automatically identify any query ending with a question mark (?) or beginning with one of the common words associated with human questioning: e.g., where, what, who, how, when, can, ... etc.

Data Analysis

A query was defined as a set of one or more search terms; it may include advanced search features, such as logical operators and modifiers. A business query was one that included a company name, product name, brand name, or business-related word or phrase. A user search session included all queries from the beginning to the end of the users' interaction with the Excite search engine.

To determine the extent of business-related Web queries, we qualitatively analyzed and subject categorized a random sample of 10,000 queries by Excite users from 19 December 1999. Each query was analyzed and classified into one business-related category. The categories were developed iteratively from the analysis of the data and evolved as the analysis progressed.

To examine in more detail business-related Web searching, a subset of 95 business-related user search sessions from 19 December 1999 were randomly selected and qualitatively analyzed.

We qualitatively analyzed a random subset of 72 user sessions containing business-related question queries from 19 December 1999.

To determine the extent of business-related Ask Jeeves question and request queries, we qualitatively analyzed and subject categorized a random sample of 10,000 queries by Ask Jeeves users from 19 December 1999.

The results of our analysis provided insights into the nature and subjects of business-related queries, sessions, and queries in question format.

Proportion of Business-Related Queries

Table 1 shows the proportion of business-related queries in the Excite and Ask Jeeves data set analyzed.

	Number of Queries	Percentage of Queries
Total Number of Excite Queries	10,000	100%
Business-Related Excite Queries	1,118	11.1%
Total Number of Ask Jeeves Queries	10,000	100%
Business-Related Ask Jeeves Queries	606	6%

Approximately one in six Excite queries (approx 12%) and one in twelve Ask Jeeves queries (approx 6%) were business-related. Although these figures relate to only two Web search engines, this represents a large amount of business-related Web searching on a daily basis. Excite processes approximately 30 million queries per day (Xu, 1999) of which 12% represents more than three million business-related queries per day. Our figure is lower than the 24.4% business-related queries identified in a 1999 data set of 2,500 queries by Wolfram, Spink, Jansen, and Saracevic (2001).

Excite Business-Related Query Structure

We conducted a detailed analysis of 95 Excite user search sessions containing business-related queries (Table 2).

Mean Number of Excite Queries per Business Session	2.7
Mean Number of Subjects per Excite Business Session	1.5
Mean Terms per Excite Business Query	2.8
Mean Grouping of 10 Web Sites Viewed	2.4
Mean Minutes per Excite Session	20.4

Business Queries in Question Format

	Number of Queries	Percentage of Queries
Total Excite Queries	10,000	100%
Excite Queries in Question Format	1,322	13.2%
Excite Business Related Question Queries	199	2%
Total Ask Jeeves Question Format Queries	10,000	100%
Ask Jeeves Business Related Question Format Queries	1,450	14.5%
Ask Jeeves Business Related Request Queries	30	0.3%

The proportion of Excite queries in question format for business-related queries is high compared to the proportion of queries in question format for the non-business-related queries in the data set. The mean number of terms per business related query was 3.5 (the high percentage of queries in question format contributes to this high mean because they have a higher mean number of terms per query).

We examined in detail the 72 Excite sessions containing the 199 business-related queries in question format. We found that

Each Excite question query was analyzed as either a fact or more general information query. Table 4 shows the results of our analysis.

Type of Excite Question Query	Number of Excite Question Queries	Percentage of Excite Question Queries
Fact	21	29.2%
General Information	51	70.8%
Total	72	100%

Overall, Excite users were looking for general information on companies, products, and industries, rather than specific facts.

Types of Business-Related Information

Business-related information seeking covers a broad range of information problems.

Classification	Number of Queries	Percentage of Queries
Company, Industry or Association Name	470	42%
Tourism, Travel, and Places	217	19.4%
Company URL	137	12.2%
Trade and Commerce	83	7.4%
Financial Information	73	6.5%
Employment	50	4.7%
Legal Issues	49	4.3%
Real Estate	39	3.5%
Total	1,118	100%

Related to the previous category, many were looking for company information and searched using a specific company URL. Many company URLs were search repeatedly, including www.yahoo.com and www.MSNBC.com. If you combine the first two categories together, more than 50% of business queries were company, product, or industry related.

If we include travel and tourism under e-business, then this category represents a sizable proportion of the business related information problem. Queries such as "travel cheap" were common. Many travel related sites on the Web, such Travelocity and PriceLine.Com, are now competing for this business and the growth of Web travel booking sites is increasing.

This category represented searches for product information or shopping that did not include a company or product name. Queries such as "Where can I buy an automobile?" or "Shop online" were common. This type of query involves a more complex response than a company query to assist the user in refining their information request. Many Web services, such as Ask Jeeves, are trying to cater to these users through a question and answer format.

