What’s all the Buzz about?
First Monday

What’s the Buzz about? An empirical examination of Search on Yahoo!



Abstract
We present an analysis of the Yahoo Buzz Index over a period of 45 weeks. Our key findings are that: (1) It is most common for a search term to show up on the index for one week, followed by two weeks, three weeks, etc. Only two terms persist for all 45 weeks studied — Britney Spears and Jennifer Lopez. Search term longevity follows a power–law distribution or a winner–take–all structure; (2) Most search terms focus on entertainment. Search terms related to serious topics are found less often. The Buzz Index does not necessarily follow the "news cycle"; and, (3) We provide two ways to determine "star power" of various search terms — one that emphasizes staying power on the Index and another that emphasizes rank. In general, the methods lead to dramatically different results. Britney Spears performs well in both methods. We conclude that the data available on the Index is symptomatic of a celebrity–crazed, entertainment–centered culture. Readers are invited to blog this paper at http://yahoobuzz.blogspot.com.

Contents

Introduction
Methodology
Results
Conclusions

 


 

++++++++++

Introduction

Individuals use search engines such as Yahoo and Google to find what they want in the vast ocean of content that is the World Wide Web. The words Web users enter into a search box provide us with a window to their mind. Search engines match individuals with information needs with sites that provide that information. In this way, search engines help facilitate the exchange of information between publishers and users.

Users arrive at search engines with various agendas. A teenager may be searching for "Britney Spears" to locate her latest song lyrics while her mom may be searching the same term to keep in touch with what kids see in Britney. Others may want images or gossip or may want to participate in online discussions about her. Aggregating the searches conducted by many users of the Web provides us with a window to not just the mind of one user, but also to culture as a whole (Brooks, 2004). As a result, it is a powerful way to understand societal trends.

Wiggins (2001) provides us with an interesting glimpse of search patterns in Google on 11 September 2001. Surprisingly, the number one search term on Google that day was "cnn" followed by more commonly expected terms such as "World Trade Center" and the "Pentagon." This information now exists as a crucial historical record of what transpired on that day and how the collective online mind constructs searches.

In this paper, we study the data presented in the Yahoo Buzz Index over a period of 45 weeks. Our original intent was to study a period of 52 weeks. However, Yahoo did not provide data for seven weeks; details to follow. This index chronicles the top search terms on Yahoo.com’s search engine every week. The data provides us with a unique vantage point to study trends in today’s culture.

 

++++++++++

Methodology

The Yahoo! Buzz Index (See Figure 1 for a screenshot) observes, measures and tracks the search behaviors of Yahoo users. Yahoo Buzz totals the number of people searching for certain terms on the Yahoo search engine on a given day, compiles them and then ranks them in relation to the most searched words (or terms). On the Index, the search terms are ranked 1–20 with number one being the most searched of term that week on Yahoo.com. The editors of the Index filter out heavily searched adult–only sexual queries.

The Google Zeitgest is similar to the Yahoo Buzz Index. However, they do not provide a systematic time trend of top 20 search terms. Therefore, we chose the Buzz Index for our study.

figure1.gif
Figure 1: Yahoo’s Buzz Index.

Yahoo! Buzz Index tracks not only the most frequently searched terms, but also "Buzz Movers," "Buzz Leaders," and "Breakout." "Buzz Movers" are the terms with the greatest percentage increase in the buzz score from one day to the next. Significant boosts in the score do not automatically indicate large overall interest in the subject for a prolonged period of time. As we will illustrate in this paper, most "Buzz Movers" do not stay on the Index for more than one week. However, "Buzz Leaders" are the "lucky" few terms that remain on the Index for an extended period of time. When a search term moves from very few searches to a large number of searches it becomes a "Breakout." It holds the largest percentage increase from the previous day and moves straight to the top of the list of "Movers."

We collected data from Yahoo Buzz Index covering a period of 45 weeks from 22 June 2003 through 22 April 2004, focusing on the United States only. There are seven weeks of data missing from the Yahoo Buzz Index, from 5 July 2003 through 24 August 2003. Therefore, we compiled the archived data on the top twenty search terms for each of the available 45 weeks, resulting in 155 unique search terms.

