OTOinsights "An Analysis of Digg.com Engagement and User Behavior"

Social Media for Marketing:

an analySiS of digg.coM engageMent

and USer Behavior

je f f rey bardzel l , ph .d . shaowen bardzel l , ph .d . ty ler pace

Social media is a broadly used term that includes a wide range of techniques and tools

for leveraging and supporting the sharing of information among people. Internet

enabled tools are the locus for social media activity as they provide low-cost access

to many people allowing for a new wave of targeted advertising to develop. As social

media continues to grow at a near exponential rate, marketers and researchers alike

are becoming increasingly interested in the practices of the users who help shape the

content at highly participatory social media sites. Digg.com is an exemplar site for

user participation in social media and is the focus of this report. Understanding the

behavior of Digg.com users will help marketers to better represent and promote their

material on Digg.com as well as provide insights into the practices of social media

users in general.

In this report, we present an analysis of user engagement with social media via

Digg.com. Viewer engagement was measured with OTOinsights Quantemo neuro-

marketing research system. Quantemo utilizes a multi-modal approach that combines

self-report, physiological and neurological data to holistically and reliably measure user

engagement with digital media. Analyzing the results from the Quantemo sources,

we present a set of four insights concerning how users engage with social media and

how the cues in social media systems positively inform user behavior.

inSightS

1. Users Seek Tailored News Experience and Not Mass Media Stories

2. Story Headlines Are the Chief Determinant of Digg Behavior

3. Short, Direct, and Branded Headlines Dominate Digg.com

4. Positive Engagement Scores Associated with Desirable Digg Behavior

execUt ive SUMMary : :

Copyright 2008, One to One Interactive www.onetooneinteractive.com 1


introdUct ion : :

Social Media

Social media is a broadly used term that includes a wide range of techniques and tools

for leveraging and supporting the sharing of information among people. Internet enabled

tools are the locus for social media activity as they provide low-cost access to many people

allowing for a new wave of targeted messaging to develop. Many consider social media to

be a significant divergence for mass media which is traditionally controlled by corporate

interests. The inherent expense of participating in traditional media (television, print, radio)

prevents the average consumer or small, niche business from fully leveraging the media.

However, social media with cheap and distributed tools enables anyone to participate in the

dissemination and discussion of information (marketing related or otherwise).

Table 1 offers a breakdown of social media systems. It is not exhaustive by any means

social media tends to reinvent itself on a regular basisbut the list does provide concrete

examples of popular social media systems used at the time of this reports development.

taBle 1: examples of social media platforms.

Social Media eMphaSiS platforM type exaMple

communication Blogs, Microblogs Wordpress, twitter

collaboration Wikis, Social Bookmarking Wikipedia, delicious

Multimedia Video, photo youtube, flickr


digg

Launched in December 2004, Digg.com is one of the oldest and most successful social

media platforms (Figure 1). Digg.com facilitates the social promotion of news via the

Internet. News is broadly interpreted on Digg.com to include essentially anything that can

be linked to Digg.com. Pictures, videos, technical reports, blog posts, corporate websites,

almost anything and everything can serve as acceptable and desirable news on Digg.com.

The diversity of content on Digg.com makes it a very broad-based social media system,

which covers several of the platform types listed in Table 1.

figUre 1: Screenshot of digg.com.

The essential activity on Digg.com is the submission and propagation of stories. Any user

can submit a story to Digg.com by providing a URL, headline and description. New stories

are then added to the upcoming pages on Digg.com where users can browse through

and vote, or digg, stories they believe belong on Digg.com and should be seen by other

users. As a story accumulates diggs (votes), it rises towards the top of the upcoming

page and eventually transfers to the main page of Digg.com. Main page content is seen

as the premier content of Digg.com which has been vetted by the larger Digg.com

community and deemed worthy of viewing by all Digg.com users. It is worth noting that

the specific mechanics by which a story is elevated from upcoming to main page are not

public knowledge so as to prevent some users from gaming the system to harmfully

promote their own content.

