Phil Lynch, Managing Director at Newton Insight shares in this interview his experience on how media intelligence makes crisis management effective and outlines the new challenges that suppliers face with the speed and scale of modern media when it comes to crisis communications and general communications. Phil also outlines how artificial intelligence and machine learning are changing the landscape for media intelligence and what are the risks.
Q: How media intelligence makes crisis management effective?
Phil: We are currently in the middle of AMEC Measurement Month and Newton has been focusing on giving clients lots of practical guidance on how to get the best from their media analysis resources. With this in mind, the first point to make is that many day-to-day media monitoring and measurement tools can be deployed effectively to support crisis management. Secondly, it is important for clients to understanding how measurement can be used strategically and tactically.
Let’s start with the big picture; if a client is tracking its media profile using measurement tools, it has a model of how much coverage it generates under normal conditions. The client also knows its most active media outlets. Over time the data becomes predictive. During periods of crisis, the client can measure deviations from the norm, to calculate the momentum and trajectory of the media response. This can then be broken down to analyse which sections of the media are showing the greatest variation (ie who to talk to first and where to focus your efforts to achieve change).
If we take the Volkswagen emissions issue as an example, we know from historic data that a crisis event in the automotive sector has a typical half-life of 48 hours in the media. The coverage trend follows a straight-up / straight-down pattern. However, in the first week of the VW issue, media activity did not follow the expected model. The initial peak was higher and the drop-off in media interest didn’t happen. This would have given VW the clearest indication that the issue needed the fullest possible response.
Q: For crisis communications and general communications, how do suppliers cope with the speed and scale of modern media? What new challenges do we face?
Phil: We are seeing more use of technology to harvest content and to produce basic quant metrics. The challenge for the measurement sector is to process more content, faster, without losing sight of our primary goal – to give clients the insights they need to perform better. Yes, we can cover more channels, but these channels have to be relevant to our clients’ needs. I sometimes wonder if the sector is doing itself a disservice by placing so much emphasis on the number of media outlets available to clients. It’s what we do with all of this stuff that really matters.
A big challenge moving forward is access to relevant data from channels which are important to clients. For mainstream media, we have to navigate our way through paywalls and licensing agreements, which can be a costly exercise, especially for smaller agencies. Access to content in social networks is also very uneven, partly due to privacy controls but increasingly because of the proprietary interests of the networks themselves, who are seeking to establish direct data relationships with end-users.
The nature of content is changing too. Traditionally our sector has focused on textual analysis, but images are now a major part of communication and agencies have to find ways to incorporate visuals into their workflow at scale. I know of several initiatives to automate image capture and analysis at supplier level, and Google is ramping up its own image classification capability. This may give the wider industry access to image segmentation at a basic level.
Q: How is Artificial Intelligence changing the landscape for media intelligence? Are there any risks?
Phil: If there is a risk, it is that we ask too much of AI too soon. It’s important that we do not over-promise on what AI can deliver.
I see AI as a source of support rather than as a replacement technology. AI is good for specific tasks with defined parameters. Agencies and clients alike are using AI to automate daily tasks such as collating coverage reports and creating media lists, freeing up their time to focus on added-value work.
The more sophisticated AI becomes, the more tasks we can ask it to undertake. At a more advanced level, AI and machine learning can calculate what messaging and content works best, based on previous activity. Machine learning techniques can help us tailor the language we use for a particular audience or topic based on what we know has worked in the past. The result is higher relevance messaging, tailored to a specific audience.
I think AI will also help to solve the fake news problem by looking at what is being shared, to who and how fast, using data from previous fake news patterns as a reference.
AI can help us to discover things by scanning unstructured data for specific signals that match patterns we are interested in. AI could identify issues and opportunities on the horizon based on what we know of the past. This goes back to what I was saying about the strategic value of measurement. It’s all about helping our clients to perform better.
About Phil Lynch
Phil Lynch is Managing Director of Newton Insight and Visiting Lecturer in media analysis at the University of Westminster.
Phil has 20years’ experience of delivering research and analysis solutions to market-leading companies and forward-thinking public organisations.
About Newton Insight
Newton Insight is a full-service reputation research agency, using the scale and speed of social media insights to understand how people feel about organisations and brands. Newton Insight delivers analytics on social media trends and deeper insights into audience attitudes, opinions and emotions. For more information visit the company website and follow on LinkedIn.