Communication Practitioners’ Perceptions of Big Data and Automation: A Comparative Study between Europe and Latin America Percepciones de los Profesionales de Comunicación sobre Big Data y Automatización: Un Estudio Comparativo entre Europa y Latinoamérica

The big data revolution has changed the way organisations operate. The implications have been phenomenal for public relations and communication management professionals who are trying to understand and manage the realm of big data and what it means for them. This study is an attempt to dive deeper into the discussion on how professionals are managing the world of big data. A large survey of European and Latin American countries reveals comparative findings on the knowledge and usage of big data and automation and demonstrates large gaps between the continents. Implications for theory and practice are finally drawn.

están tratando de comprender y gestionar el ámbito del big data y lo que significa para ellos, han sido relevantes. Este estudio trata de profundizar en la discusión de cómo los profesionales están gestionando el mundo del big data. Una amplia encuesta en países europeos y latinoamericanos revela resultados comparativos sobre el conocimiento y el uso del big data y la automatización y pone de manifiesto grandes brechas entre los dos ISSN: 2174-3681 The term Big Data is used to describe, "the overwhelming volume of information produced by and about human activity, made possible by the growing ubiquity of mobile devices, tracking tools, always-on sensors, and cheap computing storage" (Lewis et al., 2013, p. 2).
However, big data is mostly defined by its four "V"s: Volume, Velocity, Variety, and Value (Gandomi & Haider, 2015). The amount of data and its granular nature describes the volume of data. The speed with which an organisation receives data and needs to handle them in real-time outlines the velocity characteristic of big data. The unstructured, structured, and semi-structured data variety affects the data requirements and how organisations might summarise and analyse the data (variety & value). Therefore, we can assume the definition of Wiesenberg et al. (2017, p. 96): "Big data denotes huge volumes and streams of different forms of data from diverse internal and external sources and their constant processing, which provide new insights".
The intrinsic value of data helps organisations to derive meaning from the data, recognise patterns, and make informed assumptions in their decision-making. The terminology might change in the coming years but the need to strategize, collect, and analyse data will remain a top priority for organisations and their management (Wamba et al., 2017;Yaqoob et al., 2016). In its early stages, organisations focused primarily on the amount of data (volume) and how to manage the data streams. However, the question shifted from the size of data to its importance and the value derived from the data itself. At this stage, many organisations struggle with big data because it needs further analytical tools, skills, structures as well as resources (Fan et al., 2015). Applications like large-scale text analysis, such as automated content analysis, data mining, machine learning, topic modelling and sentiment analysis are still are uneasy to make them accessible for certain fields (Arcila-Calderón et al., 2016). The next section approaches some of the big data challenges as well as the different kinds of data that one needs to be familiar with.

Big Data Challenges
Given the nature of big data, there are multiple challenges that organisations and professionals have to deal with. In a conceptual classification of big data challenges, Sivarajah et al. (2017) define data challenges as the challenges related to the characteristics of the data itself, e.g. volume, velocity, variety, variability, veracity, visualisation, and value.
The main challenges are faced while processing the data like data acquisition and warehousing, data mining and cleaning, data aggregation and analysis as well as modelling applications.
Furthermore, there are management challenges related to privacy, security, governance, data ownership, and lack of skills of understanding and analysing data (Camargo Fiorini et al., 2018).
The challenges of big data not only lie within its scale of complexity, but also within issues like heterogeneity, timeliness, and even privacy problems (Holtzhausen, 2016). These aspects heighten the challenges in creating value from the data. The huge volume of data is represented by heterogeneous and varied dimensions. In the same vein, the huge volumes of data also increase the complexity and the relationships within the data. The complexity and the disparate origins of the data often result in incomplete or flawed data. In order to avoid this, it is important to understand how structured and unstructured data work.

Structured and unstructured data
One of the biggest challenges when it comes to big data is the integration of structured and unstructured data. According to Taylor (2018), about 80% of data held by an organisation is unstructured data, comprised of information from customer calls, e-mails and social media feeds. Since data in communication management practice increases significantly in a digital format, there is a greater need to identify ways to link the data and transform data for analysis. Unstructured data continues to grow, and organisations have to find ways to automate and improve their ability to understand their business.
The problem that unstructured data presents is one of volume; most business interactions are of this kind, requiring a huge investment of resources to sift through and extract the necessary elements, as in a web-based search engine. Hence, it is key to find ways for meaning creation by using connections that demonstrate specific patterns. Analysing unstructured data requires analytical tools and newer approaches based on machine-based learning. Machine-learning approaches can help analyse complex large volumes of data, both structured and unstructured, with multiple variables to make accurate predictions. The question is not whether the data should be unstructured or structured, but rather how to use its internal value in a meaningful way. The paper addresses the use of big data by communication professionals in Latin America and Europe by posing important questions around big data skills and knowledge, attention to the debate about big data, the communication professionals' familiarity with the concept of big data, and big data expertise among communication professionals.

