Social Media Usage at Universities

How Should it Be Done?

Jennifer-Carmen Frey, Martin Ebner, Martin Schön and Behnam Taraghi

Social Learning, Computer and Information Services, Graz University of Technology,

Münzgrabenstraße 35a, Graz, Austria

Keywords: Social Media, Facebook, Google, Twitter, Higher Education.

Abstract: The social media hype these days is omnipresent, encouraging even public institutions to participate. This

study seeks to reveal, which factors have to be kept in mind, when doing social media work at universities.

It also is an attempt to provide a list of recommendations and possible fields of action to ensure an efficient

presence in social web. Therefore we analyzed the present situation of university efforts and evaluated the

success by measuring user engagement concerning different aspects of social media activities (e.g. content,

publishing time, frequency of activities, existence of visual elements, additional links, etc.) The study

shows, that it seems less important how many times a week a university is publishing, or how long the text

messages are in detail, but that there is a significant relationship between the contents of a post, the time of

its publishing and the used elements, pointing out that users actively perceive and interact with social media

activities that encourage contact between both: the profile-owner with the community and the community

amongst itself - especially if made in a personal, emotional or funny way, offering people ways to identify

with the institution and to connect with it through well-known habits and traditions.

1 INTRODUCTION

Social software is still one of the most promising

technologies with continuing success since the rise

and popularization of the term Web 2.0 in 2005

(O’Reilly). Since these early stages, the social soft-

ware objective has been named in nearly every

summary or outlook of important technologies, as

for instance in Gartner’s top 10 strategic technolo-

gies for the years 2007 to 2012 (cf. Gartner Inc.) or

in the IEEE Spectrum magazine’s top 11 technolo-

gies of the decade (placed as number two) (2011).

On the other side social software has a very broad

range and therefore Ebner and Lorenz (2012) de-

fined a three-dimensional cube represented by the

axis Identity and Network Management, Information

management and Information and Communication.

In this cube the best fulfilling social software are so

called social networks colloquially social media like

Facebook, Google+ and Twitter.

For marketers and organizations social networks

provide the opportunity to reach a broad audience

with the advance of direct communication to the

target group and a low spreading loss. On the other

side individuals get the chance to participate and

find new communicational possibilities for social

interaction as well as new ways to filter and assess

the massive overload of information mankind has to

cope with nowadays. Therefore it seems naturally,

that more and more sectors of human life – contain-

ing companies, non-profit-associations or govern-

mental and public organization – are entering this

field of operation: trying to use this software for

their own purposes and benefit of its indwelling

chances.

While the success of the first pioneers in social

networks is founded on their innovative and reckless

use during the boosting time of the upcoming social

media hype, today’s newcomers won’t benefit from

that boost anymore. Who tries to join the social

community now has to know about the characteris-

tics of the present situation and start an individual-

ized and conceptualized approach to establish a solid

performance in the whole social web (Evans, 2010).

This is also true assuming the fact that those who

haven’t risked starting in a totally new field of op-

portunities while there were no conventions or in-

herent standards will most probably not risk starting

in a settled system without concrete instructions.

608

Frey J., Ebner M., Schön M. and Taraghi B..

Social Media Usage at Universities - How Should it Be Done?.

DOI: 10.5220/0004515506080614

In Proceedings of the 9th International Conference on Web Information Systems and Technologies (FWP-2013), pages 608-614

ISBN: 978-989-8565-54-9

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

According to this, there is a pressing need of re-

search supplying approved and validated knowledge

of the given field of expertise as current resources

are most times based on personal experiences rather

than scientific evidence. As for example there is a

vast number of weblogs or other web resources,

giving advice for social media marketers, e.g. Por-

terfield (2012), Honeysett (2012) or Baer (2013).

This research study tries to supply verified an-

swers for the persons in charge of the social media

channels with a special focus on universities, giving

an insight into the prerequisites and requirements

and providing a set of recommendations for estab-

lishing and maintaining a stable and valuable per-

formance in the social web.

In a first step we analyzed the present field of so-

cial media performances of universities in an attempt

to identify major factors for user engagement and -

as a result - for efficient social media activities.

