S M S

possehastyΜηχανική

5 Νοε 2013 (πριν από 3 χρόνια και 11 μήνες)

89 εμφανίσεις

S
OCIAL

M
EDIA

S
CORECARDS

C
HAPTERS

07 & 08
OF

S
OCIAL

M
EDIA

A
NALYTICS

W
/
OTHER

M
ATERIAL

1

Ravi K. Vatrapu

Director
, Computational Social Science Laboratory (CSSL)

Associate Professor
,

Dept. of IT Management

Copenhagen Business School

Phone: +45
-
2479
-
4315

Email:
vatrapu@cbs.dk


(preferred)

Web:
http://www.itu.dk/people/rkva/


Zeshan
Jaffari

zeshan.jaffari@gmail.com


Friday, 04
-
November
-
2011

T17: SMA:
Lecture 13

ITU: Auditorium 3, Copenhagen, Denmark

S
OCIAL

M
EDIA

S
CORECARDS


Provide


Monitoring


Measurement



Include


Social Data


Web Analytics


Organic Search Queries



Also Include (
Local

Social
Mobile
)


QR Code


Mobile Check
-
Ins


House Data




2

S
CORECARDS

& A
UTOMATION


Meaningful & Actionable



Coding: Manual vs. Automatic



Time + Expense vs. Efficiency + Accuracy




Scorecards as Automations


But then: What is Automation?

3

Definition of
Automation

Chapter 16 of “An Introduction to Human Factors Engineering (Second Edition)”


4

“Automation characterizes the circumstances when a machine
(nowadays often a computer) assumes a task that is otherwise
performed by the human operator” (p.418)



Aviation Industry Joke: Dog and Pilot in the Cockpit



Home Automation


Office Automation


Factory Automation



Social Media Analysis Automation





Reasons for Automation

5


Impossible or Hazardous




Difficult or Unpleasant




Extend Human

Capability




Technically Possible






Stages of Automation

6

Stage 1
: Information Acquisition
, Selection, and Filtering



Stage
2:
Information Integration



Stage
3:
Action Selection and Choice



Stage
4: C
ontrol and Action Execution

Levels of Automation

7

From Complete
Manual

Control to Complete
Automatic

Control


1.
Automation Offers No Aid


2.
Automation Suggests Multiple Alternatives


3.
Automation Selects an Alternative


4.
Automation Carries Out an Action if the Person Approves


5.
Automation Provides the Person with Limited Time to Veto Action before it
Carries Out the Action


6.
Automation Carries Out an
Action and then Informs the Person


7.
Automation Carries Out an Action and
Informs
the
Person Only if Asked


8.
Automation Selects
M
ethod, Executes
T
ask, and Ignores the Human




Problems in Automation

8


Automation Reliability



Calibration and Mistrust



Overtrust

and Complacency



Workload and Situation Awareness



Training and Certification



Loss of Human Cooperation



Job Satisfaction

Functional Allocation : Person and Automation

9


Fitts

List

http://
www.cse.sys.t.u
-
tokyo.ac.jp/furuta/teaching/csd/CSD06.pdf



Progress of Automation

10

http://
www.cse.sys.t.u
-
tokyo.ac.jp/furuta/teaching/csd/CSD06.pdf



Principles of Human Centered Automation

11

1.
Keeping the Human Informed


2.
Keeping the Human Trained


3.
Keeping the Operator in the Loop


4.
Selecting Appropriate Stages and Levels When Automation is
Imperfect


5.
Making Automation Flexible and Adaptive


6.
Maintaining a Positive Management Philosophy

Implications for Social Media Scorecards

12


Participant Observation



Qualitative Analysis



Interpersonal Communication



Establishing “Ground Truth”



Coding + Counting




Three Phases of Social Media Scorecards

13


Data Gathering



Data Analysis



Data Reporting



Examples of Social Media Scorecards

14


Ogilvy Conversation Impact



Razorfish

Fluent



Zocalo

Group’s Digital Footprint Index

Ogilvy Conversation Impact

http://blog.ogilvypr.com/2009/06/introducing
-
conversation
-
impact
-
social
-
media
-
measurement
-
for
-
marketers
/


15


Reach and Positioning


Unique monthly visits


Time on site


Overall volume


Share of voice within category or brand family


Search
visibility



Preference


Sentiment
index (% positive
-

% negative) in social media


Share
of positive voice in social media, within category



Relative
Net Promoter Score, absolute or within
category



Action


Registrations


Sales


Advocacy

Razorfish

Fluent

http://fluent.razorfish.com/publication/?
m=6540&l=1



16


Social Influence Marketing (SIM)



Reach


Sentiment




Social
ads
are
“about
infusing social content and a user’s social graph
directly
into the
ad unit itself
.”




Social
graph
is “the
network of personal connections through which
people com
municate and share information online. These personal
connections can be based on common interests, professional experiences
and offline social relationships.”


Zacolo

Group’s Digital Footprint Index

http://zocalogroup.com
/


17


Volume
of
Conversations



Location of Conversations



L
evel
of
Engagement



M
essage Adoption



Tonality




Height + Width + Depth

Example of DFI

18

http://www.webmetricsguru.com/archives/2009/09/a
-
social
-
media
-
scorecard
-
based
-
on
-
digital
-
footprint
-
index/

Questions to
Clien

19

http://
vigilmetrics.com/App_Themes/Default/images/gallery/Social%20Media%20Scorecard.PNG


Questions to Clients

20


General Requirements



Specific Requirements



Keywords and Volume Analysis



Reporting



Social Media Maturity of the Clients

21

Level 1
: Monitoring


Level 2
: Online Research


Level 3
: Social Targetting and Data Management


Level 4
: Social Business Collaborations

Three Analytical Phases

22


Culling



Classifying



Contextualizing



Q&A for the Social Media Analyst

23

1.
Which organizational unit is your
analytics report
oriented towards?


2.
How sophisticated is the case company’s industry regarding social
media?


3.
How mature is the case company when it comes to socia media
marketing, analytics, and management?


4.
What is the budget, if any , allocated for
socia media marketing,
analytics, and management
?


5.
How much time is allocated to deliver results?


6.
How much integration with other data sources?


7.
What are the languages and regions for monitoring and measuring?



Discussion

24