Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
How Raju Vanapala is using AI to reinvent the business of news
221 Plays1 year ago
With over 2 decades of entrepreneurial experience, Raju's journey from bootstrapped beginnings with Way2sms to pioneering the hyperlocal news platform Way2News showcases his remarkable resilience and vision. Gain unique insights and lessons learned as Raju discusses conceiving business models, organic growth strategies, leveraging vernacular content, and capitalizing on events to drive engagement. Tune in for an insightful exploration of how Raju continues to disrupt and redefine the landscape of Indian news consumption.
Get notified about the latest releases and bonus content by subscribing to our newsletter at www.founderthesis.com
Read more about Way2News:-
1.Way2News taps into rising demand for vernacular content with local flavour
2.Way2News founder-CEO Raju Vanapala on serving hyperlocal news for Bharat
4.Today news isn't a matter of language, but rather how quickly it is available
5. How Way2News transformed the hyper- local news ecosystem in India
Recommended
Transcript
Introduction to Raju Vanapala and Way2News
00:00:00
Speaker
Hi, I am Raju Vanapala, founder of Way2News.
00:00:15
Speaker
We here at the Founder Thesis Podcast love stories of entrepreneurs who have spent decades building ventures. There is a special humility and clarity of vision that comes from having seen a journey of entrepreneurship for so long. This is what makes this conversation special.
00:00:31
Speaker
Your host Akshay Dutt speaks to the veteran founder, Raju Vanapala. Raju became a bootstrapped entrepreneur way back in 2005 and his first venture, Way2SMS, attracted so many users that he actually had to throttle growth to manage the computing infrastructure.
00:00:48
Speaker
Raju is now building a very unique, hyper-local news platform, Way2News, that is disrupting how millions of Indians get their daily news briefing.
Way2News: A Mobile-First News Platform
00:00:57
Speaker
Stay tuned for a rollercoaster ride into the world of entrepreneurship on The Founder Thesis Podcast. So, to start with, can you give me like a
00:01:17
Speaker
You know, like an elevator pitch of what is Veto News? Well, Veto News is a digital platform of bringing short, relevant and reliable news stories to a billion Indians in their local language by local sources that can be trusted. Okay.
00:01:39
Speaker
So I mean, what category would it fall into? Is it like a social media platform with user-generated content or is it a news? It is a perfect news platform crafted for mobile experience or mobile-first audience. So you would be competing with something like a Times of India? Those would be comparables? Or would it be like a... Well, partly we have three competitors. One is aggregators. I mean, the digital aggregators that I'm talking about.
00:02:08
Speaker
Give me an example, help me understand the space a little more. A big daily hand is an aggregator. They aggregate content from a lot of sources. And then in shorts, they also aggregate a lot of content from other sources. And then when it comes to traditional, the second one is the traditional media houses. Typically the local language, traditional media houses, be it Mano Ramatanti or Iradu or Sakhshi or Jyoti. Times of India.
00:02:38
Speaker
Times of India belongs to English language, right? They are not a direct competitor to us,
Challenges and Competitive Landscape in Local News
00:02:44
Speaker
because we belong to a language. But something like a Hindustan would be... Yeah, Hindustan would be a competitor to a Hindi language. Very true. The third one is digital content creators. Typically, they are content creators for internet audience, right? So, these are the broad... What are examples of the third?
00:03:03
Speaker
The third, like an independent journalist will be writing content for users. Like the wires, all these guys crawl. Yeah. There's so many such sources. There are thousands of such sources available out there. Okay. Okay. I'm mostly aware of the English ones, not so aware of the vernacular ones, but yeah. Okay. Got it. Okay. So many thousands of such destinations, even in local language.
00:03:30
Speaker
Hmm. Okay. Because it's essentially just a blog, right? Anyone can start a blog and it becomes a news. There is no entry barrier in this business. Anyone with a decent journalistic knowledge can start off a blog or in the form of a website and will become a destination to users to consume content. Yeah. Like you can have a Bollywood gossip news site, which would have started as a blog and then it became a news site. So got it. Okay. Okay. That's how many of these became popular.
00:04:00
Speaker
Right. Right. Right. Okay. So these aggregators you spoke about daily hunt and the other one, sorry, which was the other one you mentioned? In shots does a little bit of aggregation. Yeah. So they don't do their own content, but they essentially curate and give you a personalized feed. Their USB is about recommendations, algorithms, suggestive algorithms, all that. Generally this was content from
00:04:30
Speaker
various local newspapers or blogs or independent journalists and they aggregate at their destination. And they will show you like a preview and then you click to read the whole article. Is that how it is or is it like a endless scroll where you can just read article after article, just keep scrolling. What is their format? Everyone has their own way of publishing content. Some of these guys have one story in one screen and some of these have infinite scroll.
00:05:01
Speaker
wherein some of these, they provide just the headline initially, how to click on it. You can then read the whole story. And why do the news platforms work with aggregators for traffic, like they bring them traffic? Obviously, right. They also require a distribution and aggregators and somehow got funded well, they have a better reach than independent websites. So they require distribution.
00:05:26
Speaker
Okay. Okay. Although, I'm guessing that publishers would be at the losing end of the deep because a lot of people may not click through and the aggregator would be earning more from ad revenue than the publisher would be earning from ad revenue because many users will just read the headline, get a broad idea, then move on. See, you know, for a new guy, it's a win-win because he does not have a distribution.
00:05:55
Speaker
Okay. Platforms, aggregators will provide that distribution. Maybe for a matured guy who is in the market for a long time, who has a brand or a build already. For them, it is definitely not going to be made. Okay. Initially, they might be looking for some revenue through their content distribution, but eventually it will kill their business.
00:06:18
Speaker
Hmm.
Editorial Process and Content Sourcing at Way2News
00:06:19
Speaker
Right. Right. Which is what's happening with Google now. Google news, that Google has this product called Google news, which allowed you to just within the Google news app, you could read a lot of articles, which now publishers are kind of fighting back against. Yeah. There's a massive backlash. Even in other countries, they were able to pull up some revenue from Google as well. Hmm. So Google news is also one of your company. This would be an aggregator. Yeah, obviously. Yeah.
00:06:47
Speaker
Okay. Okay. Okay. Got it. Interesting. Okay. So what is the, how do you make revenue then? Is it like ad supported? Well, yes, we monetize through advertising. We run ads in between news articles, but before going into monetization, et cetera, I'll explain you how our app functions. Basically we provide short summarized news in local language. We provide, uh,
00:07:17
Speaker
content that is vetted by a tutorial staff. We source content from traditional networks such as writers, ANI, PTI, all these guys. And we summarize content and we provide the most important news of that day to our consumers. And we also crowdsource content, a lot of these hyper-local content from Stringer network and also from Citigen journalists.
00:07:46
Speaker
So you run a newsroom or you run an asset like newsroom in a way that none of these people are on your payroll. Yeah, not none of them. We have our own set of editors here. I have more than a hundred plus editorials team here setting out. And editorial team does curation or do they also? Yeah, they also summarize content. Okay.
00:08:11
Speaker
giving it a catchy headline, writing a short summary, things like that. Summarizing an art, Akshay, so it's not an easy job. Yes. We realized it's not an easy job. Initially, we thought, okay, it's easy. I agree. Then we realized it's very tough to craft content, especially skimming the content is not that easy. And without losing the essence of the article, and I would say we have a craftsman here, they would be doing crafted content for our users.
00:08:41
Speaker
Okay. So the editing and curation team is on payroll. The people who are actually creating the articles are paid per article or something. This is how it works. As I mentioned, right? We provide content from a national geopolitical content. We provide movies content. We provide international content. We provide content from business and sports categories. All this will be crafted by our editorial team.
00:09:08
Speaker
And we also provide the hyper local content. This will be crowdsourced from a city in generalists or a string the network. So what is this trigger network?
00:09:18
Speaker
We have built a Stringer network from various towns. What does that word mean? Stringer? I've not heard that word. Stringer is an independent guy who will be roaming around for sourcing the news content. And he will then share that news content with various publications. And if that gets published, then he will get his payment.
00:09:39
Speaker
Okay. Got it. So like the paparazzi who click celebrity photographs and they sell it to the highest bidder. So got it. Okay. And traditional newspapers or all of these guys have a single network built already. But in a smaller scale, when it comes to greater news, we have a large scale network. We have our network went deep into a village level.
Growth Strategy and User Metrics for Way2News
00:10:06
Speaker
Okay. So you decided to be more hyper-local basically. Well, that's where we felt media today is not inclusive is what we seriously felt. And then we figured out that is the gap that we have to address. And as traditionally we have seen a lot of stories that are coming out of the city will get disseminated till village level. That means from top to the bottom, the content is disseminating.
