In early human civilizations, governments didn't need to analyze news media, because they fully owned and controlled all publications. As early as the Han dynasty in China, around 200 BCE, imperial bulletins known as "dibao" (邸報) were circulated amongst bureaucrats, providing them with tidbits of regional news and propaganda to publicize at local meetings. When government was the only media, institutional reputation management was a breeze.
A couple thousand years later, the rise of free press in democratic societies had started to make reputation management harder. One day in the early 1880s, a Latvian immigrant named Henry Romeike was chatting with an acquaintance on a park bench in Paris. He was unemployed and broke, hoping not to have to return home defeated. Don't companies, politicians and celebrities need a systematic way to know when they've been mentioned in the newspaper, he thought?
Romeike would move to London and open the world's first news clipping bureau. Today, media intelligence is an estimated $4 billion global addressable market, encompassing thousands of employees serving every major company.
In my career, I've worked closely with more than 100 companies specializing in the interpretation of news, social media and other public opinion datasets. A confluence of forces is quickly reshaping this space, most obviously generative AI, which promises to make business insights cheaper and faster. Methodologies are growing more advanced as well. Where a few years ago most companies where simply gauging sentiment polarity, custom solutions are now enabling researchers to identify nuanced themes and key messages in social media trends and news coverage. Smart companies are asking not "do people like us?" but rather "Do people see us in the specific way that furthers our objectives?"
Best practices around the use of new AI tools are emerging. Analytics platforms that provide real-world, verbatim citations from human-generated text to back up their assertions are proving more useful than those which offer pronouncements without proof.
Meanwhile, researchers across the private and public sectors are facing 3 related headwinds with regard to data quality:
1) Short-form shortcomings. Short-form video and other bite-sized content formats offer less detailed insight than long-form sources, such as consumer reviews and blog posts.
2) Audience fragmentation. Where social communities previously congregated on just a few platforms, a broad range of niche platforms and enthusiast communities are rallying with a vibrancy not seen since the 2000s. While that's great for social media users, it presents new challenges for researchers and analysts.
3) Data gentrification. Platforms seeking to monetize AI-based use cases are creating new revenue programs which present new license opportunities for technology builders. Like an up-and-coming neighborhood with cool restaurants and high rents, these programs sometimes come with increased costs, requirements and restrictions.
We've come a long way from the one-dimensional media environment of the dibao. The complexity of the modern landscape is inherent to an unprecedently diversified media ecosystem, one in which content around imaginable topic of interest is available to explore.
Comments