Chapter 33

The Public Relations Power of “Big Data”

Simon Collister

According to IBM 90% of the world's data has been created in the past two years, due to social media's rapid growth.280 This chapter argues “Big Data” has practical applications for PR and can secure a future for the industry.

Public relations, according to the CIPR, involves “the planned and sustained effort to establish and maintain […] mutual understanding between an organization and its publics”.281 The context within which these relationships must be established and managed, however, is highly fluid.

As PR academic Anne Gregory points out “society is changing: new issues and trends arise, some of them very quickly.”282 For PR practitioners, being able to identify and track such trends, recognize their potential impact and adapt communication strategies accordingly is a daily challenge. This challenge is made all the more complex by our contemporary networked society where the proliferation of social technology, an increased willingness to share information and the empowerment of individuals is transforming political, economic, social and technological environments faster than ever before.

This chapter will suggest that the ever-increasing flow of information driven by socially networked individuals – increasingly referred to as “Big Data” – is both the catalyst for such breakneck dynamism as well as a potential solution to help PR practitioners manage this rapidly changing organizational environment. The chapter will give an overview of Big Data and how it presents PR practitioners with strategic and practical opportunities, using a case study of Big Data analysis within the Financial Ombudsman Service. It concludes by arguing that, if adopted and applied pragmatically, Big Data could potentially establish PR as a lead function within organizations and the wider communications, marketing and management sectors, thus helping secure a prosperous future for the industry.

What is Big Data?

Our starting point for this exploration of “Big Data” and its impact on PR must be with a definition of the term. The most fruitful approach, perhaps, is to break down the concept and tackle each of the term's component parts one at a time. However, rather than deal with the individual terms in order, I will initially address the issue of “data,” as this will help us comprehend why the scale of Big Data is so significant.

Firstly, then, what exactly do we mean by “data”? Technically, big “data” can include website traffic, search engine marketing performance data, online polls, online product reviews, tweets, photos and videos uploaded or shared via social networks, CRM databases, purchase transaction databases, GPS signals from our mobile phones – ultimately anything that contains information which may be useful when combined with other data sources. For our purposes, it is perhaps better to refine our understanding of data to specify the freely and publicly available textual data generated through social media that more and more of us are sharing each day, via an ever-increasing number of social channels and platforms. This, arguably, offers the most immediate potential and practical uses for PR.

Then there's the issue of size. Big Data is, well, “big”. But how big? There are a number of ways to deal with this. Firstly we can attempt to address the sheer scale of Big Data through numbers. According to technology and consulting company, IBM, each day the socially networked public generates 2.5 trillion bytes of data. That's 2.5 exabytes. Not megabytes, or gigabytes or terabytes – not even petabytes, but exabytes!283 But, while this gives us an idea of raw size it is not necessarily a helpful way to address the issue.

To provide a more comprehensible understanding (and to root the notion of Big Data firmly in the realm of social media), the Canadian social media analytics company, Sysomos,284 claims that its software collects more than 16 million posts every hour from social media sources, such as blogs, forums, Twitter, Facebook, YouTube, Flickr, LinkedIn and others.

This interpretation, while helping us get to grips with “Big Data” at a literal level, needs further refinement if we want to recognize fully Big Data's strategic and practical potential for PR. As Danah Boyd and Kate Crawford have observed, the term “Big Data” is misleading as the real value it offers is not primarily in its size or type of data, but “from the patterns that can be derived by making connections between pieces of data, about an individual, about individuals in relation to others, about groups of people, or simply about the structure of information itself”.285 To put it more succinctly, while we may have access to increasingly vast amounts of information through the rise of social media, it is the ability to gather, analyze and interpret this data that brings about ground-breaking opportunities for PR practitioners.

Big Data and text mining applied to PR

With the inexorable rise of data available for analysis, PR practitioners must understand how it can be analyzed – or “mined” – to find the vital insights that will yield competitive advantage. Data mining, or more specifically text mining, is a process that allows the analysis of a diverse range of public and proprietary information, such as tweets, product reviews and even individual purchase histories.

At text mining's core is usually high-powered analytical technology that enables users to make connections and spot patterns in the vast amount of data that is simply not possible through manual analysis due to the sheer scale of resources and human effort this would require. This allows analysts to not only identify emerging issues or trends relevant to organizations but also begin to predict potential future events. Text mining of Big Data can help us identify the elusive “unknown unknowns” that former US Defense Secretary, Donald Rumsfeld, so eloquently described and in theory develop an appropriate response. An example of how text mining of Big Data has helped deliver PR outcomes for a national UK organization might help illustrate its potential benefits.


