1. CSR communication strategies
for organizational legitimacy
in Social Media
Marcin Szewczyk & Elanor Colleoni
Danish Technical University
Copenhagen Business School
3. Defining Corporate Legitimacy
• The creation of a “congruence between the social
values associated with or implied by [organizational]
activities and the norms of acceptable behavior in
the larger social system” (Dowling & Pfeffer, 1975:
p.122) lies at the core of the legitimacy of business in
society.
• Obtaining legitimacy by “aligning corporate behavior
with stakeholder expectations” (Dawkins, 2004:
p.108) is necessary to guarantee the corporation’s
continued existence.
4. Research Question-
RQ1
Theoretical Question: Is there an
alignment between companies CSR
agenda and stakeholders social
expectations?
Empirical Question: Do the companies
and their audiences discuss about
similar arguments and with similar
vocabularies?
5. Research Question-
RQ2
Theoretical Question: Do the different
communication strategies lead to
different convergence outputs?
Empirical Question: If we identify the
topics of discussion, do the companies
and their audiences share the same
attention and opinion towards the
topics?
6. Data-
• Why in social media?
• We use Twitter from 2009
• We identify the CSR accounts of seven companies based
on the 100 Best Corporate Citizens 2009 list redacted by
Corporate Responsibility Magazine (2009) :
Starbucks, Ford, SmarterPlanet, Microsoft, Green Mountain
Coffee Roasters, Campbell, Xcel Energy
(Thanks to our colleague Michael Etter )
• We download the networks of these companies (i.e. the
online audiences) and their conversations… about 10K
users and 300M tweets
7. Method– Identifying the
communication strategies in Twitter
Corporate
@DIRECT
Communication @RETWEET
messages
Flow
Strategies Reciprocity
No direct
Self-Centered No No
communication
Mediated Experts Experts Experts
Dialogical Audience Audience Experts Audience Experts
8. Method- Identifying the topics of
discussion
• We first selected the CSR-related tweets from
the general conversations using a Bayesian
Classifier
• We use Latent Semantic Analysis in order to
extract the “conceptual spaces” that
represent the “topics” (i.e. CSR themes).
9. Method- Identifying the attention
and the sentiment towards the topics
• We define the attention as the relative amount of
conversation generated by a topic
• We use sentiment analysis to identify the affective
orientation towards a topic (i.e. opinion).
Sentiment analysis is a data mining techniques that
allows subjective perceptions to acquire an objective
existence as observable and measurable forms (Colleoni
et al., 2011).
10. Results- We didn’t find evidence of
the mediated strategy….
Degree of Reciprocity
Features Degree of Conversation
with
Communication
Experts Audience Strategy
Corporation Experts Audience @DIREC @DIREC
@RT @RT
T T
Ford 95.11% 94.63% 0,00% 1,12% 7,87% 8,99% Self-centered
Starbucks 94.46% 89.25% 0,67% 0,33% 44,15% 5,02% Dialogical
Smarter Planet 82.05% 70.24% 0,00% 3,17% 17,46% 36,51% Dialogical
Green
Mountain 61.80% 45.15% 1,08% 0,22% 43,87% 12,47% Dialogical
Coffee Roasters
Xcel Energy 41.66% 33.61% 8,15% 0,00% 11,11% 16,30% Self-centered
Campbell 23.07% 14.50% 2,38% 0,00% 2,38% 9,52% Self-centered
Microsoft 0% 1.90% 1.08% 3.60% 1,80% 31,65% Self-centered
11. Results- We found six topics of discussion
between companies and audience
Topics
Theme Keywords
Information seeking Power, wind and solar associated
Energy & green & us
with free, look, share
Information seeking Energy and solar with thank, share,
Thank & Share
great, day
Green, solar, energy but both
Sustainable Debate Energy & Solar & Renew without associations with news or
business
Sustainable Debate
Health & Care Us, health, energy, solar, care
Business, make, market, work,
Market Free & business & make blog
Twitter, support, show, Iran,
Democracy Democracy & support
democracy
12. Results- The vocabulary of the companies
does not vary among strategies, but….
Dialogical Strategy: Most frequent words for Companies and Audience
Companies Audience
13. Results- The vocabulary of the companies
does not vary among strategies, but….
Information Strategy: Most frequent words for Companies and Audience
Companies Audience
14. Results- Is there congruence between
companies and audience?
Informational Strategy- Attention and Affective orientation’s
Congruence by topic, Audience and Companies
0,4
0,35
0,3
0,25
Sentiment
0,2
Companies
thank & share
solar & renew
Audience
0,15
health & care
0,1 energy & green & us democracy
business & make
0,05
0
0 1 2 3 4 5 6 7
Topics
15. Results- Is there congruence between
companies and audience?
Dialogical Strategy- Attention and Affective orientation’s
Congruence by topic, Audience and Companies
0.45
0.4
0.35
0.3
solar & renew
Sentiment
0.25
2
0.2
Companies
0.15 Audience
health & care
0.1 democracy
1
business & make
0.05
0
0 1 2 3 4 5 6 7
Topics
Notes de l'éditeur
Corporate legitimacy has become a pressing issue inasmuch as the stakeholders’ perception of the role of business in society has been significantly redefined. New ethical expectations have risen along with a set of contingent social responsibilities that corporations are now asked to fulfill by the various groups of stakeholders within the society …
We want to investigate our hypotheses in an online environment where companies can choose which strategies they want to follow..
For instance, the term “sand“, may be occurring in same documents as “beach”, therefore LSA model trained on such corpus may return us documents which contain term “beach” and do not contain term “sand” if such documents were close in the “concept space”.