“You will never find any other app on Playstore that makes a millionaire within two clicks,” details a video on YouTube channel Technical Boss. This is just one of the hundreds of videos on how you can earn money from NewsDog, a news aggregator app that is making waves in India.

Don’t let the name NewsDog fool you. The Chinese app backed by internet giant Tencent, is not a news aggregation app. Topping the charts on Google Playstore with 50 million Indian users, NewsDog is popular for its entertainment value and regional content — and money generating schemes.

NewsDog, launched in 2016 by CEO Forrest Chen Yukun, a computer science graduate from Tsinghua University, and Yi Ma, who holds a PhD from Princeton University and previously worked at Baidu and Goldman Sachs, ran into unexpected success in India. But if you look at the Chinese playbook of entertainment apps, the popularity of NewsDog in India is not as sudden as it seems.

Toutiao, a Chinese equivalent of NewsDog, has had a terrific run in China. It has around 200 million daily users, with the platform touted to be highly addictive among Chinese users. Known for its racy content, the platform has grown big enough to attract the government watchdog’s attention to censor inappropriate content.

NewsDog is taking a leaf out of the Toutiao playbook in aggressive user acquisitions in India. The app often crosses the line when it comes to racy versus risque. Sample some of its daily videos and posts: “Hot photoshoot of an actress,” “Heroine goes nude,” “One of the sexiest adult star’s pictures,” “Actress forcibly kissed by an actor,” “Bikini pictures of an actor”… all presented under the name of entertainment content.

“The news business as it is, is a latent need, not an immediate need, which is why notifications about Aishwarya Rai and Akash Ambani get hits,” says Sandeep Amar, a senior Indian digital media executive who is building Inaaj, a startup that is deploying artificial intelligence to better serve users.

The app runs in nine Indian regional languages besides content in English: Hindi, Bengali, Gujarati, Marathi, Kannada, Punjabi, Tamil, Telugu, and Malayalam.

Under its hood, NewsDog has a mix of artificial and human intelligence at work, says Chen. “We drive downloads a lot from WhatsApp campaigns, as well as from Google and Facebook. Also, NewsDog is adopting blockchain technology to incentivise user for referring. We are working with a blockchain project Content Neutrality Network (CNN) to make it happen,” he says in one of the two conversations he has with FactorDaily.

Much like Toutiao that used the power of algorithms to identify consumption patterns and push relevant content to increase average user time spent on the app, NewsDog, according to Chen, starts understanding the user’s preference from day one to push likeable content to the user.

Deep tech, human smarts not enough

Half of its 50-member team in India and China works on building algorithms for a better user experience. The code uses machine learning to build user profiles and classify content on the app. Metrics such as what people read in the past, keywords extracted from “liked” and “shared” articles, gender and location of the user, among others are fed into the algorithm.

Chen is candid when he says NewsDog is not a news app. “It is more about content,” he says. “You can call it an entertainment app.”

He adds that the company is aware of its “content problem”.

“Currently, we are labelling all our content using both our algorithms and through manual checking,” says Chen – in all the languages it operates in India.

If there is an inappropriate keyword like “sexy” which in some context might not be considered vulgar, it is labelled inappropriate level 1. But, if the inappropriateness is higher, like in case of a word like “nude,” it is labelled inappropriate level 3 and it goes through further moderation to determine whether or not that content should be published.

NewsDog is also using open-source image recognition tools such as Yahoo’s open source NSFW (Not safe for work) model to label pictures. On a range of 0 to 1, the tool recognises the level of inappropriateness of the picture. If the range is close to high, the images are checked manually.

Also see: How LinkedIn uses Artificial Intelligence to keep NSFW content out

Chen admits that the team hasn’t been able to nail the moderation of inappropriate content on the app due to its limited understanding of Indian languages and content. The team now is heavily investing on language-based moderation, he says.

“I don’t want it to be known as an app for semi-pornographic content,” he says. “That’s not our message and vision for the company.” The vision of the company, he adds, is to build a platform where people come and contribute in form of regional content and NewsDog as a technology company will ensure that the content reaches a relevant audience.

Content market not so content

Despite its avowed “not news” positioning, NewsDog competes with DailyHunt, Times of India, Inshorts, and other popular news aggregator apps in India. DailyHunt, a local language news app that says it has 90 million users in India is backed by the likes of Sequoia Capital and Toutiao’s parent company Bytedance. Inshorts, a news curation platform has a user base of 10 million users and is backed by Tiger Global.

Much of the content on NewsDog, as also of its peers, reflects market needs: young and local language users. Regional content platforms are soaring in popularity with growing internet penetration in India. Nine of every 10 new internet users in India are likely to be Indian language users, according to a KPMG-Google report in April 2017. Industry group Internet and Mobile Association for India identifies young students as the most prolific cohort among India’s over 450 million mobile internet users as of December 2017.

Also see: Inside DailyHunt’s bid to become India’s content King

According to an app tracker App Annie, India recently topped the list of world’s fastest market in terms of maximum mobile app downloads. It ranks Twitter, DailyHunt and Times of India as the top three news apps on basis of monthly active users as of March 2018.