These factors are subsequently used to develop con-

crete social media strategies for the sector of aca-

demic institutions.

2 METHODS AND RESEARCH

QUESTIONS

2.1 User Engagement and the

Efficiency of Social Media

Activities

To find out what are the main elements of an effi-

cient and reasonable performance in social media for

the specific field of universities, we analyzed the

present approaches of engaging academic institu-

tions. Therefore all activities of a university’s profile

on a specific social media platform were document-

ed and evaluated in relation to their corresponding

user engagement.

User engagement is an important instrument to

measure the value and sustainability of social media

activities. First and foremost it shows how many

people actively perceived a piece of message and are

affected by it in one way or the other. Furthermore it

also defines the reach of the message, as generally

user engagement increases the spread of a posting

within the social media platform and – e.g. in the

case of Facebook – also the spread of future postings

through recommender algorithms that reward overall

site activity with further reach.

Hence, analyzing the characteristics of a post in

comparison to its yielded user engagement should

allow identifying important influencers of a post’s

success or failure, which has been one of the prima-

ry goals in this work. For this purpose, a list of post

characteristics is defined that might influence the

chance of reception and engagement of the stake-

holders. These characteristics have been used later

on for an accurate analysis of the collected data

aiming to identify the constituents of an efficient

social media post. All this finally is concluding in an

approach to develop a set of social media strategies

for academic institutions.

2.2 Research Questions

For the analysis of present social media activities,

the following research questions have been investi-

gated:

 Are there primary influencers for the en-

gagement of users in social media posts?

 Which influencers can be identified?

- Which characteristics determine relevant

content for messages in social media?

- Does the frequency of publishing social

media messages influence the user en-

gagement?

2.3 Present Social Media Activities

of Universities

Out of the various possibilities in social media, the

research study is strongly concentrating on social

networking sites, as they represent the primary ob-

jective of social media displaying social structures

and relationships and can be used to create and

maintain a perceived presence in social.

Within the range of social networking facilities,

the world’s best-known platforms Facebook,

Google+ and Twitter are chosen and a set of univer-

sities for every platform is defined, which should be

investigated further. The universities differ in terms

of language, origin, size and educational approach to

display a broad and representative view of universi-

ties’ present activities in the social web. The chosen

set contains US-American (e.g. Harvard University,

Ohio State University) and English sites (e.g. Uni-

versity of Oxford, University of Cambridge), as well

as sites from Germany (e.g. Goethe University

Frankfurt am Main, LMU Munich, Austria (e.g.

University of Innsbruck, Johannes Kepler University

Linz) and Switzerland (e.g. HSG – University of St.

Gallen, University of Basel). There are also a few

institutions with technical background (e.g. Massa-

chusetts Institute of Technology or Ilmenau Univer-

sity of Technology) and universities of applied sci-

ences (e.g. FH Joanneum, Cologne University of

SocialMediaUsageatUniversities-HowShoulditBeDone?

609

Applied Sciences) included in the analyzed data. A

basic rate of user engagement has been the prerequi-

site for the selections of universities to guarantee

reliable data within the examined time period.

2.4 Post Characteristics and Possible

Influencers

The research study tried to take into consideration a

various amount of possible influencers of an appro-

priate and engaging post by classifying the posting

and its characteristics into different categories.

These categories are mainly determined by the spe-

cial needs and use cases of universities to guarantee

appropriate standards for the further development of

social media strategies.

The examined aspects of a postings are charac-

terized by

 time of publishing

The time of publishing has been listed in local

time.

 addressed target group

The addressed target group contains the cate-

gories staff, students, future students and the

public.

 used elements

Used elements or components of a post can be

videos, pictures, text and hyperlinks. Beside

the influence of a single element also the im-

portance of the composition of these elements

is analyzed.

 message length

Here the number of characters used in a mes-

sage is analyzed.

 content characteristics (e.g. subject, function

and time reference of a posting)

Content characteristics contain the post subject,

function and time reference. Table 1 shows a list of

the specific categories defined for these characteris-

tics.