00:10:34
Speaker
But from bottom to the top is not happening. So that's where we thought of disrupting the media. And we have built our network. That's how we have differentiated our content today with other traditional sources. So we have stories that are coming out from villages, and that gets populated across a state nation. In fact, there are so many cases where we have broke the story first, and that became a state level issue later on.
00:11:04
Speaker
Okay. Okay. Okay. Interesting. And from the ANI and PTI, these are like, these are called news wires, right? So these are like organizations which you can subscribe to and they will continuously send you news stories. They provide national and international news feed because we don't have reporters in, I mean,
00:11:27
Speaker
across the country or across the international geographies. So we rely on these sources for national and international content. But for local content, we have built our own network. Okay, got it. And today we have the largest network than any other media. We have almost 12,000 odd citizen journalists contributing every month
00:11:50
Speaker
and we get around 15,000 news pieces every day and we publish around 7,000 to 8,000 odd news pieces from Telugu and Tamil states alone. Okay, so 15,000 is what is total published or you receive and some of these you publish? We receive around 15,000 every day but we publish only around 7,000 to 8,000.
00:12:12
Speaker
And why is that? Because your editors decide to say... Some of it may be a repetitive content, some may not be a newsworthy, some may be a promotional content, some may be a fake news. We have a mechanism where EI does the basic checks. EI discards at least 30% of the content. And again, another 20% will be moderated by our staff and then they'll publish the content.
00:12:39
Speaker
Okay, interesting. Got it. So you're only in these two states, is it? You said Telugu and Tamil Nadu. Well, the app is available in eight languages right now, but our hyper-local content is available in these two states. And we are also dominant in these two states right now. Yeah, we started off with Telugu initially, as I mentioned, because we are a bootstrap company, I have limited resources.
00:13:09
Speaker
So going deep into the villages requires a lot of resources in terms of people, money, all that. So I took one state initially, and we made it successful here. And until later on, we extended the same playbook in Tamil Nadu. It worked really well there. So overall, we are going market by market. We are not in a hurry to go after all languages at a time, because India is a vast country. It requires a massive set of, you know, massive set of
00:13:40
Speaker
Stringer network and it requires a lot of money even to build a distribution across various languages.
00:13:48
Speaker
When you say build distribution, does that mean build sourcing of news or build like a way to acquire customer? User acquisition. Yeah, user acquisition. Sourcing content is something that we have built a playbook already and we are experts in doing so. It takes hardly three to four months of time in reaching out to the existing stringer network and existing reporters or journalists.
00:14:12
Speaker
And as your platform gets into the people... How do you reach out to them? Is there like a place where they list themselves? A lot of it is through research, a lot of research that we do on area level, who are the prominent stingers. And mostly these guys are available on Facebook.
00:14:32
Speaker
and we would be promoting their articles on Facebook. Yeah, promoting their article through Facebook and we reach out to them. Initially we have built our network through Facebook and later on we started getting reporters in housing, inbound inquiries through these reporters. Word of mouth spread basically. Yeah, through word of mouth majorly. Okay, got it. So that's how you build your supply. How do you build distribution then? Well,
00:14:59
Speaker
As I mentioned in our last conversation, the product has the ability to acquire users on its own. So unlike other products, our product UI, I don't know whether you have seen the product UI or not. It's easy to use. The product encourages the users to share content. And we acquire a lot of organic users through these shares. And today, we get around 30 million shares every month.
00:15:29
Speaker
And this is like shared on WhatsApp, shared on Facebook. And we get organic users to tune off almost six to seven lab numbers every month. That's a huge number for a news app getting half a million plus organic installs. Amazing.
00:15:54
Speaker
Content is your distribution strategy. Basically, if you have good shareable content, it automatically builds distribution because people share. And then when you see the app and you read the article, it takes time. But yeah, organic is one major source and off late in Tamil Nadu, we started investing on Google and Facebook to acquire users, especially the quality audience.
00:16:18
Speaker
The platform requires an initial take-off through Google and Facebook acquisition and later on it can acquire users on its own once it reaches to a certain scale. And the same is happening in Tamil Nadu. Today we are acquiring users organically. We are not spending much in Tamil Nadu now. What scale do you need to reach? Okay, amazing. Before organic kicks in, what is the threshold?
00:16:45
Speaker
It is a million plus DAO is what we need initially. And then from that million to 5 million journey would be organic journey. Amazing. So how many do you have in Tamil Nadu and how many in other? Well, overall, we have 6.5 million DAOs and more than 13 million MAOs today. And we enjoy the highest MAO to DAO ratio of more than 50%. That is a high stickiness for a news product in a news category, especially.