Research gaps and research questions
As identified in the literature review, there is still a knowledge gap from a global perspective.

METHODOLOGY
The research study is based on a quantitative survey among communication practitioners in Europe and Latin America Zerfass et al., 2016). The questionnaire included a special section about big data and automation, which covered six questions derived from the literature review above. The online questionnaire was made available throughout March 2016 in Europe in English language, as well as in May and August 2016 in Latin America in Spanish and Portuguese language. In Europe, more than 100,000 personal invitations were sent to communication professionals working in all kinds of organisations in all 50 European countries via e-mail. In total, 3,287 respondents completed the questionnaire and 2,710 responses could be identified as communication professionals, which were used for the study at hand. Most respondents (28%) came from Northern Europe (Scandinavia and the British Isles), followed by Central Europe (19%), South-eastern Europe (18%), Western Europe (15%), Southern Europe (14%), and Eastern Europe (7%) 4 .
In a similar way, communication professionals in Latin America were surveyed between May and August in 2016. More than 20,000 communication professionals in Latin America were invited to participate in the online survey through datasets of national and regional professional associations. In total, 2,295 respondents started the survey and 914 respondents from 17 countries could be identified as communication professionals that filled out the complete survey. Most respondents (74%, n = 675) came from South America (Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, Peru, Paraguay, Uruguay, and Venezuela), followed by Central America (16%, n = 143, Costa Rica, El Salvador, Guatemala, Honduras, Panama, and Dominican Republic), and North America (11%, n = 96, Mexico) 5 .
Further details of the samples are depicted in table 9 (see appendix).

RESULTS
The following chapter demonstrates the findings of the comparative study in order to gain new insights in the field of big data and automation in communication management on a global scale.

Big data perceptions, knowledge and skills (RQ1 -RQ3)
The literature review indicated that digitisation and therefore big data and its analysis have and team members or consultants (M = 3.15, SD = 1.28). As Table 1  Note. Standard deviations appear in parentheses below means. N = 3,505 communication practitioners. Q: Please rate these statements based on your experience. 5-point Likert scale ranging from 1 = "I have not given attention at all to the debate about big data" to 5 = "I have given close attention to the debate about big data". * Highly significant difference at the p ≤ .001 level based on independent samples t-test, t(3503) = 6.91. As already outlined in the literature review, the definition in this research project derived from the four V's ("mass quantities of stored data that provide new insights which were previously not available" = Volume; "a variety of multiple data types from internal and external sources" = Variety; "a fast stream of data (data in motion) and their constant processing" = Velocity; "high and low quality data from trusted and untrusted sources" = Veracity). In order to gain insight into the cognitive dimension of the professionals, this study used the big data definition and also four deflectors that are related to the topic but did not represent the concept of big data ("customized creation of content for different stakeholders", "interpretation of relevant data for strategic decision making", "all kinds of information which is available in real-time", "a multitude of information from social media").
Those polled were asked to choose all appropriate definitions of big data. Only 0.9% classified all eight items correctly (as either appropriate or wrong) and 5.8% classified seven out of eight correctly. This summed up to 6.7% who can be categorised as highly knowledgeable. In stark contrast, 11.1% mixed up almost everything and thus does not seem knowledgeable at all. However, the majority is somehow or moderately familiar with the concept of big data (see table 2).
Note. N = 3,624 communication practitioners. Q: "Big data" is characterised in various ways. Please pick all definitions which you believe are most appropriate. Big data refer to … * Including "None of these" / "I don't know" (n = 86). Highly significant differences at the p ≤ .001 level based on Chi-square test, Cramer's V = .233. the strongest correlation exists between the expertise cluster and the overall social media skills based on the self-evaluation of twelve items (r = .11, p ≤ .001).  5-point Likert scale ranging from 1 = "Very low" to 5 = "Very high", overall value based on a battery of 12 items.

