"Platforms like ours wouldn't exist without our partners": Ravanan N, vice president, content, Dailyhunt

The man who spent his 'entire adult life' in Dailyhunt, divided the story of the country's biggest news aggregator app in four phases. "But a common narrative through our entire journey," he said addressing the audience, "would be our intention to address the greater Bharat - Indian language content - and also to work with the partner/publisher community to add value to their business."

Ravanan N talking about "The Dailyhunt story" at Digipub World

Dailyhunt has over 160 million app installs offering 100,000 news articles in 15 languages licensed from over 800 publication partners every day. The Dailyhunt group has over 90 million users with seven billion page-views spending over 3.7 billion minutes on the service every month.

Excerpts from Ravanan's speech:

Phase 1

The first phase kicked off in 2007 when we launched the first B2B product, a white label product, for telecom operators. The vision was to bridge the information divide. It sounds lofty but I will break it down. The idea was to bring internet on to mobile devices of users who do not have an internet connection. We started with SMS alerts. Let me share an anecdote about how it all started

In 2007, we went to telecom operators with a product that promised free tech search. It sounded 'kickass' to the operators as well as us. The user could go to the message box and type in what he wants and an algorithm would figure out what he wanted and give that in the form of alerts. The first one was job alerts, the operator agreed and we launched it.

We now had a problem of communicating it to the customer. In those days, the only way of using short message text to communicate was with 160-characters. So, we had to stick to this limit, craft messages cleverly and send them off to the post-paid base. We got 20,000 responses, of which 17,000 were the same text that we had sent. We then realised that most of them did not understand the product and, second, they did not understand what had been texted and so did not know what to do. Realisation: a product may sound very good on the drawing board but it may really connect with the users.

So, we moved to a menu-based product to help the customer navigate to the product. We ran it for a year-and-a-half but realised again that there was a roadblock in terms of penetration. That was because everything was in English. That is when we decided to break the English barrier and go with a voice product. Voice is technically literacy agnostic. That kind of helped our growth in the telecom space. From then on, we launched search services with telecom operators across India, went international and launched it largely in developing countries.

Phase 2

The second chapter was around 2010-11. We saw the mobile internet revolution unfolding and wanted to be there. That is when we moved to B2C with an app called Newshunt. Run by a tech company based out of Bangalore, Newshunt - back in 2010 - rendered Indian language content to feature phones. The tech company had developed a technique of on-the-fly conversion of a web page to an MMS image and deliver it on to mobile devices (largely non-smartphones).

In 2011, we raised money and invested in product technology and big data, which drove our recommendation engine. This was the pre-AI period. We were able to recommend what the user is more likely to consume based on his past data. We also invested time and energy in the delivery of content for low bandwidth scenarios. And it was very difficult to deliver rich media content on a mobile device.

At that time, we saw a lot of growth happening on Android. Data became more affordable. But there came competition as news organisations started launching their own apps. We invested on two critical elements - the customer and our partners. We realised that if the partners didn't exist we didn't exist and that we had to give them some reason (money) to be on our platform

Phase 3

We began investing a lot on ad technology and also in scaling up our ad sales team. We don't primarily depend on networks. We do our own direct sales. Our efforts bore fruit and we grew 20X on traffic and users. The number of partners we had grew from 80 to 1,400 (now we have 1,800 partners). We moved from five languages to 15 languages.

Phase 4

While this was happening the fourth major turn in our story was around 2017 and which coincided with the launch of Jio. Data became extremely affordable. People from B and C towns who did not consume content in English (India's literacy is 74 per cent but English literacy is only 12 per cent) went to consume content in their mother tongue or language. With Jio, everybody had enough data on their phones. That is when we doubled down investment again on technology and invested a lot on machine learning and AI. There is a lot of machine and human assisted work and we have a huge back-end of teams sitting and help the machine learn. My team of 450 sits and goes through 42,000 pieces of content every day to enhance it.