00:17:15
Speaker
And if you ask me the breakup by language-wise, Telugu, we have close to 5 million DAOs. And Tamil, we have close to 12 lakh DAOs, 1.2 million DAOs. And remaining languages are hardly like 2 to 3 lakh, not much. OK. So just for my listeners, so DAO is daily average user, and MAO is much the average user. DAO is a daily, yeah, daily active user. And MAO is monthly active users. Yeah, daily active users. Yeah, sorry. OK. And so you have a lot.
00:17:44
Speaker
Right. Yeah. That's the most important metric for. And there's another critical metric that I would love to discuss with you, Akshay, the retention. How many users you are retaining by next day and how many users you are retaining by 30th day. We call it as a D1 and D30. I believe we are enjoying the highest retention in the category. Our retention hovers in mid 40s, our day 30 retention.
00:18:14
Speaker
Okay. Okay. Okay. Okay. Okay. Interesting.
Revenue and Monetization Strategies
00:18:18
Speaker
And this you're comparing with aggregators and new sites both. I'm comparing, especially in the news category. I don't know much about other categories. Generally gaming category probably might have a better retention, but news category, we have the best retention in the country. Okay. Okay. Amazing. Amazing. And what is the way in which you monetize?
00:18:44
Speaker
Well, again, when it comes to monetization, we monetize through ads. There are two ways to monetize. One is through national brands. We go after the same set of large brands out there. Maybe there are around 5,000 brands who will be spending on digital. We also get some pie out of it. That is one. The second unique monetization is through local advertisers. So these are a unique set of advertisers that does not belong to Google and Facebook planet.
00:19:12
Speaker
They don't understand Google and Facebook or how to advertise in Google and Facebook. Traditionally, these are newspaper advertisers. They generally advertise in the first page or third page or last page of a newspaper. So these guys typically spend anywhere between 50,000 to maybe 50 lakh per month. So we have a big set of advertisers.
00:19:34
Speaker
We have a chain of mobile handset sellers, we have chain of hospitals, locally built their local brands, their prominent local brands here. They're spending heavily on way to news today and 50% of revenue or revenue today comes from local brands and another 50 comes from national brands. Okay, so these local brands are like the kinds who would be like a just style customer.
00:20:01
Speaker
Yeah, it ranges from just their customer to a chain of hospitals or chain of restaurants or a chain of shopping malls paid from retail education or real estate. And real estate category also spends a lot. Right. They advertise heavily for whenever there is a project launch for these guys, they spend heavy. Okay, okay. Interesting. What does your
00:20:26
Speaker
Help me understand the product journey. And I am assuming now you have two products. One is you would have a product for the leaders. And second is you would have a product for advertisers. So just help me understand. Not for advertisers, but for reporters. So like, I mean, Facebook, Google, all of them have a product for advertisers where advertisers can do self-service targeting. Yeah, but we don't have a self-service advertising platform yet. Okay.
00:20:56
Speaker
Okay, we have one is for consumers, wherein we disseminate content there. The second one is for reporters, wherein reporters can probably write their content, they can see their account balance, how much they have got
00:21:12
Speaker
how many articles got selected, published, how many got rejected, and why. We have built it very transparently, like how a Google AdSense publisher gets money and has been transparently reported, right? So we have built a similar kind of mechanism for reporters, and that's how our citizen journalism has become super successful. Okay, okay, amazing. And
00:21:36
Speaker
So what about the user product? How has that evolved over the years? You would have gone through a journey of investing in product to make it better. And so what were some of the milestones of the product for the consumers, the readers? Well, as I mentioned, it all evolved from one statement that I have read, or I came across in a news article, wherein I have read India is adding
00:22:06
Speaker
eight language audience out of 10 internet audience. That triggered and made me to further probe into the statement and I've done research and I realized there's a gap. You're saying that eight out of 10 new internet users are vernacular users. In 2015, when I thought of starting Way2News, I've read this article and I found that a lot of these audience are language consuming audience. They cannot consume English language. And I realized
00:22:37
Speaker
imagine if 12 crore English speaking audience has several thousands of destinations and look at the scale of a language audience. It's a billion language audience this country has and does not have a good quality platforms out there. And that
00:22:53
Speaker
new statement, what I have read in economic times has triggered me to start off or probably do some kind of a probe in the area. And I realized the news is the category where there's a gap. So one, I belong to a consumer tech category because my previous venture way to SMS is a consumer tech business. So I had a great learning slayer to consumer tech journey. And two, based on this, and then I've done some research and I thought like language is the area where I wanted to be.