Digg.com is one of the most broadly trafficked social media sites on the Internet with

about 50% of traffic coming from outside the US [DIGG]. Additionally, Digg.com is the

33rd most trafficked site on the Internet and receives substantially more traffic than its

closest competitors, Delicious and Reddit [COMPETE]. Table 2 provides a breakdown of

traffic of Digg, Delicious and Reddit. Even though Delicious is experiencing a phenomenal

period of growth, Digg.com has a traffic margin of over 20 million unique monthly visitors.


taBle 2: traffic data for digg.com and its peers.

Social NeWS Site UNiqUe ViSitorS (Monthly) yearly GroWth

digg 22,810,196 44%

delicious 1,595,342 10,505%

reddit 2,833,135 105%


StUdy

As social media continue to grow at a near exponential rate, marketers and researchers

alike are becoming increasingly interested in the practices of the users who help shape

the content at highly participatory social media sites. Digg.com is an exemplar site for

user participation in social media and is the focus of this report. Understanding the

behavior of Digg.com users will help marketers to better represent and promote their

material on Digg.com as well as provide insights into the practices of social media users

in general.

OTOinsights Quantemo system serves as the foundation of this multipart study on social

media practices. Quantemo allows for the simultaneous capture of multiple biophysical

responses (breath rate, galvanic skin response, heart rate) in addition to eye tracking

information. After recording the biophysical measures, Quantemo combines all of the

measures into a single, representative measure of physiological engagement. The

Quantemo Physiological Index or QPI serves as a single point of reference of the overall

level of physical engagement (or disengagement) exhibited by a research participant.

Positive QPI scores represent stronger physiological engagement while negative QPI

scores represent weaker physiological engagement. Analysis of eye tracking footage,

QPI and interrogative data provide the basis for the insights put forth in this report.

Specifically, this study asked Digg.com users of varying levels of Digg.com experience

to visit Digg.com on a control account and browse through the first three pages of the

most popular/main page content. Users were encouraged to browse, digg and bury

stories as they would during one of their regular sessions at Digg.com. After completing

their Digg.com sessions, participants engaged in a debriefing interview with researchers

in which their Digg.com practices were discussed using events from their session as

stimulants for questions.

Seven individuals participated in the study. The average age of participants is 28 with

1 female and 6 male participants. On a scale of 1 (low) to 5 (high) experience with Digg.com,

participants reported an average of 4. Finally, participants in this study visit Digg.com an

average of two times per day and have been members of the Digg.com community for

more than one year.


inSightS

In the following sections, we summarize the findings from this study under the headings

of the primary insights derived from analysis of the QPI, eye tracking and interview data.

inSight 1: Users Seek tailored news experiences and not Mass Media Stories

Every participant in the study commented that they enjoyed very specific categories of

Digg.com content. Whether it was news about current events, video games, lifestyles,

etc., Digg.com users are looking for stories which specifically appeal to them and not a

broad spectrum of news representative of the structure of traditional media network

sites like CNN.com or MSNBC.com. Additionally, participants expressed that they

expected the Digg.com system to automatically present them with the tailored news

experience they desire.

Participants identified two methods that create the tailored news experience provided by

Digg.com. First, participants shared a universal belief in the ability of the computer-based

story promotion system to properly filter good stories to wider audiences and to remove

bad stories from the Digg.com system. Our previous research on emotion, engagement

and internet Video identified a similarly steadfast belief in the capability of computer-

based systems to successfully filter and recommend Internet videos. Similar to our research

on Internet video, participants in the Digg.com study noted a lack of trust in their fellow

Digg.com users to always filter stories correctly. Instead, participants believe that the

computer-based system will compensate for the errors of their peer human users.

Regardless of how stories were delivered to participants, they were expected to be

tailored to the interests of user. In other words, Digg.com users desire context sensitive

stories that appeal directly to the desires and interests (context) of the individual user

and not a more general demographic. Insight 3 highlights an important method through

which Digg.com users facilitate the promotion of contextually appropriate stories

through the Digg.com system.

inSight 2: Story headlines are the chief determinant of digg Behavior

All stories on Digg.com are composed of four primary components: headline, description,

image and the digg button which both displays the current number of diggs for a story

as well as serving as the interface element through which users digg a story (Figure 2).

