Listening to your audiences has become a truism in both public and private sectors. And the concept is hardly new. Long ago and far away, b...
Listening to your audiences has become a truism in both public and private sectors. And the concept is hardly new. Long ago and far away, businesses and organisations would send researchers out with a clipboard and survey to knock on doors and assail people in the streets to ask them their opinions.
However, for many decades, the techniques involved remained largely unchanged. Pollsters may have taken to calling people on the phone, and with the rise of the internet, the use of email and Web surveys ostensibly brought down the potential cost and reach of such exercises. But at root, this was still largely a case of surveying random samples of an audience in order to better understand what it was thinking and feeling – and to use this to anticipate how best to meet its needs – or, in a PR context, determine what messaging strategy and tactical activity to develop.
The use of market research and opinion poll data has been grist to the PR mill for some time – having said that, there have always been a number of caveats in its use. The old shoemakers’ maxim comes to mind: fast, quality and cheap – but not all three at once. Clearly, the bigger the survey, the higher the cost involved – and typically the longer it would take to gather the data. Quick dipstick surveys may provide fast results, but the quality and depth would obviously be lacking. How much store would any organisation lay by such insight?
Not only this, but there was always the nagging doubt that those being surveyed were giving the pollster what they wanted to hear rather than what they honestly thought. Or rather, there was a potential disconnect between what people said and what they did.
What has changed in the last five years is the inexorable rise of social media. From the standpoint of the public, they now have an easy means of expressing their views and opinions about pretty much anything. Brands and institutions make for natural conversation fodder.
Conversely, organisations now potentially have immediate real-time insight into the thoughts, views and feelings of their stakeholder audiences in a way that would have been unthinkable in earlier times – not just customers, but journalists, politicians and other important influencers. And, in many cases, this comes at a trivial cost.
In light of which, the case for social media monitoring would appear self evident. What organisation would not want to listen to its audiences to determine what they think and feel about them? Happier customers tend to be more predisposed to buying a company’s products, or voting for a political party. On the other hand, negative sentiment is a big red warning flag that says something is amiss and needs fixing.
However, given the almost universal agreement that social media monitoring is a good thing, what are the key challenges facing PR practitioners in harnessing the power of social media monitoring to better inform communication strategies and campaigns?
Perversely, one of the main issues is the sheer volume of data freely available. Whereas in the past, high quality real-time data was only available at great cost (if at all), pretty much anyone can create a low-cost social media monitoring dashboard with a tool like TweetDeck – for nothing. But anyone who has spent any time trying to monitor all but the most esoteric Twitter conversations will soon become numb to the torrent of Tweets cascading down a TweetDeck column – let alone being able to interpret what is being said and using this as the basis for action.
Even assuming that an organisation is able keep on top of what is being said, it raises a number of vital questions about resourcing and responsibilities. If a customer makes a complaint on Twitter, who should respond to it? Should it be responded to in a public environment? Are some people’s opinions more important than others? How much resource should an organisation allocate to monitoring social media and who should be responsible for it?
To give an example, BT (one of the world’s leading communications services companies operating in over 170 countries) has a team of more than 30 full-time employees dedicated to monitoring what is being said about BT online (including on social media). And they are empowered to react and respond. Interestingly, this team falls under the remit of customer service rather than PR. Clearly, an organisation the size of BT has its own business objectives and challenges – but it is a useful example of the kind of issues and challenges that any organisation will face when attempting to determine the level of investment in social media monitoring.
However, for many decades, the techniques involved remained largely unchanged. Pollsters may have taken to calling people on the phone, and with the rise of the internet, the use of email and Web surveys ostensibly brought down the potential cost and reach of such exercises. But at root, this was still largely a case of surveying random samples of an audience in order to better understand what it was thinking and feeling – and to use this to anticipate how best to meet its needs – or, in a PR context, determine what messaging strategy and tactical activity to develop.
The use of market research and opinion poll data has been grist to the PR mill for some time – having said that, there have always been a number of caveats in its use. The old shoemakers’ maxim comes to mind: fast, quality and cheap – but not all three at once. Clearly, the bigger the survey, the higher the cost involved – and typically the longer it would take to gather the data. Quick dipstick surveys may provide fast results, but the quality and depth would obviously be lacking. How much store would any organisation lay by such insight?
Not only this, but there was always the nagging doubt that those being surveyed were giving the pollster what they wanted to hear rather than what they honestly thought. Or rather, there was a potential disconnect between what people said and what they did.
What has changed in the last five years is the inexorable rise of social media. From the standpoint of the public, they now have an easy means of expressing their views and opinions about pretty much anything. Brands and institutions make for natural conversation fodder.
Conversely, organisations now potentially have immediate real-time insight into the thoughts, views and feelings of their stakeholder audiences in a way that would have been unthinkable in earlier times – not just customers, but journalists, politicians and other important influencers. And, in many cases, this comes at a trivial cost.
In light of which, the case for social media monitoring would appear self evident. What organisation would not want to listen to its audiences to determine what they think and feel about them? Happier customers tend to be more predisposed to buying a company’s products, or voting for a political party. On the other hand, negative sentiment is a big red warning flag that says something is amiss and needs fixing.
However, given the almost universal agreement that social media monitoring is a good thing, what are the key challenges facing PR practitioners in harnessing the power of social media monitoring to better inform communication strategies and campaigns?