00:23:23
Speaker
my journey should be in. And that's how I picked a way to use. And then initially, I mentioned, you know, in the beginning of the story, like I said, like, it's about relevant, it's about short, it's about reliability. There are three factors, right? So short is what the format that I have picked, and I have picked consciously after doing a pilot,
00:23:54
Speaker
I did a pilot for almost three months to understand whether the long-form content or the short-form content that works well with the new age digital audience. And before launching, and I know that I am going after two sets of audience. One set of audience that are traditional newspaper readers. They got accustomed to certain kind of an experience. One consistent and reliable experience that newspapers have offered. The second one is a new set of a mobile first audience.
00:24:24
Speaker
these does not understand newspaper all that and wanted to consume more. These guys are more multitasking guys. They consume multiple categories at a time. So I need a platform that can be appealing to both kind of audience.
00:24:43
Speaker
So that's why we picked short content because short content would be appealing to new age audience. At the same time, we wanted to offer the essence of the old newspapers. That is the credibility and trust that these guys offer. So we picked that property from newspapers and that's how Vietnamese has evolved as a short news in local language.
00:25:04
Speaker
So my journey started initially with short news in local language, providing that most important news. Instead of the traditional newspaper reader, when this guy turns to a digital, he will be having thousands of options to go through. And the audience are confused on what should be relied on or what can be trusted, what not.
00:25:34
Speaker
So because of this fragmented experience, people are not happy with the overall digital experience. So that's where we thought of providing that one experience in one platform by curating and giving the most important news of that day so that we can probably extend that old newspaper experience onto our platform. And at the same time, we thought of keeping the platform appealing to the new age digital audience by keeping it short.
00:26:04
Speaker
Define short, what does short mean? Short means you provide content in 60 words. We just give the essence of the news article and the user will get a sense of fulfillment after reading it. The difference between the long form content and short form content is, imagine you're publishing content on the mobile device and on mobile user does multitasking and his average attention span is not more than 10 seconds.
00:26:34
Speaker
and imagine you are watching an article and you got a WhatsApp message, obviously you will attend to WhatsApp message and you will leave here. If it is a short message, it is a short news, which can be consumed within 10 seconds, then you leave this and then you'll go happily to other part. I mean, to your WhatsApp message, if it is a long-form content, right? You'll always be having, imagine it takes two to three minutes to read that story and you've got a WhatsApp message that needs to be attempted.
00:27:02
Speaker
and you'll directly click on it. So you'll always be having that sense of fear of missing out something in the article. And at the same time, you have something to attend here. So this sense of fulfillment and the fear of missing out factors is something that we have played out really well on our platform by addressing this short content.
User Engagement Through Hyper-Local News
00:27:23
Speaker
Okay. And can people click for a longer version of it? Well, initially we started off with giving a long-form article, then we realized the user does not need a long-form article and user is presuming more video content. So we started giving it in the form of a video. Okay. It's like our screen is divided into half image and half text.
00:27:50
Speaker
and half image communicates what the story is all about, or it can be a video also. And then there will be a headline and then a small body with a 60 watt body, body of the story. Imagine if you wanted to learn more about the story, like you probably, we probably provide a video to go through it. And this would be like a human talking on the screen kind of a video.
00:28:17
Speaker
Uh, not always. It probably may be a real audio. Sometimes imagine like, no, something accident happened. Okay. We'll put visuals. Okay. So, okay. Footage of.
00:28:27
Speaker
Yeah, footage of it, all that. Okay, got it, got it. Okay, interesting. And you swipe up for the next story. Over the years, yeah. And the two things that we have addressed, okay, in the definition, the initial definition that I have given, one is short, and then the second one is reliability. These two factors are something that we have addressed initially. And then relevancy is what we have not addressed, because relevancy will be a little costly affair.
00:28:54
Speaker
during those days. So I thought of just giving it a try initially with the short news platform. And it worked out really well. I took some time. And then I started off the hyper-local content. We started building this single network, all that, in order to bring that relevancy. For instance, if I started giving local news based on your mandali or taluk that you belong to, you'll be able to relate the content.