Requests for financial information were less common. Some common queries were "Currency exchange rates", "Personal investment" or "Electronic banking". These queries are more complex that just seeking a specific Web site. This type of user may be seeking financial advice and require much more interaction to assist them in refining their queries.

Legal issues were also a small proportion of business-related queries. A common example is a query related to disability lawsuits and discrimination or a question such as: "Where can I find information on immigrating to the USA?"

Employment related queries were also a small proportion of the queries. Many employment queries were in question format, such as: "How can I get a job?" or "What kind of a job is right for me?"

Real estate related queries were less than one in twenty business-related queries. Typical examples are: "How much is my house worth?" and "luxury homes in Maryland."

Ask Jeeves Queries

Table 6 shows the distribution of starting terms for Ask Jeeves business related question queries.

Ask Jeeves Question Query Starting Terms	Number of Queries	Percentage of Queries
Where can I buy?	1,278	88.1%
Where do I find?	45	3.1%
Where do I buy?	26	1.8%
How do you/I buy?	21	1.5%
What is the best buy?	16	1.1%
How should I buy?	8	0.6%
What should I buy?	7	0.5%
Where is the best place to buy?	7	0.5%
Which ... should I buy?	4	0.3%
Should I buy?	4	0.3%
Can I buy?	4	0.3%
Where would I buy?	4	0.3%
Where should I buy?	3	0.2%
Where is the Web site to buy?	3	0.2%
Where do I buy?	3	0.2%
What do I look for when I buy?	3	0.2%
Where should I start in buying?	2	0.16%
How much do people buy?	2	0.16%
Total	1,450	100%

Most commonly users looking for business related information phrased the beginning of their queries as "Where can/do I buy/find ... ". An example query is "Where can I buy a gun?" A "where can I buy/find" question implies a desire to find a place or physical location for the item desired. However, few users asked "which" Web site contained the item they sought. A "where" question implies a desire to find a place or a physical location for the information desired.

Ask Jeeves Question Query Starting Terms	Number of Queries	Percentage of Queries
I want to buy	16	53.3%
I want to find	3	10%
I would like to buy	3	10%
I need to buy	3	10%
Help me buy	1	3.3%
List sites where I can buy	1	3.3%
I want to read	1	3.3%
Buy	1	3.3%
I'd like to know where I can buy	1	3.3%
Total	30	100%

Instead of the common question format query "Where can I buy/find", requests are generally in the form "I want to buy/find ...". Some users make fairly interesting requests, such "I want to buy women" that could require quite a bit of query interpretation, and "Help me buy illegal drugs" that implies possible legal complications. Some users may prefer to use a question format while others like a request format. Request queries were shorter and more concise than question format queries. Some users may feel more comfortable using a request rather than a question format. This is an area for further research.

In general Web search engines are currently trying to cater to the broad, rather than specialized, nature of human information needs. Many are increasingly searching for business related information with general Web search engines. The nature of their business related information needs are broad and relate to all aspects of corporate interests. Our results show that company (and industry) related queries account for more than half of all business related Web searches. Many of these searches are attempts to locate specific corporate Web sites. A need exists for new search tools that more easily facilitate company and product searches. Excite allows industry search, but it would be easier for people to use it if Excite moved this feature to its main page. A number of searches are aimed at shopping sites as a number business-related queries appeared to be directly shopping-related. These searches are aimed at brand names as well as generic types of products, such as "cars" or "golf clubs".

We found that great majority of Web business queries posed by the public are short, not very modified, and very simple in structure. Few queries incorporate advance search features and when they do many contain mistakes. Despite seeing results that, as a rule, list a large number of Web sites as answers to their queries, Web users for business information view few result pages, tend not to browse beyond the first or second page of results. Overall, a small number of terms are used with very high frequency, while there are great many terms that are used only once. The language of Web queries is very rich and varied, with many unique company and product names appearing as queries. A few high frequency terms include general words like "shopping" and "pizza".

Implications

Many searchers were seeking advice or product comparisons while others were just not sure what they wanted or how to find it on the Web. Many users need a conversation with an "expert" to guide them in expressing and organizing their information needs. Many Web search engines and e-commerce sites are beginning to provide some forms of interaction for customers to help them sort through the plethora of Web sites, services, and products that confront them. Some search engines, such as AltaVista, are providing some form of comparison shopping.

We also compared the results with two other related studies of large queries. Overall, business-related queries are slightly longer than general queries. Many searchers are increasingly querying for e-commerce information in question format. The Ask Jeeves type of Web services are growing and encouraging this trend. The development of Web services that focus on helping people with business-related information needs in question and answer format is the next step in improving the Web as an e-business tool. There is a great market for niche Web services and search engines that facilitate more effective and targeted access to e-business sites. Our study with search log data as the only available data cannot answer other questions about the results of these queries or the performance of different search engines. However, they do provide a snapshot for comparison of public behavior while searching, a behavior that can also serve as a clue for improving Web services.