One might argue that the Yahoo Buzz Index is not a complete picture of society since not everyone has access to the Web and not all Web users use Yahoo regularly. We certainly recognize these limitations in our analysis.

Our original intent was to look at world data. However, the lack of archival data and the nature of the data indexed prevented this effort. No reliable time trend of world Buzz data is available. Moreover, the world data is based on searches done in English, and those searches done in international sites are given no weight. Given these limitations, the terms do not necessarily reflect the culture of the countries included.

We categorized the search terms for purposes of evaluation into the following subject catagories: entertainment, sports, news, and seasonal. The entertainment category includes celebrities, performers, movies, music, games, television, and books. The sports category contains athletes, teams, associations, and events. The news category includes world events, commentary, politics, health, and weather. The seasonal category consists of holidays, taxes, and daylight savings time. Appendix 1, "List of All Search Terms Sorted by Longevity on the Yahoo Buzz Index," provides a list of all 155 terms ordered by the number of weeks each search term appeared on the index.

 

++++++++++

Results

Figure 2 provides the distribution of the number of weeks on the Index. Table 1 provides a quantitative description of the same data. We find that the data is very asymmetric, exhibiting power-law characteristics (Huberman, 2003; Barabási, 2003; Watts, 2003). Only two terms (Britney Spears and Jennifer Lopez) were found on all 45 weeks. It was most common (41.94 percent of terms) for a term to be on the Index for only one week. In fact, 67.1 percent of terms were on the Index for one, two, or three weeks and 80.65 percent of terms were on the Index for eight weeks or less. Of the 155 search terms, only two, Britney Spears and Jennifer Lopez, made it on the top twenty searches for all 45 weeks.

figure2.gif
Figure 2: Distribution of search term longevity on the Yahoo’s Buzz Index.

 

Table 1: Frequency distribution of search terms on Yahoo Buzz Index.
Number of weeks
Frequency
Cumulative frequency
Percent
Cumulative percent
1
65
65
41.94%
41.94%
2
27
92
17.42%
59.35%
3
12
104
7.74%
67.10%
4
11
115
7.10%
74.19%
5
4
119
2.58%
76.77%
6
2
121
1.29%
78.06%
7
1
122
0.65%
78.71%
8
3
125
1.94%
80.65%
9
1
126
0.65%
81.29%
10
3
129
1.94%
83.23%
11
3
132
1.94%
85.16%
12
4
136
2.58%
87.74%
13
2
138
1.29%
89.03%
15
3
141
1.94%
90.97%
16
1
142
0.65%
91.61%
20
1
143
0.65%
92.26%
21
2
145
1.29%
93.55%
23
1
146
0.65%
94.19%
26
1
147
0.65%
94.84%
27
1
148
0.65%
95.48%
38
2
150
1.29%
96.77%
41
1
151
0.65%
97.42%
43
2
153
1.29%
98.71%
45
2
155
1.29%
100.00%
Grand total
155
 
100.00%
 

Next, we were interested in classifying the search terms on the basis of their "star power" in the culture. At first glance, it may appear that computing the average weekly rank is the best way to do this. However, since we know from Table 1 and Figure 1 that most terms stay on the Index for only one week, this analysis does not provide us with diagnostic results.

We provide two ways of computing search term power. The first technique involves computing a "Star Power" for each search term using the equation below:

Search term star power(1) = (20–Rank)*(Number of Weeks on the Index)

Then, the lowest term’s search term power is set to 1 and the other terms are recalibrated to create an index. This approach to star power emphasizes staying power over great performance in one week. The results for our data are shown in Figure 3. The index varies from 1 to 39.73. The top term (Britney Spears) could be thought of having about 40 times as much star power as the lowest term. The top 10 terms using this approach are shown in Table 2; there is some displacement from the data based on the longevity on the index only.

figure3.gif
Figure 3: Distribution of Star Power(1) across search terms.