Table 1: Analysed content characteristics.

subject function time reference

• research

• study

• university

teams,

projects

• non

academic

• information

• promise

• interaction

• contact

• fun

• expression

of emo-

tions

• announce-

ment

• news/reports

(after event)

• seasonal

• serial

• without time

reference

Additionally we also tried to find a correlation

between the user engagement and the frequency of

published contents of the universities.

2.5 Measurement of User Engagement

on Facebook

During the examined time period, it became obvious

that Facebook can be seen as a sort of exemplary

standard for present social media efforts (further

details are explained in the next chapter). As a con-

sequence the study concentrated on a detailed inves-

tigation on the findings of the Facebook results. For

that reason a method was designed to measure user

engagement based on the characteristics of Face-

book’s edge rank.

To evaluate the user engagement of a given post

in comparison to other posts, it is necessary to ex-

plain the process and prerequisites of user engage-

ment in Facebook posts. An overview of the process

is also shown in figure 1.

Figure 1: The process of user engagement.

WEBIST2013-9thInternationalConferenceonWebInformationSystemsandTechnologies

610

Any published Facebook post has got different

preconditions concerning the possible engagement

rate. This is primary due to the reach of that specific

post.

There are different ways to reach the audience

within Facebook. The post does either reach them in

an organic, a viral or a paid way. While paid audi-

ence can only be reached by Facebook ads and pro-

motions, the viral reach, meaning how many friends

of fans did see that any of his or her friends engaged

with the site, depends on a high percentage on the

organic views of a Facebook post. The organic Fa-

cebook audience represents a subset of the site’s

Fans. This subset is determined by the Facebook

edge rank, a recommender algorithm assigning every

posting a certain value for every single user of Face-

book. The edge rank algorithm defines which users

see a posting in his/her information stream and

which do not.

According to different resources on this topic

and with due regard to the fact, that there are still

some settings, that are not issued in these resources

to keep some secret of the news stream composition,

the Facebook edge rank consists of the following

values (What is Edgerank?, 2013) and (Tarbaj,

2013):

 Affinity (How strong is the relationship be-

tween the site/posted content and the user:

how often does he/she interact with the site,

how many friends of the user do interact with

the site, is it a content the user typically tends

to interact with, etc.)

 Weight (Value to promote specific content in

comparison to other content types, the specif-

ic values are not commonly known for that

factor.)

 Time decay (How many time has passed since

the publication of a post.)

Considering all these factors, the possible reach

of a Facebook post is determined by:

 fans of a site

 the interaction rate (visible as the talk-about

count of a site)

 further settings of the Facebook edge rank,

which are not visible for users

To evaluate the Efficiency of a specific post p we

therefore calculated an estimated user engagement

est(p)considering the existing conditions of publica-

tion of that post and compared that value with the

actual reaction act(p) it has achieved.

Efficiency(p)= 100*act(p)

est(p)

(1)

Assuming recent statistics, that show a signifi-

cant decrease of interaction rates within large

fanbases (Jochemich, 2013) proving that a grand

amount of fans is influencing the overall interaction

rate in a substantial way, we covered this fact by

dividing the analyzed universities by scale into three

different types.

 universities with less than 5,000 fans

 universities with more than 5,000 but less

than 100,000 fans

 universities with more than 100,000 fans

For those types the average amounts of post reac-

tions (sum of likes, shares and comments) per fan

are calculated as well as the average amounts of post

reactions per talk-about to provide comparable rela-

tive factors for the estimation of user engagement.

These factors are shown in Table 2.

Table 2: Relative factors for the estimation of user en-

gagement.

scale reaction per

fan

reactions per

talk-about

>100,000 fans 0.0015 0.0626

5,000-100,000 fans 0.0021 0.0622

<5,000 fans 0.0076 0.1207

According to that, the estimated user engagement

can be calculated by estimating an overall reaction

with the actual fan and talk-about count of a posting.

Est(p)= (fans(p)*fanfactor(size(p))+talk-

about(p)*talk-aboutfactor(size(p)))/2

(2)

2.6 Field Study

The data, we gathered from university-posts on

Facebook, has been systematically documented and

categorized in regard to the above-mentioned post-

ing characteristics.