The Digg.com system limits the amount of text that can be used for the headline

and description (350 characters) fields. Additionally, only one image may be selected

from a storys content page to represent that story on Digg.com. During the course of

this study, we identified 73 pairs of headlines, descriptions, images and digg buttons

with sufficient eye tracking data for analysis. For this study, participants were shown

their eye tracking data and assisted the researchers with the interpretation of

their eye movement and Digg.com behavior.

Our analysis reveals that the digg button and the number of diggs associated with each

study are the least important criteria for viewing or promoting content. Only in the rare

cases where the number of diggs reaches an extreme high (4000+) did users consider

it to be a valuable criteria. For the vast majority of stories, participants spent no time

considering the current number of diggs. Additionally, the image component, while

regularly viewed, was considered to be of minimal value by all participants when

evaluating content.

diGG BUttoN headliNe deScriptioN iMaGe


figUre 2


Headlines and descriptions were more important to users and commanded much more

of their eye tracking time. Interesting, headlines and descriptions were almost equal in the

amount of viewing time received. Descriptions and headlines were similar (p > .05) in all

measures of eye tracking calculated for this study. Figures 3 and 4 depict heat maps and

gaze plots of digg activity while Table 3 outlines the eye tracking results for headlines

and descriptions.

figUre3: heatmap of a digg page.

figUre 4: gaze plot of a digg.com page.


The most compelling part of the eye tracking data is the fact that they are so similar

considering how descriptions contain much more textual content than headlines. All

participants spent dramatically more time looking at each headline word compared to

each description word. Participant self-reports upon viewing their eye tracking data

confirmed the claim that headlines were the single most important factor for influencing

their Digg.com behavior. Additionally, participants identified the critical role of headlines

in setting expectations for the associated content. If expectations were set too low, users

avoid the content. However, if expectations are set too high then the disappointment in

the content makes it unlikely the user will digg submission. Insight 3 offers a detailed

analysis of Digg.com headlines with recommendations for maximizing the effectiveness

of headlines for the promotion of content via Digg.com.

inSight 3: Short, direct, and Branded headlines dominate digg.com

As noted in Insight 2, headlines are the single most important factor for influencing

user behavior on Digg.com. The importance of headlines on Digg.com has not gone

unnoticed by the larger Digg.com community and a complex and nuanced use of

headlines has developed on Digg.com. For this study, we analyzed the headlines of

all 91 stories dugg by our participants. Open-ended analysis of the headlines led to

the creation of a set of six heuristics which were then used for further analysis of the

91 dugg headlines. Headline heuristics were developed along two dimensions:

categorical and descriptive. Categorical heuristics (Table 4) represent the rhetorical

structure of the headline. Descriptive heuristics (Table 5) represent the type of

content emphasized in the headline.

taBle 3: analysis of eye tracking data from digg.com.

MeaSUre deScriptioN headliNe

aVG. fixatioN coUNt 4.78 4.96

aVG. Gaze tiMe 913.59MS 1146.08MS

aVG. fixatioN dUratioN 179.19MS 230.61MS


Even though the heuristics were not intended to be exclusive of each other as headlines

might overlap in their structure or content, no single headline was simultaneously a list,

question and/or statement. Digg.com headlines are clearly delineated along their categorical

heuristics. The stark clarity of the rhetorical structure of Digg.com headlines reinforces the

fact that Digg.com users are constantly looking for ways to reduce the effort required

to read and interpret headlines. Conforming to the primary rhetorical categories of lists,

questions and statements facilitates quick headline consumption by Digg.com users.

Conversely, multiple descriptive heuristics were regularly identified within single headlines.