00:29:20
Speaker
Today, a guy sitting in some village in Srikakolam district does not understand about the G20 summit and how it can impact his life. And he won't bother much. For us, it's an important news. I can provide that news to the user, but it probably may not relate. The guy who is sitting out there in a village cannot even feel the relevancy of that news.
00:29:45
Speaker
So for him, what matters is the local weather updates, local market-related news, maybe news around what's happening around him, news related to temple festivities, local festivities, all that. So that's where the second phase of waiting news is all about, and we have built it successfully. Initially, my team is a little skeptical about the citizen journalism, because globally, it has not become successful citizen journalism.
00:30:16
Speaker
Then we got in deep into it, why it has not become successful, all that. And then we made a transparent system to stringers. And stringers also have the problem with the newspapers. They work very hard, the stringers out there on the ground. They room around 10 to 12 hours throughout the day for sourcing the news. But they are not sure about the money that they get from the news publication, whether someone will publish it or not at the end of the day.
00:30:45
Speaker
So there we have built a solid system, transparent system. Even if something is not published on our platform, we'll give them a reason why is it not published. And then we also conduct so many training sessions for these serious reporters time to time. All that helped us to make creating that solid
00:31:06
Speaker
a presence in our villages today. And today we were able to publish 40 to 50% of our content is original and unique to us and is exclusive at least for several hours. It is all because of trusted reporters that are working for us on the ground. So that helped us to create such a large vast network. So that is our second phase. Overall, we started initially addressing two things. One is a short, the second one is
00:31:35
Speaker
Reliable content. The third one is, I mean, the relevancy we achieve through local content. Okay. So I want to understand two things here. The first is, how is the system aware of the location of the news and the location of the user? So location of user, you probably get from the device, right? The GPS. Well, yes, user has to give us the permission.
00:32:03
Speaker
Once we get the permission, we get the lat and long from the user. And our system maps his lat and long to a town that he belongs to. And then it maps back to his Mandal or Talu. And further, it maps back to a district. And we automatically start matching. Based on matching, we start publishing the relevant content to him.
00:32:25
Speaker
And how does the system map the stories that this story belongs to this location? We have additions like how a newspaper functions, right? Like a newspaper has a district edition. So the traditional newspapers have gone to the extent of a district. When it comes to a way to news, we have a district edition, we have a taluk or mandal edition, we have a town level edition. What is a taluk or mandal?
00:32:50
Speaker
Taluk or Mandal is a subset of a district. Generally, a district can have anyway. It's like a bunch of villages. Yeah, it's a bunch of villages. Okay. A bunch of villages will be treated as a Mandal or Taluk. And a Mandal can have a typically anywhere between one and a half lakh to two lakh of your population. And a set of Mandals will be called as a district generally. And the district generally will be having more than 10 to 20 lakh kind of a population.
00:33:20
Speaker
Okay. Understood. So you have a Mandal level news edition and a district level news edition. District level. And if it's a town with more than 50,000 population, we also do a town level content publishing to make it more relevant to the audience.
00:33:40
Speaker
So this year Stringer network, when they upload a story, then they will probably select at the time of uploading. Yes, they'll select at the time of uploading and AI also does the basic checks, whether the content belongs to that area or not. Over the years, we have got several millions of content pieces and AI has learned and has become far more accurate and robust. It is now able to discard even bad actors from the system.
00:34:07
Speaker
fake or motivated content will be eliminated. AI eliminates all such kind of a content and the manual activity has become very limited these days.
00:34:18
Speaker
Okay, interesting. So the second thing I wanted to ask you, these stringers are essentially relying on way to news for their earning. Like how much do they earn on an average? Or is way to news one of the places from where they earn and they earn from multiple sources? Both cases. There are people who are heavily relied on way to news to earn and there are people who contribute to multiple sources.
00:34:43
Speaker
But I would say someone who's giving you 20 stories a month, which are published, how much would you be earning, for example? I will put it this way. I have users who are earning anywhere between 20 to 30,000 per month. And I have users who are just earning anywhere between 1000 to 3000 as well. Those who are contributing to multiple sources, then they probably would be contributing a few stories to us.
00:35:08
Speaker
They probably earn anywhere between 2,000 to 3,000 a month. But those who are contributing heavily to us, and we also mark them as senior reporters or senior contributors, because they contribute a lot of trusted content, and their test score on the platform is very high. We have an AI that ranks every reporter. And based on that rank, also, the system accepts more content from reporters. And those trusted reporters will always enjoy
00:35:38
Speaker
more payouts from way to news than compared to relatively new reporters. Okay. So the payout per article will be higher for a reporter with a higher trust score. Right. And payout for article varies anywhere between from 10 rupees to almost 50 rupees based on the exclusivity, uniqueness of the content and based on the format as well.