Big data analytics and the need of automation (RQ4 -RQ5)
In total, 19.1% of the respondents declared that their organisation has implemented big data activities in the communication field. In 15.1% of the cases, the departments or agencies planned to start such big data activities by the end of 2017, while 42.2% indicated that their department or agency is not conducting big data activities, and 7.7% stated that they do not know how their organisation handles the issue. Moreover, 20.5% report that their agency or department consults (internal) clients and colleagues in the field of big data. The analysis revealed statistically significant differences for the current status of big data activities in the two regions (see Table 5 In both regions, only a minority has already implemented big data activities as described above. However, as demonstrated in Table 6, big data analytics are mostly used to plan overall strategies (e.g. to use insights to guide future campaigns or for foresights). However, big data analytics is also frequently used to justify activities (e.g. by measuring results and demonstrating effectiveness). Analysing big data to guide day-to-day actions (e.g. targeting publics with specialised content) is also used more frequently in Latin America, compared to Europe.
Note. N = 1,320 communication practitioners (including only respondents who have already implemented big data activities and/or consults clients and colleagues). Q: How does your department or agency use big data analytics? 5-point Likert scale ranging from 1 = "Never" to 5 = "Always". * Significant differences at the p ≤ .05 level. ** Highly significant differences at the p ≤ .001 level. Strongly connected to the topic of big data is the question of automation, as it is necessary at least for big data analytics (to make sense of big data). Table 7 presents an overview of the differences between the two regions regarding the importance and the implementation of algorithms in public relations. All results are strongly linked (correlated) with the implementation of big data activities as well as big data analytics.
Note. N ≥ 3,111 communication practitioners. Q: Search engines and social media platforms use algorithms to select and display content. Similar approaches might be used by organisations to automate their communication activities. What is already used by your department/agency? 5-point Likert scale ranging from 1 = "Not at all important" to 5 = "Extremely important". And what is already used by your department/agency? * Highly significant differences at the p ≤ .01 level based on independent samples t-test, t(3167) = 2.84. ** Highly significant differences at the p ≤ .001 level based on independent samples t-test, t(3205) = 7.73.

Transformation of the field (RQ6)
The effects of big data on the PR field are seen quite contrary. While the European professionals is convinced that big data will change the communication profession (M = 3.85, SD = 0.83, N = 2710; 5-point Likert scale ranging from 1 = "will not change at all" to 5 = "will substantially change"), professionals in Latin America do not agree with this impression in general (M = 2.31, SD = 1.29, N = 774; highly significant differences at the p ≤ .001 level based on independent samples t-test, t(3482) = 39.53). The results reported so far demonstrate the importance of big data and automation for the communication profession on the one hand and a very diverse view on the other hand.
However, the implementation of big data, its analytics as well as automation imply some intrinsic challenges and even risks especially for communication management that works for trust of the entity in the public sphere. Hence, the findings will end with the question about the major challenges when working with big data from the perspective of PR professionals working in Europe and Latin America (see table 8  The findings demonstrate the low importance and attention given to big data and automation in both regions. Especially in Latin America, the inter-correlations between attention and understanding of big data as well as individual social media skills demonstrate that the differences between these two regions is a question of individual competencies and personal interest in innovation. However, other studies comparing these two regions regarding innovation also explicitly refer to the level of innovation of the company as whole (see e.g., Raffo et al., 2008).
Nevertheless, those who have already implemented big data activities in Latin America use them to a higher extent, compared to Europe. This represents a significant difference

LIMITATIONS
The study is exploratory in nature and provides an insight into the spreading of big data and automation in public relations from a comparative cross-cultural perspective. However, the author(s) understand that there are some limitations in the study that need to be addressed.

One of the main limitations is that the total population of practitioners both in Europe and
Latin America is unknown, along with the voluntary answer to the survey. The second limitation is that the study only provides panoramic differences between the two regions in 2016 and does not analyse the countries' specifics in-depth (e.g., developing, emerging, and developed countries). Further studies are required that dive deeper on the organisational level on the one hand, as well as studies specified in specific countries on the other hand.
For instance, in the case of Latin America, the diversity of economic, political and social context could generate interesting results.

IMPLICATIONS AND FUTURE PERSPECTIVES
The author(s) undertook this comparative study to understand the world of big data and automation especially in regions other than North America and Europe. Hence, this study provides great implications for theory and practice alike.
Communication practitioners can use the results of the study to better understand the use of big data and automation and the implementation rate of big data activities and algorithmic tools. The results can also be used to understand the differences and similarities in the two regions regarding acceptance and implementation of such new tools, structures etc. in these regions. As questions have been very general, further insights into the activities and tools they use and how they assess these tools in their own department/agency as well as in between departments are required. In addition, from a communication management perspective, it must be asked to what extend these activities and tools are already implemented in a sound management process.
Future research needs to explore the impacts and challenges of big data and automation in the professional field. It would be interesting to ask questions about how the communication profession might become more data-driven and to what extent this needs to be addressed profoundly for the field of communication management. Thus, scholars also need a comprehensive understanding to come to grips with algorithms and big data to gain deeper insights in the impact of big data and algorithms used by PR for inbound and outbound reasons, wherefore this study hopefully could be inspiring.
To conclude, two big challenges have been identified for the professional field: On the one hand, producing valuable insights for communication from structured and unstructured data seems challenging, and, on the other hand, educating communication practitioners poses challenges as well. The author(s) see this study as a starting point to a bigger discussion and research topic for the field of communication management not in a specific region, but on a global scale.