Dainik Bhaskar’s in-house digital arm, DB Digital has come up with an AI-based powerful content recommendation engine called ‘Wisdom Next.’ The software helps targets its readers through advertisements using algorithms and AI. At the recently held Digital Media India 2018 event in Mumbai, Gyan Gupta, chief executive officer, DB Digital the working of the new software and how it can help predict whether a story will gain significant mileage.
Sharing insights into the power-packed engine, Gupta said, “Wisdom Next is an AI-based software that addresses difficulties related to audience engagement. It enables users to map real-time analysis of the content on the website. It provides details on visitors, page views, top stories and current trends besides generating reports regularly and offering ad-personalization based on the section, news and browsing history.”
AI-based approach
Gupta shared how voice is used as a medium to search for news or to read newspaper. Some of the AI-based systems have been installed by Google and will be soon adopted by digital publishing as well. DB Digital is currently working on voice and video integration in the system as it will give a feel of video being played after the text. Even the search algorithms for video will be different. Talking about how Wisdom Next’s AI-based applications have helped the company, Gupta said, “One of the crucial aspects that we as a team have been able to crack is that of video. Showing results based on a search for a video is quite a difficult task. So, we have developed a breakthrough solution by taking frames from the video and including the same for better search results. Moving further, we have also integrated a speech-to-text option, to make it easier for journalists to extract text from the same video. This also makes it easier for a converging medium like website that handles both text and videos.”
Such a system can not only save time but also valuable resources of a company. The new software has been able to bring down error rates by offering automated processes. If one has to categorize a sports veteran getting married to an actress, with human intervention, it may take time to decide the section the story will fit in but machine learning makes it easier by getting the job done faster and without any errors.
Being a Hindi daily newspaper, the audience Dainik Bhaskar attracts is mostly from the Hindi-speaking belts. Other than English, Hindi is preferred by 60% population in India to read newspapers. Moreover, a research conducted by KPMG and DB Group suggests that in India, the Hindi and vernacular newspapers are more loved by news readers. There is a vast majority interested in consuming news in Hindi and other vernacular languages on digital platforms. Given the demand, DB Digital’s Wisdom Next team is working on personalization of content along with ad-personalization.
Due to their efficient team on board, the Dainik Bhaskar page is the fourth most engaged page on Facebook globally while it tops the list in India. Speaking on their future endeavors, Gupta said, “We are still developing the engine based on insights we receive. We are delving deeper into the analysis, so that our advertising can move with the readers and follow their pattern. Soon we will introduce this software to the industry. We feel that if we are able to generate better results with Wisdom Next, we can also pass this knowledge across the industry for a better tomorrow.”
For the editors and reporters, the software offers several updates on a story, based on which one can select the story angle or choose to select the right detail and decide on the presentation perspective. It can even predict, based on the article, whether the story will receive good mileage or not. Interestingly, it can also suggest the right time to take the story online for better page views. It can curate a list of top 100 stories which are popular on social media. These top stories can be further bifurcated into a dedicated list for stories on Facebook, website and UC Browser. Almost 40% of digital traffic is driven through Facebook to the DB website.
Mantra for audience engagement
As per Gupta, the following are the mantra to go by for better audience engagement:
– Target keywords as indicators of engagement and identify high-relevancy keywords
– Use Associate Mining to identify stories that are doing well; a score based on performance is given for future analysis.
– User and their interest stories are monitored. At DB Digital, the reader/user history is taken into consideration to identify their patterns of engagement based on what they have read before.
– Showing similar stories to users. Here story stitching can be of great help. This brings people together who have shared similar patterns and allows displaying to readers the most frequent stories trending in their pattern.
“Today, with a rise in digital spends, there is a growing need for a system that can track the engagement and manage ad and content. This is what every publisher will need in the coming future and we are ready for it,” concluded Gupta.
