27 August 2019
News Startup Idea ('NewsCo')
by Wojciech Gryc
If you’re reading this, we’ve likely chatted or exchanged emails about the news industry and a few potential startup ideas. I wanted to share a more structured piece of writing to help with discussions and brainstorming.
NewsCo is an AI-driven fact extraction and analysis product that will help alleviate issues around fake news, enables workflow automation for journalists, and aims to be a billion-dollar content company.
The news industry is struggling with business models, high quality content (versus fake news or algorithmically generated spam), and other issues1. After conducting dozens interviews with individuals that spend over $1000/year on news content personally or via work, a number of themes are becoming clear:
- Content models are disaggregating, with subject-specific news sources replacing “generalist” ones2.
- Increasingly, people want to spend money on analysis rather than breaking news.
- People buying news content have no clear “job to be done” and use it for procrastination. Furthermore, spend is driven by ritual and identity (i.e., “I support the New York Times”) rather than business need.
- Investment professionals, business executives, and strategy executives spend significant funds on news feeds and sources.
Solution and Vision
NewsCo is a new type of news media company that creates high quality investigative content to help business leaders and individuals with decision-making. It is the first “tech-enabled” journalism company
NewsCo is building a new type of information product that scans content and automates aggregation of breaking news, facts, and stories. This aggregation is at the fact level3, and is directly tied to software components that can be automated to respond to these facts via workflows.
From a consumer perspective, three things are provided:
- Personalized facts and real-time breaking news.
- Deeper analytical content written/produced by NewsCo, written by tech-enabled journalists who can be better and more productive in their content creation.
- API access to the software and machine learning platform, so that NewsCo components and research can be integrated directly into other software and information systems4.
This sort of tech-enabled journalism company has a Total Addressable Market (TAM) above $1 billion USD5.
Timeline and Goals
Below is a short summary of our plans:
- Launch an Asian/Western news service to help business leaders, investors, and others track developments across both markets. This is a small-scale content creation and fact checking pilot, and we can validate whether there is a market for such news.
- Build the underlying platform to enable news aggreation, fact extraction, and workflow automation. An initial version has already been built using Airflow, Django, Postgres, and off-the-shelf machine learning tools6.
- See my earlier writing for an overview of the challenges this industry is facing. There are a lot of changes in the space today, and I believe a major part of the problem is the lack of structured thinking around how people read/consume the news, what they are willing to pay for, and how they incorporate takeaways into their daily lives.
- For example, we have sports-specific startups like The Athletic, The GIST(for Canadian women interested in sports), investor-oriented news tools like Eikon and Bloomberg, Sinocism for China news for North Americans, Stratechery for tech investors and executives, and so on.
- Fact-level news aggregation would facilitate automated fact checking for detection of fabricated stories, surfacing of less-popular stories, or other use cases. As far as I know, no one has framed news personalization or research in this way before.
- This technology layer is critical to enable workflow automation and profitable monetization of such a system. Such automation would be extremely valuable for investment professionals or business managers, and would help fund the broader analysis and content creation component of the company, especially early on.
- Via the combination of content creation, AI-driven fact extraction, and workflow/API automation, I believe this could become the biggest news/media/information company in the world.
- Below is a diagram of the current system architecture: