bridge_ci has been designing public affairs strategies based on machine learning and open-source intelligence for the last fifteen years for the largest companies and institutions in the world. The technology has only improved, and we make it a priority to incorporate the most bleeding-edge technologies into our offerings. These investments enable our clients to understand EU policy trends that would be impossible with traditional desk research and intuition alone.
Please get in touch if you want to know more.
Domain Analysis™
bridge_ci has designed a process called “Domain Analysis™,” which leverages world-class machine intelligence, OSINT, and alternative data to extract intelligence on any policy, institution, policymakers, or organization to identify influential narratives and trends. The insights are then used to set the strategy using data, not intuition or experience alone. As a result, the insight provides a comprehensive overview of the landscape of any policy and identifies potential risks and opportunities that might otherwise be missed.
Advantages:
Stress-tests strategic hypotheses, i.e., “this is connected or important to that” in hours, not months, before making large investments in time, assets, and resources.
See all the intelligence on a topic or issue structured in the context of all the information available, not just a few news articles or data points. Personal and biased information feedback loops are filtered out at the beginning resulting in “cleaner” and less noise within the information.
Analytically measure “Soft” trends, such as political events or regulatory risk, not intuition.
Machine Summarization
Many of our techniques leverage advanced natural language processing. With AI tools, we automatically summarize hundreds of documents and extract insights related to the policy of interest, such as the example below that covers the “EU Energy Prices” domain. This approach offers many advantages because, unlike traditional human-curated research, machines can look at all the content or data on a given issue and contextualize it versus looking at data (news articles, social media posts, and research papers) in isolation, which would lead to a total bias based on personalized search results that Google and most other information streams offer. Effectively this is a monitoring system that can provide immediate context & insights into any public issue in real time. Machines can also structure complex events in order of how events happen in near real-time. In some more sophisticated projects, we can use these time stamps and classifiers to build models to predict outcomes based on the type of event - much like a hedge fund would with the stock market, but for political outcomes.
Policy Indices
Much like how a stock trader looks at asset/stock prices, it's also possible to create what we call “policy indices,” which track online mentions related to the policy in question enabling public affairs teams to rank and model emerging trends. The data can come from various sources such as Twitter, news, blogs, and forums like Reddit or Google Search Trends. Below the chart is looking at total mentions across all channels of the policies areas in the top right of the chart. EU Digital Sovereignty and Cloud computing appear to generate the most engagement (mentions) in March and May of 2022.
Taking the same data, we can apply regression to find which domains (topics) are most correlated to one another. Matrix correlation charts (like the one below) are often used in finance and economics to understand the relationships between different variables and to make informed decisions based on the data. They can be especially useful for identifying trends and patterns in large datasets and help researchers and analysts better understand the underlying relationships between dozens of variables (think multiple columns in Excel).
The correlation coefficient measures the strength and direction of the relationship between two variables and can range from -1 to 1.
A value of 1 indicates a perfect positive correlation, meaning that the two variables are strongly related and change in the same direction.
A value of -1 indicates a perfect negative correlation, meaning that the two variables are strongly related and change in opposite directions.
A value of 0 indicates no correlation.
The insights can then be used for planning an advocacy strategy. For example, in the chart above, Chips have a lower correlation to Cloud Computing (.12 correlation) than it does to Digital Sovereignty (.52 correlation). This means that if we were interested in getting more computer chip manufacturing sites made, it would be wise to include how this benefits the Digital Sovereignty strategy versus using cloud computing narratives. Alternatively, the same data can be used to measure how close or far a company’s position is from another policy area of interest and reverse engineer the most efficient ways a campaign can embed within the said policy area.
Why The Data Is Useful:
Understand which policies are competing for mindshare or are connected.
Draft of correlated topics, e.g., if AI is an issue that is trending, cloud computing or quantum policy areas would also be relevant and organic topics to bring up within the macro (AI) conversation.
Save time and money by creating a strategy that combines multiple related policy areas of relevance (high correlation) to maximize resources.
Measure communication effectiveness by seeing if your topic is becoming more correlated with the policy conversation of interest over a period of time.
Machine Learning
Once policy indices are created and then correlated, the data can also be used to build an “influence model” using machine learning to rank the order of how the variables influence the prediction. In the example below, the algorithm ranks the Chips, Cloud Computing, and AI as the most influential second-order narratives connected to Digital Sovereignty across Twitter, News, or Blogs.
Once the categories are established, they can also be utilized to identify the channels of influence. This same machine learning approach that can find influential policies or topics can be extended to communication channels such as websites, Twitter, online advertising, and Reddit. By analyzing these channels, we can determine which channels are most effective in influencing public opinion and driving engagement - Google Searches in this case. For instance, our analysis shows that the most influential channels for Google searches about the European Parliament are Reddit and blogs.
Channel Impact In Driving Google Searches of EU Institutions
This information can be highly valuable in allocating time, ad spend, or other resources to each channel, depending on the specific goals or objectives. It is worth noting that each policy area has a unique dynamic regarding how information is dispersed and consumed. Our data also highlights that Twitter and News have limited influence in this context, suggesting that they operate as largely separate media ecosystems in their own right. By utilizing this type of analysis, we can better understand the various channels of influence and design effective communication strategies that reach the right audience through the most impactful channels. And just like the policy indices mentioned, each channel can be correlated to one another for further insights into channel engagement drivers (chart below).
Channel Correlation
Network Analysis
Another technique we leverage at bridge_ci is network analysis which, in our opinion, is the best way to understand the relationships between topics or entities (person, place, region, organization) within complex systems. The network below clusters (groups) European Commission official communications (the same can be done with any Twitter, News, or text data) based on their topical similarities. More related topics are connected or embedded with one another. Topics that do not have the same policy thematics or key sub-domains of interest are farther apart. For example, the network shows that democracy is not connected to energy or cell research clusters at the bottom.
With this information, you can:
Know where you are starting by measuring how embedded your issue is within a target conversion.
Find out what topics or policies “bridge” other topics of interest.
Post-mortem analysis of how your policy of interest issue becomes more central after advocacy campaigns.
For example, maybe an organization that wants to promote democracy wants to embed the energy and some of the lack of democracy and human rights within many energy-producing states into the conversation. The data could be used to find what topics to communicate, potentially linking the two clusters more efficiently. In this case, it appears to be innovation, resilience, and recovery are the key connecting topics.
Algorithms can also identify the people, places, and regions that are important. This can help to give a more accurate picture of public opinion on certain issues and can help to identify any potential areas of concern. This is next-generation stakeholder mapping that can be done in real-time versus months or weeks using traditional desk research.
The benefits of using data-driven public affairs techniques are clear: more efficient decision-making, increased accuracy in advocacy campaigns, and a better understanding of the EU trends and policy ranking or the sub thematic associated with them. Ultimately, data-driven EU affairs insights make the EU policy landscape more accessible to EU advocacy teams – leading to better decision-making and improved outcomes.