Douglas Chapin, CEO at CyberAlert LLC, shares in this post his views on why brands need predictive analytics and what is contributing to this evolution, describes in which way does the rise of being data-driven affect media monitoring and visualization tools, explains how the increasing blend of data and software is affecting the relationship between media intelligence and business intelligence and comments on what he sees as the biggest obstacle that keeps sentiment analysis from reaching its true potential.
Q: Why brands need predictive analytics and what is contributing to this evolution?
DC: Every element of marketing should be focused on delivering improvements to brand awareness, revenue growth, profitability and value of an organization. The improvements are generated through allocation of resources and then execution of plans which should positively impact drivers of brand revenue growth and profitability and/or increase value in the future. This means that every choice is based on either good or bad prediction of future outcomes of a decision.
When I started my career at Hewlett Packard we did product planning using the “next bench model.” This model was incorrectly vilified by people who didn’t really understand it. The truth was that it worked great as long as the engineer sitting next to you was a good predictor of the customer’s problems. The predictive analysis only broke down when HP customers stopped being electrical engineers.
Since that time marketers have become more sophisticated in their approaches to understanding how to model and predict the impact of products, campaigns and communications on revenue growth and profitability. However the collection of the data necessary to develop good models has been very expensive because it was primarily based on aggregating data from disparate sources or assembling customers and prospects to test marketing, promotional and product designs.
The internet reduced the costs of communications significantly and additional change has occurred with the explosion of social media. Social media has provided a continuous stream of data from millions of customers commenting, promoting or denigrating brands, products and campaigns in real time. This treasure trove of real time data combined with the advances in communications theory and analysis tools enables smart organizations to access real time predictive models of consumer behavior and provide advance warnings for brand mangers before negative events have exploded onto the top story listing on Google News.
Q: In which way does the rise of being data-driven affect media monitoring and visualization tools?
DC: There are many challenges for media monitoring companies in a data-driven world.
One is that the same old standard sets of charts are no longer adequate. Smart customers who understand data collection and modeling want to be able to see their specific visualization of the collected data. They demand that their provider be able to both customize the charts to their specific model and also provide them with the ability to access data to support their own analysis tools. At CyberAlert we are currently supporting several clients that have in-house tool sets such as Tableau. They do their own visualization and integrate our measurement data with in-house systems. We also created the capability in our new software to create completely customized client-specific dashboards which were initially modeled using Tableau.
Another key challenge is the ability to integrate much more than just media data into client dashboards. In my experience, sophisticated customers who understand the value of measurement of their communications program want their monitoring and measurement provider to be capable of delivering real time services which integrate web analytics, multiple social media channels and even financial systems into their measurement systems. All of CyberAlert’s current dashboard clients require the integration of data collected through multiple systems into their dashboards in order to measure the effectiveness of their communications and social media programs.
As we support our clients desire to measure outcomes which can be directly linked to business objectives, we expect the need to integrate with other systems will continue to grow.
Q: As media monitoring becomes savvier with data and software, how do you see that affecting the relationship between media intelligence and business intelligence?
DC: As mentioned in the prior answer, at CyberAlert we already integrate many business intelligence data sources into the analysis dashboards that we have created for customers. As analytics progress, media intelligence will become a key component of overall business intelligence. Media teams need to expand their view of monitoring and their knowledge of data analysis or else will be left behind and unable to contribute to the organizations success. In-depth understanding of statistics and statistical analysis has become imperative for PR, marketing and communications staff.
The interesting questions today are no longer “How many mentions did we get last month?” because that may not have any impact on business results. Media professionals now need to be asking questions like “How can I predict the Nielsen rating of a TV show based on our social media efforts?” or “How can we best measure the impact of our actions on qualified lead generation?”
CyberAlert believes that we are here to make our clients more successful by building media monitoring and measurement capabilities well beyond what the average customer currently uses.
Q: In your opinion what is the biggest obstacle keeping sentiment analysis from reaching its true potential?
DC: The biggest obstacle for current sentiment analysis tools is the ability to align the analytics with a client’s specific objectives. The value of general sentiment analysis is limited for many customers because it is not statistically valid to the specific outcomes they desire.
As soon as a client wants to be able to measure specific outcomes based on their media analysis, they find that generalized positive and negative sentiment measurement frequently has low correlations to the outcome they are trying to measure. The measurement of outcomes requires either analytics driven by specific taxonomies which are still very expensive to develop or the customization of sentiment definitions to match individual client needs.
This is the reason that CyberAlert still uses human analysts to provide the media analysis for clients that desire finely-tuned measurement dashboards. A trained human analyst is still much more cost effective than a sentiment analyzer for measuring the effectiveness of communications program because they can easily determine key message uptake and other key metrics in addition to providing results of tonality specifically defined for individual clients.
About Douglas Chapin
Douglas Chapin is the current CEO of CyberAlert LLC. He has over 25 years of experience in successful startup and global company leadership, including two decades as an executive at Hewlett-Packard.
Doug led the turn-around and sale of PR and social media measurement powerhouse, KDPaine and Partners LLC, to News Group International. He served as Chief Operating and Technology Officer at NGI during the integration period.
A passionate entrepreneur, Doug also shepherded the medical device company, AlterG, from initial startup through creation of its first production model and venture funding. His first startup, GlobalSight – the linguistic software company – won the Red Herring Top 100 and Upside 100 to Watch awards and was sold to Transware Plc.
During his tenure at HP, Doug held several roles in marketing and field management and he launched their IT outsourcing business, growing it to $1.2 billion in five years.
CyberAlert is today’s most cost-efficient media monitoring solution to find and “clip” what’s being said about your organization, its products, people or competitors in the news and social media or by consumers/customers on the Internet.