00:36:00
Speaker
format like if it's video or just a text content we pay less okay if it's a video content we pay a little more but in exclusive video we pay a little higher than that and these are decisions made by AI or humans who decides how much to be AI does all that based on the submission okay okay
Advertising Approach and Market Expansion Plans
00:36:20
Speaker
Interesting. Very interesting. So I understand now how you built the delivery of news and how people are getting served personalized news. How do you, I mean, without an advertiser sales service product, then how do you run advertising? Like how does that happen? Well, I have a sales team out there in the market.
00:36:44
Speaker
And they reach out to every advertiser. And whenever there is a campaign, we get creatives. And I have an AdOps team here. The team does everything. We have an internal product built, but that has not been given away to advertisers. I have an internal product that runs like a self-serve product, but has not been given away to advertisers yet.
00:37:06
Speaker
So the thing about digital advertising is, and digital advertising is called performance marketing because you can track performance. You can see like a dashboard which says that your ad was viewed by a thousand people out of which 800 people clicked through. And then you can also track how many of those 800 people actually signed up and how many made a purchase or whatever it is. So that whole performance marketing funnel that currently you're not offering to advertisers.
00:37:36
Speaker
Well, we fundamentally, we don't belong to that category. As I mentioned, we are a natural exchange extension to a newspaper. As the newspaper subscriptions are going down and people are the same set of our people are migrating to wait and use. And advertising, traditional advertising advertising on newspapers, they are migrating to wait and use.
00:38:00
Speaker
One, we are not a performance advertising platform. We are an impact platform. That's why I said right, scale is important for us in every market, in every community. As I mentioned today, I think we are enjoying the highest scale in Telugu market and is growing in Tamil Nadu as well. We are reaching masses. We have all kinds of audience. Be it anger audience or like a elderly audience.
00:38:28
Speaker
Any kind of a campaign will work, but specifically for an impact or a brand building exercise and not for performance. That is very clear. The message to the advertiser is very clear and is proven also. I'm not just saying that, okay, this is how we want it to position it, but the platform has proven already that it can deliver better results like newspapers.
00:38:53
Speaker
Okay. Interesting. Yeah. Newspaper doesn't, is not a performance marketing channel. You don't get analytics and you can't track who saw the ad and who purchase it. But we heard from our advertisers that Red News is driving better footfalls than traditional newspapers. Ah, okay. Okay.
00:39:11
Speaker
For a small advertiser, it's relatively easy because they would stop newspaper advertising, they would advertise here. So they would immediately see the results. So it is not very difficult for them to judge efficacy. Yeah. And these, these advertisers, local advertisers, very smart guys, they measure the media very well. Traditionally, they have been advertising on multiple platforms, right? They know how to measure.
00:39:34
Speaker
It's like one day they will give an advertisement on a newspaper, the second day they will give an advertisement on written news, and they'll see the fake footballs, and they'll compare the medias, and they'll look at the ROI against each media. And today 99% of my revenue is totally a brand advertising, I mean the impact advertising revenue.
00:39:57
Speaker
Okay, interesting. I mean, how do you price the advertising? You would probably price it per thousand views or what is the way in which that happens? Like, you know, I understand how Facebook, Google pricing is like, you know, that CPM cost per thousand views is how they price their advertising. How does it happen in your case? Well, again, the same case. We don't act like a digital platform rather we act like a natural extension to the newspaper.
00:40:28
Speaker
We don't sell in CPM or CPC or CPL. We do a cost per day, like how a newspaper sells their advertising space, like they sell in a cost per day.
00:40:41
Speaker
It's the same way we also sell the cost per day. Our first a few properties, especially the position one to position five, like newspaper cells, like a page one, page two, page three, and a sports page. There will be a back page. There will be a middle page, right? The same way we sell up position-wise, our first five positions are heavily for brand building, creating that impact among the consumers. And we charge a premium.
00:41:07
Speaker
in first five positions and then there is a cost associated with every position. Generally cost goes down a little, it starts at some X amount and then as the position goes a little deeper, the cost also goes down. Okay, so these positions essentially are like a bundle of like if someone is buying in the first five positions then they will get, it will be shown to more number of people, something like that. Very true, the reach will be higher generally in first five positions.