After that we evaluated the thus enriched data in

relation to the quantitative measurement of user

engagement described previously, therefore using

statistical techniques (predominantly correlation

methods, e.g. Pearson’s correlation, rank correlation,

etc.) to give significant and hence validated state-

ments. A logarithmic relativization has been made

over the values for efficiency as to balance out the

spread of engagement data, which has been very

inhomogeneously distributed over the given field.

3 RESULTS

For the analysis of present university approaches in

SocialMediaUsageatUniversities-HowShoulditBeDone?

611

Table 3: Overview of the analyzed data.

Facebook Google+ Twitter

universities 20 12 26

posts/activities 1.196 231 2.301

comments/replies 7.772 121 1.090

range of fans/followers 457-2.051.296 59-8701 422-163.578

range of average posting count 15-107 8-61 3-658

max. engagement (sum of interactions) 14.375 56 80

range of average engagement per university 5-2060 0,2-20,7 0,05-17

social media, all activities of the selected universi-

ties between 15.09.2012 and 07.11.2012 were gath-

ered. This results in over a thousand Facebook posts,

about 300 Google+ activities and about 2,500 Twit-

ter statuses. An overview of the gathered data is

given in Table 3.

The results show the unquestioned reign of Fa-

cebook in the social media sector. This leadership is

also stated by other research regarding Facebook’s

market share (StatCounter, 2013) or the usage of

communication behaviour in social media (Ebner

and Nagler, 2013).

Therefore it seems to admit of no doubt, that Fa-

cebook represents the state of the art of present so-

cial media usage within the university context. This

has been the reason for the decision to concentrate

our statistical research on Facebook as a generic and

exemplary social media platform. The following

results apply to the use of Facebook as a social me-

dia facility for universities, the overall survey of the

other two platforms observed in this study does

support these results.

According to the statistical evaluation the re-

search questions can be answered as described in the

following chapters.

3.1 Are there Primary Influencers for

the Engagement of Users in Social

Media Posts?

The examined posting characteristics have been

various. The user engagement doesn’t correlate with

a single factor of these characteristics. It is more or

less the composition of characteristics that defines a

good and efficient post. Nevertheless there have

been characteristics that seem to be more important

than others, with some of the categories having a

significant correlation to the achieved user engage-

ment while others have not.

3.2 Which Influencers can Be

Identified?

Having those characteristics in mind, which have

shown a significant correlation to the user engage-

ment, the following influencers are identified:

 Time of publication

 Used elements

 Content characteristics (containing subject,

function and time reference)

Time of Publication

Weekend and night posts get more user reaction than

others, while morning posts get less.

Used Elements

The best combination of post-elements is sharing a

text with a photo and a link. Videos don’t seem to

have a significant effect on the user engagement. For

each of the other elements the results show that posts

with a text, link or picture have better reactions than

those without.

There is also a negative correlation between user

engagement and uncommented pictures as well as

messages without visual elements.

3.2.1 Which Characteristics Determine

Relevant Content for Messages in

Social Media?

Analyzing the content of the posts, the study shows

that the content characteristics are highly related to

the user engagement.

Posts with the inherent function of contact have

higher engagement rates than all others. In second

place fun posts get good reactions too.

While the research subject doesn’t get good reac-

tions, posts subjected to the university as a place to

live and work in, as source for user identification, as

well as posts subjected to a special group of the

community like sport teams, clubs, etc. get signifi-

cant higher engagement rates.

WEBIST2013-9thInternationalConferenceonWebInformationSystemsandTechnologies

612

Regarding the inherent time reference of a post,

messages with seasonal content and messages with

recurring subjects like weekly sequels tend to have

higher user engagement than messages announcing

or reporting events.

3.2.2 Does the Frequency of Publishing

Social Media Messages Influence the

User Engagement?

The user engagement rates don’t show an unambig-

uous disposition in terms of the frequency of post-

ings of a university. There are slight differences

indicating that more than 3 posts a day relate to a

lower engagement for further posts. It also seems

reasonable, that less than one post per week leads to

a lower engagement rate as the interaction value

included in the Facebook edge rank is calculated

over the period of a week.