We also invested a fair amount of efforts on our ad deliveries. We are probably the only company that can deliver ads at a pin code level. We use GPS and we try to give relevant ads and relevant content thanks to our association with mobile operators.

Till recently, we were on rich text but we realised that we have to give more reasons for users to come and consume our content and engage with our platform. So, we decided to pivot into videos and are moving to a multimedia platform. Our mission is to be the Indic platform, empowering a billion Indians to discover, consume and socialize with content that informs, enriches and entertains.

Was it all hunky dory?

We have had our fair share of failures from Day One. Our first failure was when we designed a product but overestimated its acceptability. Our first pitch to Airtel was that we wanted to be the Google of Mobile. From then on, we have moved really far but also had adventures.

We went into e-commerce and launched the e-books portal. We realised that no Indian language content existed in digital form that can be consumed. We went around the country, spoke to publishers, picked up manuscripts, handwritten notes, PDFs, Corel files and digitised them for distribution.

We were doing fairly well, but were probably way ahead of time. We had to heavily discount and subsidise the book. Selling books at '1 and it did not make economic sense. At some point of time we had the largest e-bookstore in India, bigger than Amazon and Flipkart. The whole thing folded up as we realised that heavy discounting was not going to work. The entire VC community had moved from GMV to gross margins and we had to toe the line.

We also launched subscription on books and magazines and also solved the problem of payment through credit and debit card on mobile by launching iPay, which allowed users to pay for the digital goods through a prepaid balance or post-paid bill. That was a kind of game changer. With that we were able to launch chapter-wise billings on magazines. With Delhi Press' magazines we sold each chapter at '1. We sold each chapter of Amish Tripathi's books at '5. We also had an exam prep product that helped aspirants prepare for government exams in their own language.

Key learnings

I quote founder Virender Gupta: "Keep your head down and your eye on the two important stakeholders - the customer and your partners." Platforms like ours wouldn't exist without our partners.

Second, there is a lot of distraction. In the start-up space you find a lot of noise being made by new kids on the block. Keep your head down. Finally, if you have to let go take a hard call and let go and move fast. If you hold on, it could be very dangerous for you.

In the end, only someone with a focus on providing content which the consumer is more likely to consume and engage and thereby making it monetisable for us will be the winner.

How Dailyhunt uses AI

We get around 118,000 articles a day and have round nine million people coming to us. Right now, it is man-assisted, but machine-driven. After we launched an LR based model (a statistical model that tries to predict what content is more likely to be consumed), the time spent has gone up by 25 per cent.

On increasing interactivity with video

These are very early days for us as far as video is concerned. Right now, we are going to the bottom of the pyramid, to the long tail of content on a platform and that itself is giving us huge numbers. I think this month we should cross a billion video views. The biggest problem we face is how to make the content discoverable and are in talks with tech partners to make video viewing more engaging.

Why direct selling and not agencies?

If you have three parties in the entire ecosystem, more money is made and that is better. The direct sales CPM is much higher than a network CPM. The idea is to get more involved with the brands and not just to display advertising and to see if we can do 'solutioning' (we do a lot of solutioning-based deals now). Unless you have a direct sales team, it is not going to work. It has helped numbers. We own OneIndia, the largest Indian language platform for content on the web. It was primarily network driven when we took over. We moved it to direct sales and saw a 45 per cent growth in our overall yields.

Difference in what people read, watch and consume in various languages

There is a huge difference. Movies and entertainment are No. 1 in Telugu, news is No. 1 in Tamil and Bangla. In Marathi, politics is bigger than anything else. For Kannada, it is music and sports that does well. There are unique behaviours that happen across geographies. And it is also to do with the amount of content available. In Kannada, for example, we have a problem with not enough content being available. What is available is news, sports and music. Probably that is why it is like this.

Digipub World 2018 was partnered by - Timesnownews.com (platinum partner); Akamai and Facebook (silver partners); and Freshworks, Vidooly, comScore, Quintype, Times Internet, and 24 Frames Digital (bronze partners).