exploring caller dialogue: analyzing information search...
TRANSCRIPT
![Page 1: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/1.jpg)
Bing PanDaniel Fesenmaier
eKnowledge GroupNational Laboratory for Tourism and eCommerce
University of Illinois at Urbana-Champaign
Exploring Caller Dialogue: Analyzing Information Search through Call Center for Tourism Products
![Page 2: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/2.jpg)
Communication Channels with Customers
• Traditional printed media – books, brochures, etc.
• Travel agencies, etc.• Computer based - kiosks, etc.• The Internet (Web sites, email-lists)• Call centers
![Page 3: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/3.jpg)
HighLevels of information Richness
Levels of flexibility of information
structure
Low
High
Low Printed Media Database Inquiry
Travel Agencies
Computer Agent
Web Site
Email list
Call center
Communication Channels with Customers
![Page 4: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/4.jpg)
•In rational decision making theory, information searching is used to maximize the utility of the outcome of decision. All the evaluation rules can be classified as either attributes-based or alternative-based (Bettmanand etc, 1998)
•Ongoing search vs. prepurchase search(Bloch, 1985)
Background in Information Search and Decision Making
![Page 5: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/5.jpg)
Background in Information Search and Decision Making in Tourism
•Tourists search information for various kinds of needs (Vogt and Fesenmaier, 1998): functional; hedonic; innovation; aesthetic; sign.
•Travel decisions have been shown to be relatively complex, involving a number (10- 18 ) of sub-decisions or facets (Jeng, 1998): destination(s); travel group; timing; route; activities and etc.
![Page 6: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/6.jpg)
A Framework of Tourist Information Search
DestinationDecision
NeedRecognition
Ongoing Search
DestinationChoice Search
PlanningSearch
Most FunctionalLeast Functional
Most Hedonic Least Hedonic
![Page 7: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/7.jpg)
1-800-2connect
• Call center: part of Customer Relationship Management System (CRMS)
• Provide up-to-date tourist information about:– location: cities/towns, directions/maps, etc.– interest: activities, events, etc.– amenity: prices/discounts, etc.
• Provide travel counseling services:– availability of services, etc.
![Page 8: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/8.jpg)
Search Format
ID Seq. Location Amenity Interest1 1 Chicago Hotel Aquarium2 1 Chicago Hotel None2 2 Springfield Camping
Session and inquiries
![Page 9: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/9.jpg)
Location and Interest Depth
• Illinois General (0)– Illinois Regions (1, Chicago Land, Northern Illinois…)
• Illinois Counties (2, Cook County, Peoria County…)– Illinois Cities (3, Chicago, Urbana…)
• General Interest (0)– Categorical Interest (1, Accommodation, outdoors…)
• Detailed Interest (2, Land, Water…)– Very Detailed Interest (3, Hunting, Picnicking…)
![Page 10: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/10.jpg)
Focus of Study
• Amount of Search (how many inquiries? What elements determined the amount of search?)
• Directions of Search and depth of Search(Location dominated or interest dominated? How detailed are their inquiries? Can we infer different decision stages from the search?)
• Marketing Implications. (How can we use these information?)
![Page 11: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/11.jpg)
0
5
10
15
20
25
30
35
40
45
50
1 14 27 40 53 66 79 92
Number of Searches
Per
cent
age
Number of Inquiries
From Jan to Sept, 1999
74,850 caller sessions for Chicagoland;
169,426 total inquiries;
Average inquires during a session: 2.3
![Page 12: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/12.jpg)
Top Countries of InquiriesCountry Calls Percentage Lasting timeUSA 65422 97.99 10.2Canada 1144 1.71 11.1United Kingdom 74 0.11 13.9Germany 23 0.03 10.8Australia 15 0.02 14.2Ireland 13 0.02 10.2France 8 0.01 7.7New Zealand 8 0.01 8.4Netherlands 6 0.01 20.4Puerto Rico 5 0.01 7.7Brazil 4 0.01 40.1Japan 4 0.01 8.6Belgium 3 0 11.7Iraq 3 0 8.9
![Page 13: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/13.jpg)
Travel Distance and Inquiries
•No significant correlation exists between distance and number of inquiries and between distance and session duration;
•Illinois callers made less inquires ( significant) compared with all USA callers; USA caller make longer sessions compared with foreign callers (significant)
![Page 14: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/14.jpg)
The First Inquiry – Location and Interest
Levels of Location in First Search
14%
21%
1%64%
Illinois GeneralIllinois RegionIllinois CountyIllinois City
Levels of Interest in First Search
2%
91%
6%
1%
General InterestCategorical InterestDetailed InterestsVery Detailed Interest
![Page 15: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/15.jpg)
• The percentage of those people who made their first inquiry about a city or a categorical interest drastically decrease among the multiple search people.
Proportion of Inquirers according to Their First Search
0
20
40
60
80
100
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Searches
Per
cent
age
GeneralRegionCountyCityGeneralCategoricalDetailedVery Detailed
The First Inquiry
![Page 16: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/16.jpg)
6.8231.26F Value
Prob > P 0.0090.0001
Depth of interest search of first inquiry
Depth of location search of first inquiry
Distance and Depth of Search
![Page 17: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/17.jpg)
Change of Levels of Location and Interests
Mean Location and Interest Levels in Different Search
0
0.5
1
1.5
2
2.5
3
search1
search2
search3
search4
search5
search6
search7
search8
search9
search10
LocationInterest
![Page 18: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/18.jpg)
Different Dominance for Multiple Inquiries
558112343Location-dominant
394215472Not location-dominant
Interest-dominant
Not interest-dominant
![Page 19: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/19.jpg)
Change of Proportions of Different Callers
Changes of Search Pattern by Number of Inquires
0
10
20
30
40
50
60
2 3 4 5 6 7 8 9 10 11 12 13 14Search
Perc
enta
ge
Same Location DifferentInterestSame Interest DifferentLocationBoth Dominance
Location Dominance
Interest Dominance
![Page 20: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/20.jpg)
Marketing Implications
• We can infer different search patterns but hard to infer difference search stages;
• Branding at city level seems to be most effective; very few people are using county as search cues;
• The first search has the role of leading to more search; a large amount of first search are mainly functional, focusing on one city and one categorical interest;
• From 8th to 12th inquiry is the most effective time (largest percentage of prepurchase search) for cross-marketing and cross-selling.
![Page 21: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/21.jpg)
System Design Implications
• Information Organization: location-based and interest-based; more direct links from detailed location to more general location (network of locations);
• Most important element for the first search is speed; after the 8th element we can more focus on the hedonic part of the dialogue.
![Page 22: Exploring Caller Dialogue: Analyzing Information Search ...personal.psu.edu/bup63/pan/enter2001-present2.pdf · Bing Pan Daniel Fesenmaier eKnowledge Group National Laboratory for](https://reader033.vdocuments.us/reader033/viewer/2022060300/5f0825b37e708231d42092bb/html5/thumbnails/22.jpg)
Limitations
• In-direct inferences;• System error (more than 20 hours calling time);• More information is needed to explore the
relationship between decision process and the search pattern.