The second approach to computing star power involves adjusting the rank based on the number of weeks on the Index. Specifically, star power using this approach (called "Star Power(2)" for convenience) would be defined as shown below:

Search Term Star Power(2) = (Average Weekly Rank)*45/(Number of Weeks on Index)

This approach overemphasizes weekly performance in contrast to staying power. In other words, if one term does well in the one week it shows up, it will shoot to the top of this list. This leads to the graph shown in Figure 4.

figure4.gif
Figure 4: Distribution of Star Power(2) across search terms.

 

Table 2: Comparison of top 10 terms using different approaches.
 
Top 10 terms based on longevity
(i.e., number of weeks on Index)
Top 10 terms using Star Power(1)
Top 10 terms using Star Power(2)
Ranking
Search term
Number of weeks
Search term
Number of weeks
Search term
Number of weeks
1
Britney Spears
45
Britney Spears
45
Fallujah
1
2
Jennifer Lopez
45
Jennifer Lopez
45
Chinese New Year
1
3
Beyonce Knowles
43
Beyonce Knowles
43
John Ritter
1
4
Christina Aguilera
43
50 Cent
38
Miss Universe
1
5
Eminem
41
Christina Aguilera
43
Pamela Anderson
3
6
50 Cent
38
Eminem
41
U.S. Open
3
7
Linkin Park
38
Kazaa
27
Britney Spears
45
8
Kazaa
27
NASCAR
26
Texas Chainsaw Massacre
2
9
NASCAR
26
Paris Hilton
20
Golden Globe Awards
1
10
NFL
23
NFL
23
Grammy Awards
1

Star Power(1) generates results similar to the number of weeks. However, there is a displacement in the relative ranking. For instance, "50 Cent" shows up at number four under Star Power (1) and at number six for longevity. Star Power (2) is remarkably different since terms that show up on the Index for one week get a chance to make it. Fallujah is the number one search term. It is striking to see Britney Spears show up in all three categories. She displays staying power as well as a high rank on the Index.

Next, we provide case studies of key "Buzz Movers" and "Buzz Leaders." We describe two "Buzz Movers" — Fallujah and Linkin Park — and two "Buzz Leaders" — IRS and Janet Jackson. While examining the graphs shown in Figures 5–8, one must keep in mind that "low rank" indicates superior performance in terms of number of searches.

"The Buzz Movers"

figure5.gif
Figure 5: Fallujah on the Buzz Index.

Fallujah was on the list of top twenty search terms for only one week, but it was a "Buzz Mover" because it reached the number one spot, thereby giving it a higher ranking. This is also captured in the Star Power(2) data shown in Table 2.

figure6.gif
Figure 6: Linkin Park on the Buzz Index.

In contrast, Linkin Park had an average rank of 12.5 (the lowest rank for our data), yet was considered a "Buzz Leader." Even though it showed up in the top twenty search terms, it may have resided low on the list, in the number 19 or 20 spot for example, for the 36 weeks it remained on the Index.

"The Buzz Leaders"

figure7.gif
Figure 7: U.S. Internal Revenue Service (IRS) on the Buzz Index.

This graph shows that there was little interest in the Internal Revenue Service until approximately 18 January 2004, then it climbs consistently to its peak on 18 April 2004, then drops off and does not make it onto the Index in the top twenty search terms after that date.

figure8.gif
Figure 8: Janet Jackson on the Buzz Index.

To demonstrate how volatile and fickle searches can be, consider the history of the recent Janet Jackson brouhaha. On 28 January 2004, Janet Jackson promised "shocking moments" during the SuperBowl XXXVIII half–time show. On 1 February 2004, during the act, Janet Jackson’s "wardrobe malfunction" occurred. On 2 February 2004 she issued a written statement of apology. Not surprisingly, SuperBowl XXXVIII met with great search interest in the days following with increased media and FCC (U.S. Federal Communication Commission) attention. This was mirrored in the number of searches in connection to her name.

