March was as eventful as ever for the President with more high-profile hirings and firings, continued fallout from Russian political interference, and increased attention on extra-marital affairs allegations. The impact of these events was clearly evident in how the public viewed the President throughout the month.
While Trump's approval rating in March tended to stay within a range of 45 to 50 percent, the public were viewing Trump more negatively through the period of March 14 - 19 with his approval rating reaching a monthly low of 39%. The public's discussions during this period focused on the firing of former Secretary of State Rex Tillerson, the hiring of Larry Kudlow as Trump's top economic aide, and Russian sanctions. The last time the President's approval rating reached such a low level was in the aftermath of the February 14th school shootings in Florida.
Trump's approval rating bounced back to its normal range over the last 10 days of the month, despite the period including heightened attention to multiple allegations of extramarital affairs and more changes in various roles within government.
Through the analysis of online conversations, Eyesover provides real-time polls indicating the public's support (or lack thereof) for the topic of the conversation. Many of our customers use this polling ability in the political sector. Eyesover listens, analyzes, and tracks how much public support each political party or candidate is getting compared with their competition.
So how do we determine public support? To answer this question, we looked at what it means to be considered a supporter.
Consider how one might learn about a friend's favorite sports team? One way is to ask them - this is how traditional polls work. The poll would ask a question and provide options to choose from. A different way to find out what your friend's favorite sports team is, would be to just listen to them. Typically, if someone is vocal about their opinions then it generally doesn't take long to learn where that person stands on different topics. Even if someone isn't very vocal, we often learn a lot from normal, everyday conversations with each other. The key here is listening. Listen to what people are saying and learn from it. Even without asking a question, it is not difficult to tell if someone is opinionated about specific topics based on how much they keep bringing up the topic.
We took these concepts and applied them to how we find supporters within the social media landscape. Our customers simply choose which subjects, such as two competing politicians, they would like to find public support for and our system then listens to all online conversations taking place about those politicians. By listening to these conversations, Eyesover is able to determine if people on social media are (a) interested in a subject, based on how much they discuss it, and (b) how they feel about that subject.
Through our product development, we have learned how much discussion and what sentiment is required to confidently state an individual is a supporter or an idea, brand or individual. The analysis is performed for every “vote” in our poll in real-time. In fact, Eyesover's polling system calculates over 10 million voters every hour, and as the search for discussion grows through our AI technology, more voters are calculated every day.
Determining a supporter has far reaching benefits. Not only does this provide accurate predictions for such things as political elections, it also provides our customers with a deep understanding about the critical opinions influencing their audience.
As the numbers from last night’s Special Senate Election in Alabama rolled in, things were playing out exactly how Eyesover Technologies had predicted they would as Democrat Doug Jones beat Republican Roy Moore.
Heading into the vote counting Tuesday night, Eyesover projected Jones had a 52-48 advantage over Moore. The final results were Jones at 49.9%, Moore at 48.4% and 1.7% write-ins.
Eyesover uses online data to provide public opinion to their clients including their Real-Time Polling that correctly predicted the Jones victory Tuesday night. “We understand that the communications landscape is undergoing a massive shift with people cutting the cord and increasingly using online devices for their everyday communication needs. Because of this, the way we identify, measure, and act on public opinion has to change as well. As the public increasingly expresses their views online, that data can be used to understand public support for products, policies, and in last night’s case, political candidates” stated Eyesover CEO Craig Leonard.
The New Brunswick, Canada based company has recently made inroads into the US market with their software and expect their success in predicting the Alabama result will accelerate that growth. “You could not ask for a more difficult election to predict as demonstrated by the fact most polls were predicting a Moore victory. But when you have access to over 35,000 individuals expressing their opinions as we did on Tuesday, it makes it possible to accurately predict what is going to happen” Leonard said.
The Eyesover platform uses artificial intelligence to find online discussions pertaining to their customer’s area of interest. The data collected by the platform allows Eyesover to tell their customers what issues are trending, public opinion, and helps target advertising towards the most receptive audiences for the customer’s messages.
Doug Jones is heading into tomorrow's Senate Election in Alabama with a small but growing lead over Roy Moore. After being tied at the end of last week, Jones has started to add support over the past couple of days to move ahead 54-46.
Since Eyesover uses online data for its polling, determining the location of a specific account is not always possible if the account owner chooses not to publish their location. However, Eyesover can segment the individuals discussing the Alabama election online into groups of people who are from Alabama and those who have no confirmed location (they may be from Alabama or any other location).
The variance in support levels between these two groups is interesting as the Alabama group shows the 54-46 lead for Jones while the data from the Unconfirmed Location (UL) group shows a closer race at 51-49 for Jones.
When we dive down into the issues these groups are discussing in relation to the election, the Alabama group's top issue is Healthcare while the UL group's top issue is Human Rights/Women's Issues.
The strong lead Doug Jones held in the Alabama Senate race last week is all but gone as Roy Moore's support has rebounded driving Jones' lead down to 4 points.
As a result of Jones' lead last week, it appears that Moore supporters have become more engaged as there has been an increase in the number of individuals talking negatively about Jones online, while discussions about Moore have become more positive.