Perversely, one of the main issues is the sheer volume of data freely available. Whereas in the past, high quality real-time data was only available at great cost (if at all), pretty much anyone can create a low-cost social media monitoring dashboard with a tool like TweetDeck – for nothing. But anyone who has spent any time trying to monitor all but the most esoteric Twitter conversations will soon become numb to the torrent of Tweets cascading down a TweetDeck column – let alone being able to interpret what is being said and using this as the basis for action.
Even assuming that an organisation is able keep on top of what is being said, it raises a number of vital questions about resourcing and responsibilities. If a customer makes a complaint on Twitter, who should respond to it? Should it be responded to in a public environment? Are some people’s opinions more important than others? How much resource should an organisation allocate to monitoring social media and who should be responsible for it?
To give an example, BT (one of the world’s leading communications services companies operating in over 170 countries) has a team of more than 30 full-time employees dedicated to monitoring what is being said about BT online (including on social media). And they are empowered to react and respond. Interestingly, this team falls under the remit of customer service rather than PR. Clearly, an organisation the size of BT has its own business objectives and challenges – but it is a useful example of the kind of issues and challenges that any organisation will face when attempting to determine the level of investment in social media monitoring.
Sentiment Analysis
The rise in interest in social media monitoring has seen a commensurate rise in technology solutions available to meet this demand – solutions ranging from the free to the very costly. Perhaps the most interesting area is in the use of so-called automated sentiment analysis.
What is sentiment analysis? Quite simply, it is the attempt to deduce how somebody feels about a particular person, topic, issue or organisation based on what they say. Given the sheer volume of social media content available, technologies have emerged that attempt to automate this process of analysis. They fall into two broad camps. The first makes use of simple word matching based on dictionaries (or bags of words as some cynics describe them). Lists of ‘positive’ and ‘negative’ words are drawn up in advance, and any piece of content is then analysed and matched against these lists. If the balance of negative to positive words is higher, then the content is deemed ‘negative’ towards the subject.
The benefits of this kind of approach are that automated sentiment analysis tools can be created at low cost for the end user. Indeed, some tools such as Social Mention are free to use. The downside is clearly accuracy. Critics of the ‘bag of words’ approach rightly argue that this technique can’t deal with irony, sarcasm or slang – and thus the prevalence of so-called ‘false positives’ can be high, thereby adding to the resource overhead required to check the accuracy of the analysis. Having said that, there is no question that free tools such as Social Mention allow any organisation to at least test the concept of sentiment analysis in the context of social media for no monetary outlay, and can give them some idea as to what level of investment may be appropriate for them in the future. Low-cost tools such as UberVu will also give the end user the ability to easily amend the sentiment rating if inaccurate assessments are spotted.
Contextual Analysis
The opposing camp uses sophisticated computer algorithms and semantic analysis to provide a higher level of accuracy and refinement. These more contextually based tools allow for more accurate analysis of far higher volumes of content as well as being able to isolate sentiment from different viewpoints (big bankers’ bonuses may be perceived negatively by the public and media. However, a banker who is going to receive a big bonus may see this differently). The only downside with this approach is clearly the comparatively high cost of such solutions, which will make them viable only for the very biggest of organisations (although that may change over time as usage rises and providers are able to offer these capabilities to a broader market for lower cost).
In tandem with the rise of automated sentiment analysis has been the introduction of influence rating services such as Klout, PeerIndex and PeopleBrowsr. These tools have the potential to answer one of the other key challenges of social media monitoring – namely, if some people’s views and opinions are more important than others (or perhaps need responding to more urgently than others), how do you identify who those people are? A tool such as PeerIndex, for example, will attempt to analyse an individual’s online influence and then provide an overall score relative to a particular subject or topic. Using this kind of analysis could help PRs to determine whether someone who is voicing a negative opinion needs to be dealt with in a particular time-frame or in a certain way based upon their potential impact on a key stakeholder audience (although caveats clearly apply given the level of maturity of the technology).
Network Topology
The other emerging aspect of social media monitoring is the concept of relationship network analysis. PR (and marketing generally) has traditionally been a very linear process. On a simplistic level, an audience is identified, a list of relevant journalists and media titles is built and content is fired at them with the hope that key messages will stick and influence the target audience. This approach ignores the interconnectedness of an internet-mediated world. To be fair, the means to actually track, measure and understand the nature of networked relationships has never really existed until recently. And network science has traditionally focused on technology rather than people. The pioneering work of people such as sociologist Nicholas Christakis demonstrates the power and value of understanding network topology – whether of a computer network or a cadre of political bloggers. The marrying of social media monitoring with sentiment analysis, influence ranking and network analysis gives a glimpse of the opportunity for the PR profession.
In summary, it is hard to argue against using some form of social media monitoring – especially in the context of helping to guide and inform the development and execution of PR and communications programmes. Given the ability to test out approaches to social media monitoring at next to no cost, there seems no excuse for not trying it. Conversations about your organisation or brand are taking place whether you like it or not. Paying no attention to this conversation at all could be seen as negligent business practice in the modern world. The bigger challenges lie, though, in terms of how companies determine their level of investment and commitment to social media – and who should take responsibility for owning and managing those approaches. PR professionals should and must be involved in that debate. However, in order to win the battle, they must arm themselves with direct knowledge and understanding of the value of social media. Social media monitoring is a vital first step.
COMMENTS