The majority of Yahoo search engine users search for terms in the entertainment and sports category as illustrated in Figure 9 below. It was interesting to see that news terms did not dominate. We had expected to see the "news cycle" drive the Buzz Index to a greater degree. It is too early to call for the demise of the media, of course. However, this data should add grist to the conversation about the role of media in our society.

figure9.gif
Figure 9: Classification of search terms.

We looked particularly at the top 10 search results for entertainment, sports, news and seasonal for the 45 week period on the Index: See Table 3.

 

Table 2: Top 10 search terms, by subject categories.
Ranking
Entertainment
Sports
News
Seasonal
1
Britney Spears
NASCAR
NASA
Internal Revenue Service
2
Jennifer Lopez
NFL
Mars
Halloween
3
Beyonce Knowles
NBA
Nick Berg
Christmas
4
Christina Aguilera
Super Bowl
American beheading video
Martin Luther King Jr.
5
Eminem
NCAA basketball
Atkins Diet
Valentine’s Day
6
50 Cent
U.S. Open
Donotcall.gov
Chinese New Year
7
Linkin Park
Boston Red Sox
Drudge Report
Christmas cards
8
Kazaa
Chicago Cubs
California wildfires
Cinco de Mayo
9
Jessica Simpson
Kobe Bryant
Fallujah
Daylight Savings Time
10
PlayStation 2
New York Yankees
Hurricane Isabel
Mother’s Day

The entertainment category was dominated by personalities. Only one search term was a brand. Sports leagues mostly led the sports category. However, three sports franchises — Chicago Cubs, Boston Red Sox, and New York Yankees — are featured. The seasonal category is a fascinating collection of observances and important days in the calendar. Classifying these terms is tricky. Some of the entries may belong in multiple categories. For instance, Kobe Bryant has been in the news due to a sex scandal and hence, perhaps belongs in both the sports and news categories. Such overlap is inevitable.

 

++++++++++

Conclusion

The Yahoo Buzz Index displays a "winner–take–all" structure. A few terms dominate by showing up for many weeks. American society seems to be consistently interested in entertainment, scandal and sports, so searches on these topics are frequent. We also conclude that users are not necessarily making use of Yahoo for academic or research purposes, at least at a frequency to appear in the Index. More commonly, searchers wanted information about topics related to entertainment.

The Yahoo Buzz Index is highly competitive. Serious topics such as "Fallujah" have to compete with terms such as "Britney Spears" to succeed. Yahoo searches appear to be faddish, with the greatest number of searches centered on pop culture. Unexpected, weekly "breakout" events, such as a search for Fallujah or John Ritter, may shoot these search terms to the top of the Index, but they are short–lived. An event in the news or a scandal can shift the public’s short attention span away to searches on new and different personalities or topics. This is the one week of fame for these terms.

The death of Ronald Reagan, on 5 June 2004, provides an interesting case study. This event never made it into the Buzz Index over the next few weeks, while the usual suspects "Britney Spears" and "Jennifer Lopez" continued to make it. Ironically, dead presidents have to compete with pop culture icons in the world of today’s search engines for attention — a commentary in itself on American culture.

When utilizing search engines, we are a culture of voyeurism and escapism — whether it appears in searches for entertainment, scandal, or natural catastrophes. Those individuals that linger on the Buzz Index have a talent for remaining in the public eye, such as J–Lo, Britney Spears or Beyonce. Ironically, Madonna, a master marketer, showed up only four times in our survey, perhaps due to her declining influence with younger audiences. Individuals, more than events, appear heavily on the Index. Individuals are dynamic, while events are more likely to be static. Moreover, today’s celebrities tend to take on a larger–than–life personality that dominates American culture. We can conclude that those individuals who find an interest in the Yahoo Buzz Index tend to be interested in celebrities, rather than meaningful, world–encompassing events.

Our hope is that this research will spur greater interest in how users search online. Future research will certainly include a textual analysis of search terms. End of article

 

About the authors

Nicole Bladow is a student at the University of Washington, Bothell campus majoring in Interdisciplinary Arts and Sciences with a concentration in Society, Ethics and Human Behavior.
E–mail: nbladow [at] u [dot] washington [dot] edu.