For example, a headline might be structured as a question, contain a branded quote and

identify the linked media. For example, a headline might read What did Google say about

the Android phone platform?. The headline conforms to the categorical heuristic of a

question, includes the brands Google and Android, and identifies that the story links to an

image. Headlines with complex uses of descriptive heuristics are very common on Digg.com

and do not need to be strayed from by marketers when submitting content.

taBle 4: categorical heuristics used to analyze dugg headlines.

heUriStic deScriptioN exaMple

list headline suggests content in list form.

question headline written as a question.

Statement headline written as a statement.

8 cSS techniques for charting data the 10 Greatest Movie Summers in history

Best school name ever?What Sex is your Brain?

i am an honest Business ManBrain freeze

taBle 5: descriptive heuristics used to analyze dugg headlines.

heUriStic deScriptioN exaMple

identify headline identifies content type.

quote

Branded

higher! higher! [pic]results may vary [coMic]

Building the Big Bang MachineMit tool aims to cut airline delays

Google censorshipMicrosoft calling up Gurus to take on apples Geniuses

headline contains a direct quote from linked story.

headline contains an official brand name.


Statement headlines were the most common, followed much less frequently by questions

and lists. Brands were frequently mentioned in headlines, though few quotes were used.

These findings suggest that Digg.com users are both interested and accepting of news

about popular brands and prefer to have their news delivered in a direct manner (state-

ment) which, as many participants put it, reveals the gist of the story upon a quick

reading. Headlines requiring more effort from the user (question, list) are less frequently

dugg. Table 6 lists the complete results of the heuristic analysis.

In addition to the heuristic analysis of Digg.com headlines, we analyzed the number

of words per headline. As previously stated, Digg.com allows for up to 60 characters

(approximately 10 words) per headline. The average headline length for the 91 dugg

headlines in this study was 7 words. Comparing the number of words in a headline with

the number of diggs revealed a negative relationship (r = -.348, p < .001) among the two

criteria. In other words, stories with fewer words in the headline were more likely to have

more diggs than stories with more words in the headline. Table 8 provides a breakdown

of headline length and average number of diggs.

taBle 6: Breakdown of heuristic categories of dugg headlines.

heuristic % of dugg Stories

list 9

question 14

Statement 79

identify 14

quote 4

Branded 45

headline length average # diggs

5 2213

7 2025

10 1187

taBle 7: headline length and number of diggs.


Our analysis of Digg.com headlines suggest Digg.com users prefer short, direct and

revealing headlines. Underestimating the importance of headlines on Digg.com almost

guarantees the failure of your submission. As one participant, a very successful Digg.com

user and web development blogger, stated I spend almost as much time researching

and writing my headline as I do my entire post!

inSight 4: Positive engagement Scores associated with desirable digg Behavior

As previously mentioned, the Quantemo system measures a suite of physiological

signals (heart rate, breath rate, body temperature, skin conductance) and combines

those measures into a single measure of physiological engagement called the QPI or

Quantemo Physiological Index. Positive QPI numbers reflect heightened physiological

activity while a negative QPI represents depressed physiological activity. The QPI is useful

in measuring engagement because heightened physiological activity is associated with

positive emotional states while depressed physiological activity is associated with

negative emotional states.

For this study, QPI was calculated for each participant while they were engaged with the

Digg.com system (interactions with content sites were not consistent enough for valid

QPI data). Participants were then separated into high (top 25%) and low QPI (bottom

25%) groups.

Separating participants into high and low QPI groups based on their interactivity with

Digg.com itself allows us to compare engagement with desirable social media outcomes

(digging stories, viewing content). Table 8 provides a comparison of social media

behavior in relation to high and low engagement scores.

aVeraGe qpi % StorieS dUGG % coNteNt VieWed

hiGh eNGaGeMeNt 1.36 84 71

loW eNGaGeMeNt -1.72 28 13

taBle 8: average engagement scores for high and low QPi compared to social media behavior.

Participants in the high QPI group were more likely to digg stories, had a higher average

gaze time, number of fixations and fixation duration. Consequently, participants in the low

QPI group dugg fewer stories, had fewer fixations, lower fixation duration and gaze time.