While 90% of last week's discussions pertaining to the race were about Moore, we now see a 60-40 split between Moore and Jones. The issue for Jones is his sentiment scores have fallen from an overall positive to an overall negative in a relatively short period of time.
Today's Eyesover Support Index indicates that Alabama Republican Senate candidate Roy Moore has been unable to stop the erosion of his support over the past week and now trails Democratic candidate Doug Jones by 12 points.
Over the past week there has been 10x more online discussion pertaining to Moore than Jones, but the overwhelming majority of those discussions have been marked with very negative sentiment towards Moore, the cause of his deteriorating support.
Jones has been in the background while Moore has been the focus of the media over the past week, but he will clearly be attracting more attention from all sides now that the race is competitive. With his online mentions increasing over the past two days, we'll be watching to see how it will affect his sentiment and support scores.
As we often post the results from our political Real-Time Polling on our site, we thought a couple of posts discussing the current challenges traditional pollsters face and how Eyesover can help solve some of those problems would be in order.
Traditional pollsters face challenges on multiple levels these days. It has always been no easy task to deliver accurate, up-to-date information on the opinions of the public, but in today’s changing society, it has become even more difficult.
Although there have been technological advancements in the industry with automated phone calling, which greatly reduced the cost of calling voters across the country, the reality is that 2016 marked the first year where more than 50% of households in the US did not own a landline. Of course, cell phones are increasingly being targeted by the industry, but reaching the owners of these devices poses a significant challenge as well due to the wide range of rapidly improving call-blocking apps available.
Even worse for polling companies, the households that still have landlines are increasingly skewing towards those in older age groups. More than seven in ten households under the age of 35 are now cellphone-only. These issues create significant and costly problems in creating a sample that accurately reflects the population, a necessity for accurate polling.
Considering these challenges, credit must be given to polling companies that, in-spite-of these adverse conditions, still manage to produce accurate and insightful polls!
But as the landscape of telecommunications changes, we believe polling derived from other sources of information will play a growing role in the industry and we’ll explain why in our next post.
Mention count, number of views, expected reach; these are just a few of the terms used by the social media marketers, but what does this information really tell us?
A major problem with online metrics is the potential for mention count to be manipulated by multiple mentions coming from both legitimate accounts and bots. Hundreds of comments, likes, or retweets from the same account will obviously inflate metrics, sometimes to the point where wrong decisions are made based on the flawed data.
Similarly, this inaccuracy can apply to the ‘number of views’ or expected reach. Often, neither metric will give you accurate information regarding the number of unique individuals who viewed your article or video if the numbers are inflated by individuals using different devices or web browsers.
The same holds true for inbound content. Knowing exactly how many individuals are talking about a particular subject can be far more valuable than just knowing how many mentions there are. Too many times organizations will be put on full alert and devote resources to an issue that is rapidly growing in mentions online, yet the reality is the mentions are coming from a handful of accounts.
One of the unique features of the Eyesover system is that we analyze our data on an individual by individual basis. This gives us the ability to report not only on the number of mentions, but also, the far more important metric of the number of individuals that are actually talking about the subject.
This method cuts through the spam and counts high volume accounts as exactly one individual, regardless of how many tweets they posted that day. By analyzing online content in this manner, we not only identify and discover real trends as they develop, we can use the individual opinions for key features such as our real-time polling and ad targeting while ensuring users are not distracted by noise.
Eyesover Technologies Releases New Issue Discovery Software
Fredericton, New Brunswick, Nov. 22, 2016 – Eyesover Technologies Inc. is pleased to announce the release of their new Issue Discovery module as the latest addition to their online media monitoring system.
Utilizing Eyesover’s proprietary artificial intelligence, the new Issue Discovery module provides customers with a predictive tool that identifies online discussions or mentions of new issues that are relevant to the end user.
“The limitation of most social media monitoring tools is selection bias – a system will only report on what you tell it to monitor. Eyesover’s new Issue Discovery module can extrapolate from existing data to see the potential development of issues a customer may not be aware of.” said Eyesover co-founder Dr. Ali Ghorbani, Canada Research Chair in Cybersecurity, Dean of the Faculty of Computer Science at the University of New Brunswick, and the Director of the Canadian Institute for Cybersecurity.
The Eyesover system now provides customers with online media monitoring and analysis, real-time lead generation, and the new issue discovery functionality to enhance the depth of information customers are obtaining from online media.
“We’re looking to take social media monitoring into the next stage where we’re not just listening, we’re predicting what is going to occur. The Issue Discovery module is a big step in that direction that will bring benefits to a wide range of private, public and non-profit entities.” said Eyesover co-founder and CEO Craig Leonard.
About Eyesover Technologies
Established in 2014, Eyesover provides an online media monitoring and issues discovery solution for customers including political parties, utilities, public and private sector corporations and governments.
The system is the product of research from the Intelligent and Adaptive Systems (IAS) Research Group within the Faculty of Computer Science at the University of New Brunswick, led by Dr. Ali Ghorbani.
For more information, visit www.eyesover.com
Contact: Craig Leonard, CEO
46 Dineen Drive