Case study: Financial Ombudsman Service

Early in 2012 the Financial Ombudsman Service, the UK's independent organization set up to resolve complaints between consumers and financial businesses, briefed the global conversation agency, We Are Social, to undertake a strategic social media project. The project had three primary objectives which were to help the ombudsman service:

1. Use social media to engage directly with stakeholders and consumers.
2. Manage its brand reputation by monitoring and responding to mentions of the ombudsman service brand.
3. Identify and manage “unknown conversations” relevant to the ombudsman service's work.

This case study focuses on the third objective as it offers the best example of how text mining analysis can be applied to big sets of social media data for PR purposes.

The business goal for identifying conversations about financial services and providers unknown to the ombudsman service was to enable the organization to add value through social media engagement. By being able to spot and resolve potential consumer problems before they became more serious and complex complaints requiring greater resources to resolve, the ombudsman service ultimately hoped to minimize its case workload and potentially reduce costs. Secondarily, this activity would help strategic business and communications planning by spotting and potentially predicting emerging consumer finance issues and trends.

This was a complex challenge that involved human as well as automated text analysis and insight and was achieved as follows. Initially, We Are Social worked closely with the ombudsman service to identify a set of “base terms” corresponding to the ombudsman service's main areas of focus, such as bank, insurance and mortgage, and other key words and phrases used by consumers when discussing finance products or providers online. These included “credit rating”, overdraft, “over limit”, etc. Using these terms, searches of social media data were undertaken using the analytics platform Sysomos Map.286 This generated a high volume of conversations with low levels of accuracy, as the search terms identified general finance discussions that were irrelevant to the ombudsman service.

To increase the accuracy of results, additional text mining techniques were applied to the search, including:

  • Proximity constraints: ensuring that Sysomos Map only looked for conversations with a maximum of ten words between the base term and other key words.
  • Personalization: use of personal pronouns in genuine complaints, such as “My pension was missold” or “I was overcharged” identified through manual analysis of natural language patterns.
  • Everyday expressions: use of emotive language and colloquial terms likely to be used in potential consumer complaints identified through further manual analysis of relevant conversations, e.g. “nightmare!”, “get a resolution”, “what can I do?”, etc.

This helped identify a much smaller volume of conversations (e.g. 1,500 discussions over a six-month period, rather than several thousand) with much higher levels of accuracy and relevance. As a consequence, We Are Social and the ombudsman service were able to distinguish specific problems and potential complaints from general financial discussions. The analysis also uncovered three main types of conversations:

1. General complaints
2. Specific questions or advice
3. Highly emotive rants.

This enabled We Are Social and the ombudsman service to work out a strategic engagement plan with a tailored approach to the different types of conversations identified. The analysis also helped predict likely changes in conversation volumes over time, which meant the ombudsman service could help adequately plan its resourcing of future conversation engagement.


Limitations of Big Data

Despite the increased analytical power of software and technology being adopted in text mining Big Data, this technologically determinist approach cannot adequately deal with the complexities of interpersonal communication. For example, one of the biggest challenges software technology faces is how to interpret the highly nuanced and individualistic use of language. Automated analysis can rarely identify and process human traits such as sarcasm or irony with any accuracy; similar problems are encountered in sentiment analysis. While many commercially available technology products may claim to provide 80% reliability in their automated analyses, the experience of many PR professionals engaging in social media research suggests this is an over-statement.

A further potential restriction is the language used by an organization's stakeholders or publics within a globally networked environment. Although English is the internet's most widely used language, Chinese and other globally dominant languages, such as Spanish and Portuguese, are nearly as popular. This can cause problems for analytical technologies which can only function with English texts. This limitation is further exacerbated by the significant presence of non-Latin languages, such as Chinese and Arabic, online. One obvious solution for these challenges is to replace automated analysis with human oversight. Given the volume and complexity of Big Data, however, any manual analysis will have to balance using a limited data set capable of human interpretation or potentially spending vast sums of money hiring, training and maintaining substantial teams of highly-qualified analysts – a luxury not many organizations or PR agencies can yet afford.

While the incorporation of human insight can offer a potential solution to the linguistic challenges of Big Data analysis facing PR professionals, this presumes the skill set of PR practitioners is evolving to meet the need of a Big Data-driven communications environment. Unfortunately, evidence from studies into the changing skill sets of PR professionals belies this assumption. Research carried out during 2012287 indicates that while UK practitioners are adapting their skill sets to the rise of social media technology, this change is not necessarily as rapid as it could be and may lead to the industry falling behind other, more data-driven industries such as advertising, management and specialist social media consultancies.