While these trends are not yet statistically significant (p < .1), they show promise for

demonstrating a relationship between user engagement and desirable social media

behaviors. As these trends suggest, the engagement level of the social media platform

itself may impact the behaviors of users. Those who contribute to social media platforms

should consider how their involvement (via the submission of content and crafting of

headlines/descriptions) might heighten user engagement which in turn facilitates the

creation and maintenance of a more desirable and receptive audience.



conclUSion

Digg.com represents a long tail approach to marketing [ANDERSON]. As our study

demonstrates, Digg.com users intentionally seek out tailored, niche news experiences

on Digg.com. While mass media styled stories can find success on Digg.com, all of our

participants felt that the real benefit of Digg.com is filtering the niche story and turning it

into a main stream (for Digg.com) headline. The careful crafting and interpretation of story

headlines is the chief mechanism by which Digg.com users develop their tailored news

experience. Crafting proper headlines is no easy task, but marketers (much more so than

the average Digg.com users) are extremely experienced in crafting precise messages with

limited space and should have no problem crafting successful Digg.com headlines after

understanding the expectations set forth by the Digg.com community for story headlines.

The insights discussed in this study offer some guidelines on how to maximize exposure on

Digg.com and generate interest a submission. Collectively, the insights in this report form

two major pieces of advice for marketers submitting and monitoring stories on Digg.com.

headlines are the key to success. Without a contextual, direct and short

headline digg.com users are much less likely to proceed to viewing the

associated content. including brands and branded messages in headlines

is both acceptable and encouraged. digg.com users are not put off

by marketing efforts, so long as those efforts conform to the general

expectations of digg.com use outlined in this report.

Marketers have some control over how engaging the digg.com experience

is to visitors via the comment section associated with each story submission.

an active comment section on a digg.com story page increased engagement

for all participants and was a key indicator for digging or viewing a story.

Marketers are encouraged to participate in the comments on their own

digg.com submissions in order to promote activity regarding the story.

This study represents the first in a multipart series on social media. This study presents a

developing association between engagement (as measured through the QPI) with social

media systems and desirable social media behaviors (digging stories, viewing content,

etc). Future work in this series will focus on the content itself and how engaging content

influences social media decisions. Finally, the series will connect the two threads, system

engagement and content engagement, to determine how exactly the two types of

engagement interact with each other and inform user decisions on social media systems.

referenceS

Anderson, Chris. (2006). The Long Tail: Why the Future of Business is Selling Less of More.

Digg. http://blog.digg.com/?p=256

Compete. http://www.compete.com/


aMPlifying USer engageMent

New knowledge about human behavior brought to light by social and neuroscience

has fundamentally called into question the old mental models of how advertising

and marketing work. Gone is the notion that consumers make decisions in a

linear think-feel-do way and behavior is guided by rational-only principles. Instead,

memories, emotions, associations, and thoughts play a primary role in how

individuals relate and ultimately engage with brands.

OTOinsights is a primary research offering that is breaking new ground in neuro-

marketing to offer clients advanced and scientific levels of insights into how their

consumers engage with them across the landscape of new media channels.

to learn more about otoinsights, visit www.otoinsights.com

coMPlete one-to-one SolUtionS for BrandS, agencieS, and PUBliSherS

OTOinsights is a One to One Interactive company. Established in 1997, One to One

Interactive is the first enterprise to assemble a complete solution for brands, agencies,

and publishers executing one-to-one marketing strategies. By bringing together one

of the nations leading digital marketing agencies, the worlds most comprehensive

portfolio of permission marketing platforms, unique performance-based social media

networks, and cutting-edge neuromarketing research techniques, the companies

of One to One Interactive build informed and creative customer/constituent strategies

on the belief that digital medias ability to enable engaging one-to-one dialogues is

the future of marketing.

to learn more about one to one interactive, visit www.onetooneinteractive.com

OTOinsights

529 Main Street, Charlestown, MA 02129

617.425.7300

www.otoinsights.com

[email protected]


OTOinsights "An Analysis of Digg.com Engagement and User Behavior"

Technology

positive engagement

eye tracking

social media

quantemo physiological

social media

social media

internet enabled

social media