As the Web continues to become a key centerpiece for e-commerce, new tools and new ways of searching for business information on the Web are needed. The Web lacks a standardized approach to search engine functionality and terminology. Business terminology on the Web is incredibly varied and difficult for the average user to predict with total accuracy. As more Web users begin to ask questions rather than generate Boolean queries, methods and approaches are needed to analyze and process question queries in a "question and answer" format to diagnose the information requirements of average users to facilitate the effective sale of goods and services over the Web. End of article

About the Authors

References

J. Aberg and N. Shahmehri, 2000. "The Role of Human Web Assistants in E-Commerce: An Analysis and Usability Study," Internet Research, volume 10, number 2, pp. 114-125.

S.Y. Choi, D.O. Stahl, and A.B. Whiston, 1997. The Economics of Electronic Commerce: The Essential Economics of Doing Business in the Electronic Marketplace. Indianapolis, Ind.: Macmillan Technical Publishing.

cC. Greaves, P. Kipling, and T.D. Wilson, 1999. "Business Use of the World Wide Web, with Particular Reference to UK Companies," International Journal of Information Management, volume 19, pp. 449-470.

C. Guglielmo, 1998. "The Mezzanine May Be Closed for Merchants," Inter@active Week (9 February), p. 44.

D.L. Hoffman and T.P. Novak, 1997. "A New Marketing Paradigm for Electronic Commerce," Information Society, volume 13, number 1, pp. 43-54.

B.J. Jansen, A. Spink, and T. Saracevic, 2000. "Real Life, Real Users and Real Needs: A Study and Analysis of Users Queries on the Web," Information Processing and Management, volume 36, number 2, pp. 207-227.

S. Lawrence and C.L. Giles, 1999. "Accessibility of Information on the Web," Nature, volume 400 (8 July), pp. 107-109.

S. Lawrence and C.L. Giles, 1998. "Searching the World Wide Web," Science, volume 280, number 5360, pp. 98-100.

M.H. Martin, 1996. "Why the Web is Still a No Shop Zone," Fortune (5 February), pp. 127-128.

H.-I. Ng, Y.J. Pan, and T.D. Wilson, 1998. "Business Use of the World Wide Web: A Report on Further Investigations," International Journal of Information Management, volume 18, number 5, pp. 291-314.

S.C. Palvia, 1997. "The Challenges of Using the Exploding Internet for Global Electronic Commerce," Journal of Global Information Management, volume 5, number 4, pp. 3-5.

I. Phau and S.M. Poon, 2000. "Factors Affecting the Types of Products and Services Purchased over the Internet," Internet Research, volume 10, number 2, pp. 102-113.

P. Ratnasingam, 2000. "The Influence of Power Trading Partner Trust in Electronic Commerce," Internet Research, volume 10, number 1, pp. 56-63.

J. Rowley, 2000. "Product Searching with Shopping Bots," Internet Research, volume 10, number 3, pp. 203-214.

C. Silverstein, M. Henzinger, H. Marais, and M. Moricz, 1999. "Analysis of a Very Large Web Search Engine Query Log," ACM SIGIR (Association for Computing Machinery Special Interest Group on Information Retrieval) Forum, volume 33, number 3, pp. 6-12.

A. Spink, D. Wolfram, B.J. Jansen, and T. Saracevic. 2001. "Searching the Web: The Public and Their Queries," Journal of the American Society for Information Science, volume 53, number 2, pp. 226-234.

A. Spink, S. Milchak, M. Sollenberger, and A. Hurson, 2000. "Elicitation Queries to the Excite Web Search Engine," CIKM 2000: Proceedings on the Conference on Information and Knowledge Management (November), Washington, D.C., pp. 134-140.

A. Spink and H.C. Ozmutlu, in press. "Characteristics of Question Format Web Queries: An Exploratory Study," Information Processing and Management.

L.F. Spiteri, 2000. "Access to Electronic Commerce Sites on the World Wide Web: An Analysis of the Effectiveness of Six Internet Search Engines," Journal of Information Science, volume 26, number 3, pp. 173-183.

L.R. Vijayasarathy and J.M. Jones, 2000. "Print and Internet Catalog Shopping: Assessing Attitudes and Intentions," Internet Research, volume 10, number 3, pp. 191-202.

R.T. Wigand, 1997. "Electronic Commerce: Definition, Theory, and Context," Information Society, volume 13, number 1, pp. 1-16.

D. Wolfram, A. Spink, B.J. Jansen, and T. Saracevic, 2001. "Vox Populi: The Public Searching of the Web," Journal of the American Society for Information Science and Technology, volume 52.

Editorial history

E-Commerce Web Queries: Excite and Ask Jeeves Study by Amanda Spink and Okan Guner
First Monday, volume 6, number 7 (July 2001),
URL: http://firstmonday.org/issues/issue6_7/spink/index.html

Contents

Introduction