Cari Dorey is a business student at the University of Washington, Bothell.

Liz Frederickson is a student at the University of Washington, Bothell campus majoring in Interdisciplinary Arts and Sciences with a concentration in Society, Ethics and Human Behavior, and minoring in Human Rights. After 20 years in the legal field, she returned to school and will apply to the Teacher Certification program in January, 2005. She looks forward to teaching.
E–mail: bissy62 [at] u [dot] washington [dot] edu.

Pavla Grover is a student at the University of Washington, Bothell campus majoring in Interdisciplinary Arts and Sciences with a concentration in Society, Ethics and Human Behavior. Pavla is originally from the Czech Republic and currently resides in Bellevue, Wash.
E–mail: pavlag [at] u [dot] washington [dot] edu.

Yvette Knudtson has nearly twenty years of experience in the field of Business. She is currently a student at the University of Washington, Bothell campus majoring in Interdisciplinary Arts and Sciences with a concentration in Culture, Literature and the Arts, minoring in Education. Yvette plans to apply to the Teacher Certification program.
E–mail: yvettek [at] u [dot] washington [dot] edu.

Sandeep Krishnamurthy is an Associate Professor of Marketing and E–Commerce at the University of Washington, Bothell. His current research interests are in the areas of search engines, open source software, spam and online communities. He invites you to visit his Web site at http://faculty.washington.edu/sandeep and his blog at http://sandeepworld.blogspot.com. This paper was a collaboration in his "Search and the Web" class.
E–mail: sandeep [at] u [dot] washington [dot] edu.

Voula Lazarou lived abroad in Greece while working for United Arab Emirates Airlines. She has been in corporate travel industry for thirteen years and after major transformations decided to go back to school and finish her degree. Voula is currently a student at the University of Washington, Bothell campus majoring in Interdisciplinary Arts and Sciences with a concentration in Global Studies.
E–mail: voulala [at] u [dot] washington [dot] edu.

 

References

Albert–László Barabási, 2002. Linked: The new science of networks. Cambridge, Mass.: Perseus.

Terrance Brooks, 2004. "The nature of meaning in the age of Google," Information Research, volume 9, number 3 (April), at http://informationr.net/ir/9-3/paper180.html, accessed on July 23, 2004.

Bernardo Huberman, 2003. The laws of the Web: Patterns in the ecology of information. Cambridge, Mass.: MIT Press.

Duncan J. Watts, 2003. Six degrees: The science of a connected age. New York: Norton.

Richard Wiggins, 2001. "The effects of September 11 on the leading search engine," First Monday, volume 7, number 10 (October), at http://firstmonday.org/issues/issue6_10/wiggins/, accessed on 23 July 2004.

 