Such a shift in terms of PR skills, however, is not enough to fully deliver a renewed industry capable of engaging with Big Data-driven public relations. Without buy-in from senior practitioners and support from industry bodies, such as the CIPR and PRCA, the contemporary and possibly future reality of PR and Big Data may represent a further limitation to the strategic and creative contribution the sector might make to wider marketing and communication strategies. The consultant and author Martin Thomas ascribes PR's reluctance to embrace the strategic potential offered by opportunities such as Big Data to historically-driven traits like intellectual laziness, lack of confidence and executional fixation. The results have arguably been that PR has prioritized media relations over the development of strategic and creative insights designed to ensure the long-term viability of the PR function within organizations.288

Conclusions and future directions

But it needn't be like this. While some of the potential applications of Big Data to PR might seem daunting (or fanciful) to some, forward-thinking organizations are already adapting and utilizing Big Data in some way. To help illustrate this, perhaps it might be helpful to go back and restate our understanding of Big Data and sketch out how it might be applied tactically and strategically by PR practitioners.

Firstly, Big Data is arguably not a distinct, uniform concept but rather a way of understanding the proliferation of publicly available data and the ways in which analysis can be used to develop strategic and tactical insights for organizations. Some of these approaches include using Big Data insights to:

  • Identify new types of value-driven audiences or communities based on self-expression and the type of things they're saying, doing or buying in real time.
  • Identify and predict longer-term changes in audience or stakeholder attitudes or behaviours.
  • Identify, analyze and even predict internal and external events likely to impact on an organization, such as operational crises or consumer dissatisfaction.
  • Devise more creative and strategic responses to communications objectives.
  • Craft more targeted messages and create more relevant content in the format most appropriate to the audience.
  • Evaluate the impact of communications campaigns more effectively and across a broader range of organizational metrics.

In all instances, however, these opportunities will bring about new practical and theoretical approaches for PR practitioners. In order to adapt successfully it is crucial for practitioners to recognize Big Data's limitations and approach analysis and organizational application with a set of realistic and workable objectives. Done right the results offer the PR profession significant short- and longer-term opportunities.

For example, Martin Thomas suggests data-driven PR practice can help the discipline gain greater recognition and respect from clients and thus improve influence among senior management. Such status, in turn, can help strengthen PR's position both within client organizations and also with regard to other marketing and management industry competitors, such as advertising, SEO and management consultancies (all of which claim to be operating to the same organizational objectives). The ultimate result is that data-driven communications can help generate increased reputation, revenue and greater profit margins. Such tangible benefits, it should be clear, point to a successful, increasingly strategic and thus long-term future for PR.

Given the potential outcomes of Big Data it's no surprise that the World Economic Forum this year categorized data as a commercial asset in the contemporary knowledge economy. To ensure PR secures a future for itself as a strategic business function in this networked era, it is going to be vital for practitioners to tap into these assets and make the most of them in establishing, maintaining and even predicting likely changes in the mutual understanding between organizations and their publics.

Biography

Simon Collister is a senior lecturer at University of the Arts, London. He is currently conducting PhD research at Royal Holloway, University of London's New Political Communication Unit on the mediation of power in networked communication environments. Before entering academia, Simon worked for a number of global communications consultancies, planning and implementing research-led campaigns for a range of public, voluntary, and private sector organizations.

Notes

280Bringing Big Data to the Enterprises, IBM: http://cipr.co/14JLnlP

281It's important to note there are literally hundreds of different definitions of PR, although many of these consider the management of relationships between organizations and their specific publics as a central function

282The Context of Public Relations, Exploring Public Relations: http://cipr.co/11x7tKD

283Bringing Big Data to the Enterprises; IBM: http://cipr.co/14JLnlP

284Sysomos Products; Sysomos: http://cipr.co/Y8mLgz

2856 Provocations for Big Data; Danah Boyd and Kate Crawford: http://cipr.co/Xzx3da

286Sysomos MAP; Sysomos: http://cipr.co/Z9WFk2

287Recruiting for PR 2.0; Sarah Williams, Jennifer Challenor and Simon Collister: http://cipr.co/12Le7MC

288PR & Planning; Martin Thomas: http://cipr.co/XzKRRm

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
18.189.171.125