3.3 Comparison of University Efforts

The efficiency of university social media efforts has

been evaluated with the fan base, the recent interac-

tions (talk-about count of Facebook) and the overall

size of the site an estimated interaction rate in mind,

to provide a field of data, where single postings can

be compared to others, even if they are made by

another university with different prerequisites. Still

there are significant differences between the univer-

sities. Some of the universities obtain consistently

higher user engagement rates compared to their

prerequisites and seem to perform notably better

within the social media.

For example the Ohio State University also in-

vestigated in the study has got about 500 thousand

fans at the beginning of the research period, hence

gaining an average user engagement of ca. 1590

likes, shares and comments on a single post. Where-

as bigger universities like Harvard University with

about 2 million fans (four times the amount of the

Ohio State University) just achieved an average of

ca. 2060 reactions, or the University of Oxford with

ca. 650 thousand fans (33% more than the Ohio

State University) attained just an average of ca. 370

user reactions.

Observing the profiles of the universities, that

did attain a better user perception we also recog-

nized, that most of all they are strategically using the

aforementioned posting characteristics, while others

do not really vary their approaches and are chiefly

posting research information through their social

media channel.

This also confirms, what has been discovered

earlier.

4 DISCUSSION AND SOCIAL

MEDIA STRATEGIES

This chapter combines the findings of the statistical

analysis of Facebook and the investigation of col-

lected data from Twitter and Google+ in order to

develop recommendations for an efficient social

media strategy in academic environments.

During the study it became obvious that a uni-

versity’s social media account can’t be filled fully

automatic by periodically publishing news of recent

research efforts or upcoming events. The results

show clearly, that there is less interest on these top-

ics in the social web. Users might take other media

to inform themselves of these contents. Moreover it

seems of relevance to deliver an accurate setting of

content, taking into account the social characteristic

of this media.

Therefore it is of great importance to publish

contents that raise and maintain contact between the

university and their stakeholders on a personal,

maybe even emotional way (e.g. by using visual

elements) which is proven by the higher engagement

rate of the posts containing these characteristics.

Below we suggest a few examples how that

could be achieved according to the results of engag-

ing post characteristics:

 Presenting the university as a primary place

to live and work in; (cf. post subject)

 Supplying opportunities to keep in contact

with the university and the community, e.g.:

referring to the history of the university, well-

known places on the campus, community ac-

tivities and collective community knowledge;

(cf. post subject and post function)

 Accomplishment of social conventions, e.g.:

congratulations, seasonal greetings, acknowl-

edgements; (cf. post subject and post func-

tion)

 Providing Information, which is up-to-date,

even if this means posting in the night, at

weekends or holidays; (cf. time of publishing

and time reference)

 Acknowledgement of achievements and

providing contact opportunities within the

community, e.g. congratulating student

teams, sport teams, achieved awards; (cf. post

subject, post function)

SocialMediaUsageatUniversities-HowShoulditBeDone?

613

 Enabling user participation, where it isn’t

usual throughout other media, e.g.: topics of

common interest, feedback possibilities; (cf.

post function, time reference, post subject,

etc.)

 Providing linking possibilities throughout se-

quels and recurring topics; (cf. serial posts)

 Posting some contents just for fun; (cf. fun

posts)

 Avoiding information about research and

event announcements or reports

5 CONCLUSIONS

Over all it can be pointed out, that it is good to sup-

port the ties of the community and to plead on loyal-

ty and community spirit. Various contents should be

used to get the advances of a diversified and rich

social media performance by activating as much of

the related users as possible. This study also seeks to

prevent universities of overrating posting character-

istics that have shown no significant relationship to

the user engagement, for example the posting fre-

quency or the text length of a message. Especially

because the text length has previously been stated

important by other studies like the Buddy Media

Data Report (2012) or the Facebook content study of

Reimerth and Vigand (2012) both investigating the

use of Facebook pages for marketing and sales ob-

jectives of businesses. Therefore further research has

to be taken, to clarify if aspects like the higher ac-

ceptance of longer text messages are only true for

the sector of higher education or can be applied to

other non-profit-organizations too. Regarding the

other results – part of them also mentioned in the

referred studies – it is reasonable, that these results

are also reliable for a social media strategy for non-

academic institutions due to the fact that the inherent

business content of universities (research news) is

less important to become a good social media player.