Appendix 1: List of all search terms sorted by longevity on the Yahoo Buzz Index.
Ranking
Search term
Number of weeks
1
Britney Spears
45
2
Jennifer Lopez
45
3
Beyonce Knowles
43
4
Christina Aguilera
43
5
Eminem
41
6
50 Cent
38
7
Linkin Park
38
8
Kazaa
27
9
NASCAR
26
10
NFL
23
11
Jessica Simpson
21
12
Playstation 2
21
13
Paris Hilton
20
14
American Idol
16
15
Usher
15
16
Harry Potter and the Prisoner of Azkaban
15
17
OutKast
15
18
Internal Revenue Service
13
19
WWE
13
20
Janet Jackson
12
21
Tupac Shakur
12
22
Kazaa Lite
12
23
NeoPets
12
24
Hilary Duff
11
25
Sean Paul
11
26
NBA
11
27
Lord of the Rings
10
28
Chingy
10
29
R. Kelly
10
30
D–12
9
31
Halloween
8
32
The Passion of the Christ
8
33
Blink 182
8
34
Orlando Bloom
7
35
Lindsay Lohan
6
36
Slipknot
6
37
Bow Wow
5
38
Super Bowl
5
39
Toxic
5
40
Harry Potter
5
41
B2K
4
42
Christmas
4
43
Madonna
4
44
Metallica
4
45
Michael Jackson
4
46
Survivor All–Stars
4
47
The Bachelor
4
48
William Hung
4
49
Eamon
4
50
NASA
4
51
Kill Bill
4
52
Academy Awards
3
53
Mars
3
54
NCAA Basketball
3
55
Nick Berg
3
56
Shrek 2
3
57
The Apprentice
3
58
Troy
3
59
U.S. Open
3
60
Avril Lavigne
3
61
Justin Timberlake
3
62
Pamela Anderson
3
63
The Matrix
3
64
American Beheading Video
2
65
Ashanti
2
66
Atkins Diet
2
67
Big Brother
2
68
Boston Red Sox
2
69
Chicago Cubs
2
70
Donotcall.gov
2
71
Drudge Report
2
72
Evanescence
2
73
Friends
2
74
Hulk
2
75
JoJo
2
76
Kobe Bryant
2
77
Martin Luther King Jr.
2
78
Milkshake
2
79
New York Yankees
2
80
Texas Chainsaw Massacre
2
81
The Masters
2
82
The Matrix: Revolutions
2
83
Valentine’s Day
2
84
Van Helsing
2
85
Victoria’s Secret
2
86
Wimbledon
2
87
Johnny Depp
2
88
My Band
2
89
Nelly
2
90
World Idol
2
91
38241
1
92
2 Fast 2 Furious
1
93
2004 NFL Draft
1
94
Angelina Jolie
1
95
Australian Open
1
96
Average Joe
1
97
California Wildfires
1
98
Charlize Theron
1
99
Chinese New Year
1
100
Christmas Cards
1
101
Cinco de Mayo
1
102
Daylight Savings Time
1
103
Disneyland
1
104
Elisha Cuthbert
1
105
Elliott Smith
1
106
Emmy Awards
1
107
Euro 2004
1
108
Fallujah
1
109
Freddy vs. Jason
1
110
French Open
1
111
Golden Globe Awards
1
112
Good Charlotte
1
113
Grammy Awards
1
114
Halle Berry
1
115
Harley Davidson
1
116
Hurricane Isabel
1
117
John Ritter
1
118
Johnny Cash
1
119
Jonathan Brandis
1
120
Kentucky Derby
1
121
Los Angeles Lakers
1
122
Ludacris
1
123
Martha Stewart
1
124
Mean Girls
1
125
Miss Universe
1
126
Miss USA
1
127
MLB
1
128
Mother’s Day
1
129
MTV Video Music Awards
1
130
National Hurricane Center
1
131
NIT
1
132
North American International Auto Show
1
133
North Korea
1
134
PGA
1
135
Prince Charles
1
136
Robert Palmer
1
137
Roy Horn
1
138
Ryongchon
1
139
Saddam Hussein
1
140
SARS
1
141
Sea Shepherd
1
142
Shania Twain
1
143
Siegfried and Roy
1
144
South Beach Diet
1
145
Star Wars Kid
1
146
Super Bowl Commercials
1
147
Super Bowl Halftime Show
1
148
Survivor: Pearl Islands
1
149
t.A.t.U
1
150
The Day After Tomorrow
1
151
Uma Thurman
1
152
Xbox
1
153
Kellie Waymire
1
154
Paradise Hotel
1
155
Return of the King
1

Editorial history

Paper received 23 July 2004; accepted 25 October 2004.
HTML markup: Christopher Day and Edward J. Valauskas; Editor: Edward J. Valauskas.


Creative Commons License
This work is licensed under a Creative Commons License.

What’s the Buzz about? An empirical examination of Search on Yahoo! by Nicole Bladow, Cari Dorey, Liz Frederickson, Pavla Grover, Yvette Knudtson, Sandeep Krishnamurthy, and Voula Lazarou
First Monday, Volume 10, Number 1 - 3 January 2005
http://firstmonday.org/ojs/index.php/fm/article/view/1201/1121





A Great Cities Initiative of the University of Illinois at Chicago University Library.

© First Monday, 1995-2017. ISSN 1396-0466.