India, therefore, is a daily battleground for regional content players like DailyHunt and NewsDog. DailyHunt has 40 million monthly active users who on an average spend 20 minutes on the app daily.

Chen says that NewsDog has 18 million monthly active users with over 100,000 new users signing up every day.

The app gets the most traffic from Tier 2 cities – Lucknow tops the charts. Delhi, Mumbai and Bangalore also generate significant traffic. Most users fall between 25 to 30 years age group; 80% users are male.

The metrics of average time spent on the app daily is critical to a content business like NewsDog. With the overdose of content being pushed on the internet, an aggregator that curates personalised content drives traffic on the app. Global businesses are known to invest heavily in algorithms that hook users on the app. The average user of Toutiao spends 74 minutes on the app every day, thanks to its deep learning algorithms that are even capable of generating content.

To increase ad-based revenue, the company needs more and more people to sign up and spend more time on the app.

Chen says the company’s vision to achieve that is through generating more content in different media forms. Popular video content dramatically increases the average time users spend on the app. The company is trying to source more video content to cater to the liking of their audience.

The company says in the next few years it plans to invest more in sourcing regional content in form of text, images and videos.

Currently, besides news publications and wires, NewsDog has a network of about 10,000 content writers that publish on the platform. Based on their contribution, the top 100 writers make about Rs 5,000 a month on an average.

Gamifying app distribution

Arman Ansari, 22, who runs a Youtube channel where he posts app hacks has been using NewsDog for the last few months. He says that he uses the app for content as well as a means to make money. So far, Ansari says he has made Rs 6,000 through NewsDog.

The app, as a means to user acquisition, offers newbies Rs 50 as a sign-up bonus. Subsequently, you get coins for spending time on the app, which can be cashed out through a Paytm wallet.

The inner mechanics of the app are very simple. A user can browse content and based on her or his “likes” and other metrics such as, what people read in the past, keywords extracted from the article, gender, and location, the user is served personalised content.

(Also read: Inner mechanics of Bigo Live, the money spinning app Indians are going crazy about)

At this stage, the aim of NewsDog is to make users spend more and more time on the app. Besides pushing personalised content to every user, the app also incentivises the user by crediting 10 points to their accounts for every hour they spend on the app. A user can come back after an hour to check and get another 10 points. Once you accumulate minimum 200 points, you can make a withdrawal through Paytm.

The app also offers other ways to make money on the app– this includes sending app link to WhatsApp contacts, visits to the shared link, and effective usage of the app by a referred contact. This is probably the reason why YouTube is full of videos on how to hack NewsDog to make money. The company doesn’t shy about promoting its money-making schemes.

A daily scoreboard screen on the app indicates how much money its top users have made on the app. It can go up to Rs 5 lakh for some users.

Sunny Singh from Gurdaspur in Punjab has been a top earner on NewsDog for a few weeks now. Singh says he has made Rs 3.5 lakh on the app just in the last three months. The 17-year-old, 12fth grader runs two YouTube channels PayTM Men and Techno King where he posts videos about apps that help users earn money. Singh says he urges users to download NewsDog using his referral link and has got the app at least 50,000 new users.

Chen holds his cards close to his chest when it comes to disclosing the marketing dollars the company has spent on acquiring users. All he will say is that the average cost of acquiring users was as low as Rs 3 – a stunningly low number among apps in India.

He explains the math: Say, if you get 10 rupees for referring a user, you will get the money in seven days time. During these seven days, the user you referred has to use the app every day. If the user doesn’t come on the app, you don’t get the money. Most people who are not interested in the content, don’t continue to read it for seven days. If you stick to reading on the app for seven days, you are probably a real user.

“We think this is a good way to acquiring users, otherwise you need to go to agencies and pay them to get some traffic on the app,” he says.

Inaaj’s Amar credits incentivised app downloads to the high numbers at NewsDog. But, he adds, it has no connection with retention of those users. “According to Comscore data of May 2018, NewsDog ranks 28th in the total number of unique audience (in India) for the app. A business marketing merely on incentives and racy content is difficult to sustain,” he concludes.

Respecting copyright, growing

As for users who have earned as much as Rs 5 lakh on the app, Chen says it is a fair number. “Many of these guys are YouTube influencers or run targeted campaign to get more users to sign up. They actually do bring in hundreds of thousands of new users on the app every day,” he says.

Since NewsDog allows freelancers and independent content contributors to post on its platform besides using stories from news wires and other publications, it has to exercise special caution to prevent copyright disputes. The platform uses algorithms along with manual verification for this.

After the initial filtering of sensitive keywords related to offensive content like porn or race discrimination, the detection system partitions the whole article into sentences and uses each sentence and runs it through the NewsDog’s database as well as against content on the internet.

If the number of the duplicate of sentences in the article is found to be high, it will be categorised as infringement. If the number is quite small, it is ignored and cleared through the filter, and an article is directed to a manual detection process if the number is in mid-range.

The company is looking to hire an India head to manage operations in India. Of its 50 staffers, a dozen work in India. Besides this, there is also an outsourced network helping with moderation of the content.

NewsDog plans to hire around 100 more people in the next few years as part of its operations team. The company plans to spend the money raised from Tencent on building more technology, hire better people to improve the distribution algorithm and improve the quality of content.

---