Google News: the full coverage feature
The origin story of Google News traces back to the ghastly terrorist attack on the World Trade Centre on September 11, 2001. After the attacks, Google Search struggled to provide relevant and up-to-date information to the people. Here is a screenshot of the search results page 4 hours after the attack happened.
Google had to resort to using its ad space below the search box to inform people about the events.
Because of the shortcomings of Google Search, Krishna Bharat, a scientist at Google, created a personal portal on top of Google Search technology that aggregated news from different news publishers as new stories got published. The personal portal got popular within the company and was launched to the public in beta in September 2002 as Google News. The launch helped Google further its mission of organizing the world’s information and making it universally accessible and useful.
Google News has evolved a lot since then and has gone through many significant iterations. The last major iteration was in mid-2018. It was a significant redesign which significantly leveraged artificial intelligence (AI) in many parts of the application.
- The use of AI/ML to analyze the information as it arrives in real-time and connecting it to already existing information by building a graph of people, places, and things linked in the story.
- A more personalized ‘For You’ tab to give information about the top stories, local stories, and stories around your interests.
- Newsstand Tab that makes it easier for you to follow the sources you trust and discover new ones.
- Full Coverage Feature — Provides a holistic view of a particular story from different publishers/news sources.
- Newscasts — a visual format that uses AI to give you articles, videos, and quotes on a single topic.
However, today, I want to talk about only one of the above enhancements — Full Coverage.
With Full Coverage, Google is using AI to solve a pain point for those users who are interested in getting all the information they can about a topic they are interested in; Users who want to assimilate all that information to understand different perspectives about the topic to form an opinion and discuss it among their circle of friends.
“If you want to get a deeper insight into a story, the “Full Coverage” feature provides a complete picture of how that story is reported from a variety of sources. With just a tap, you’ll see top headlines from different sources, videos, local news reports, FAQs, social commentary, and a timeline for stories that have played out over time.”
Outside of Google News, the typical journey of such a user persona is very painful and time-consuming; The user has to search for the information on different sources across the web — spending sometime even hours. And then bring all the information together to have a holistic understanding of the topic.
I’ve no idea how much of editorial curation is done by Google on each of the Full Coverage stories. However, as Google can do so for a large number of topics across the web, a significant part of the heavy-lift is definitely through the use of AI — using AI/ML to understand and connect the topics and information from different sources and organize it for the user.
To those who haven’t seen the feature, let me explain how the feature works by picking up a story that has the Full Coverage feature. Below are the screenshots of Full Coverage of the Brexit Story in the Google News app.
In the first screenshot, you can see the list of most recent stories on Brexit from the well-known publishers — NYT, BBC, Vox, and Guardian. The second screenshot gives a timeline of events/news related to Brexit over the past few months. The third screenshot gives Opinion columns from some of the renowned publications. The fourth screenshot shows social media coverage associated with Brexit, where Donald Tusk, President of European Council, is tweeting about Brexit on the Social Media. All of that (or most of it) is happening without any human curation. Isn’t that cool!!
Google News is used by hundreds of millions of users across the world. That’s why I admire the audacity of vision here and appreciate Google’s efforts to solve such a wicked hard problem; Especially when you put into context the amount of data generated over the Internet. According to a research report by Domo, users of the Internet generate 2.5 quintillion bytes of data every day, and 90% of that data over the Internet was created in the last two years only.
The data Google News has to organize will seem minuscule in comparison to that scale. However, it’s still HUGE by any standard. Figuring out which topics to pick across thousands of noteworthy news items from across the world and identifying the most critical information about those topics (by building connections between people, publications, content) using algorithms is no joke. Only a few companies can even think of developing such technology.
However, even for companies like Google, who have the world’s best search technology and are known for their AL/ML prowess, this implementation is difficult to execute to feel like magic to the end-users.
Let me explain what I mean by my the last statement by taking a few examples of news stories in the Full Coverage feature.
Example 1: News about Apple Watch swollen batteries
The first screenshot finds all the relevant stories related to the swollen battery news. When I scroll further down, the second screenshot has a timeline feature. Which ideally should tell how the story about swollen batteries unfolded, but both the news articles shown are unrelated to the central story: one talks about the Apple and Samsung Watch comparison and the other about 30% larger display of the new Apple Watch.
Example 2: News about the price cut by Apple on iPhones in China
The 1, 2, and 4 screenshots show a right mix of stories and social media links explaining the reason for the price cut. However, the third screenshot with Timeline lists many different stories about Apple from patent infringement, to credit card, to Apple retail chief leaving the company — again all of them unrelated to the news that I’m interested in reading.
Example 3: News about harmful apps on the Google Play Store on Android
What is the story about casting content on Android TV from Google Home in the middle of the story about malicious apps on Google Android?
Example 4: News about justice department warning Hollywood media companies from colluding to keep Netflix out of Oscars
In 2 screenshot, I’d have loved to see how the fight between Netflix and Hollywood unfolded over a timeline, but the first news articles on the Timeline isn’t linked to the story itself.
There are many more examples of such incoherent stories, but the point I wanted to convey was that these examples show that the actualization of the grand vision of this feature still requires a significant level of execution. If executed successfully, the whole experience will feel more like magic than just a cool feature for infovores like me.
If successfully executed, there are other applications where Google can reap huge benefits from such technology. In Google Search where instead of showing a list of links, taking users through a curated/guided experience by anticipating what the user is looking for and what other information will be useful in the search topic. Other exciting areas could be the field of education (Google is already investing significantly) where kids face the challenge of navigating the Internet to find relevant information.
Any other interesting ideas for applications of such a technology on Google products?