REFERENCES

Baer, J. (2013). Social Media Strategy in 8 Steps. Re-

trieved March 4, 2013, from http://

www.convinceandconvert.com/social-media-strategy/

social-media-strategy-in-8-steps/

Buddy Media (2012): Strategies for Effective Wall Posts:

A Timeline Analysis. Retrieved March 21, 2013, from

http://marketingcloud.buddymedia.com/whitepaper-

form_review-strategies-for-effective-facebook-wall-

posts-a-timeline-review

Ebner, M. and Lorenz, A. (2012). Web 2.0 als

Basistechnologien für CSCL-Umgebungen. In Haake,

J., Schwabe, G., &Wessner, M. (Eds.), CSCL-

Lernumgebungen. München: Oldenburg.

Ebner, M., Nagler, W. and Schön, M. (2012). Have They

Changed? Five Years of Survey on Academic Net-

Generation. Proceedings of World Conference on Ed-

ucational Multimedia, Hypermedia and Telecommuni-

cations 2012, 343-353. Chesapeake, VA: AACE

Ebner, M., Nagler, W. and Schön, M.

(2013).“Architecture Students Hate Twitter and Love

Dropbox” or Does the Field of Study Correlates With

Web 2.0 Behavior? ED-Media Conference, 2013, ac-

cepted, in print

Evans, D. (2010). Social media marketing. The next gen-

eration of business engagement (1st ed.). Indianapolis:

Wiley Publishing.

Gartner, Inc. Gartner Identifies the Top 10 Strategic Tech-

nology Trends. Retrieved March 2, 2013, from

http://www.gartner.com

Honeysett, A. (2012). Planning Your Social Media Strate-

gy? Take a Cue from GE. Retrieved March 2, 2013,

from http://www.forbes.com/sites/dailymuse/2012/12/

03/planning-your-social-media-strategy-take-a-cue-

from-ge/

IEEE Spectrum (2011). Special Report: Top 11 Technolo-

gies of the Decade. Retrieved March2, 2013, from

http://spectrum.ieee.org/static/special-report-top-11-

technologies-of-the-decade

Jochemich, M. (2013). Is Your Facebook Page Performing

Above The Average For January 2013? Retrieved

March 2, 2013, from http://www.quintly.com/blog/

2013/02/infographic-facebook-page-performance-

2013-01/?utm_source=Infographic&utm_medium

=Infographic&utm_campaign=20130219_FB_PageBu

ckets_01_2013&r=25

O'Reilly, T. (2005). What Is Web 2.0. Retrieved February

25, 2013, from http://oreilly.com/pub/a/web2/archive/

what-is-web-20.html?page=1

Porterfield, A. (2012). 3 Steps to an Effective Social Media

Strategy. Retrieved March 3, 2013, from http://

www.socialmediaexaminer.com/3-steps-to-an-

effective-social-media-strategy

Reimerth, G. And Wigand, J. (2012). Welche Inhalte in

Facebook funktionieren. Facebook Postings von Con-

sumer Brands und Retail Brands unter der Lupe. Re-

trieved March 21, 2013, from http://www.knallgrau.at/

facebookcontentstudie

StatCounter(2013).StatCounter Global Stats. Top 7 Social

Media Sites in Europe on Feb 2013. Retrieved March

3, 2013, form http://gs.statcounter.com/#social_media-

eu-monthly-201302-201302-bar

Tarbaj, C. (2013).Understanding Facebook’s EdgeRank.

Retrieved March3, 2013, from http://

socialmediatoday.com/ node/1205641

What Is EdgeRank?

(2013). Retrieved 3 March, 2013,

from http://www.whatisedgerank.com/

WEBIST2013-9thInternationalConferenceonWebInformationSystemsandTechnologies

614