00:41:36
Speaker
And it's like newspaper like your first page or a jacket ad will be very high and your page three or page six will be charged less, right? It's the same way. I'm wondering why you don't make this more transparent and I mean, it should
00:41:55
Speaker
be relatively quick for you to build out a product which makes this more trans. You could just directly say that instead of saying... It is a matter of time, Akshay. There's nothing that is stopping us today to launch a self-serve platform, but it requires a massive scale. Right now we are available in two languages, dominantly. Once we are probably become maybe in two, three languages more, then I may extend that service as well.
00:42:24
Speaker
But I'm also talking in terms of transparency in pricing. You right now are not giving them a transparent CPM number. Like, you know, you are paying 50,000 to me and for that 50,000 it will be viewed by X number of users. That is not the case. You're saying that for position one, you will pay 50,000 to me. But what is the meaning of position one?
00:42:48
Speaker
Position one, we will say that it gets around X number of impressions and it reaches to Y number of audience. Okay. So you share those numbers. We tell them up front. Yeah. Okay. Okay. Got it. Got it. Okay. And the cost per day will be arrived based on the reach and number of impressions it can deliver. And the advertiser will tell you what geography he wants to advertise in.
00:43:15
Speaker
Well yeah, advertisers can target based on the language, one based on the location. He can target multiple locations as well. And probably age and all? Or is it just language and location? Right now we're not offering age-based or gender-based targeting at this moment. We're going to play in advertising, newspaper type of advertising, and that's working out really well for us.
00:43:40
Speaker
Okay okay okay understood okay and what kind of like this year how much revenue will you do like last year we did around 34 35 and we are operating at 40 to 50 percent more revenue this year.
00:43:57
Speaker
And trust me, we have not started off monetization earlier. We just started off this year actively. I don't have a sales team earlier built across various geographies. Now sales teams are on the ground and hopefully next year would be really great, a great year for us, especially for revenue.
00:44:17
Speaker
Next day you'll probably cross 100. Yeah. The focus for this year is more on our user growth in other languages. Intent is to prove that this model can be replicated across languages and we have done it successfully in Tamil Nadu. Another challenge what we have is that we can go deeper in the language that in Telugu language, especially, and we have proven that as well.
00:44:40
Speaker
So now the focus is on revenue. This year, we started building all the systems later. And then we started building the product also, but also getting ready for monetization. And earlier, the product is not ready for monetization. That's what you asked for, age-based targeting. All this is not ready yet. So we are getting now ready for better targeting, giving better. Demographics. Yeah, exactly. Better demographics to advertisers, all that. OK. Interesting. Interesting. OK. Got it.
00:45:09
Speaker
Why did you call it? I wanted to add this point as well. In a mature language like in Telugu, we have become profitable already. Okay. So profitability is calculated as cost you're paying to the stringers. That is the cost. Everything. We generate our P&L based on language and our Telugu P&L today is operating in a positive margin.
Raju Vanapala's Entrepreneurial Journey
00:45:35
Speaker
So your tech team costs you're over everything is allocated and we have given the loaded costs to other languages. Okay. Amazing. Amazing. Okay. Essentially at about 5 million users in a language that language gets profited.
00:45:53
Speaker
need not be 5 million because earlier we are heavily dependent on one language, so the load will be very high. But imagine if you have 3-4 more languages, the cost of multiple languages will come down, right? Your fixed cost is distributed over more years. Very true. The fixed cost gets distributed across languages.
00:46:13
Speaker
Right. I think, you know, maybe at a 15 to 20 lap DAO level, once you start the monetization, it will become profitable. Wow. Amazing. Amazing. Okay. Okay. Interesting. What's the roadmap? What are languages in what sequence are you planning to launch? Well, South first is the strategy that we have taken initially and we have covered two major languages in South and next would be Canada and Malayalam.
00:46:38
Speaker
And then we wanted to add Marathi as well. I also wanted to do one pilot in one Hindi language just to check, okay, whether this can be replicated in Hindi territory or not. And we probably know just to have understanding over the Hindi market as well. So which state will you do for Hindi? Not decided yet. Yeah, but Haryana is in consideration. Okay. Right.
00:47:07
Speaker
Yeah, we wanted to pick a small state initially. Right, right. I think one of the hill states could be interesting like Himachal or Uttarakhand because there are definitely more news dark states. And weather is very important there, you know, in the hills. So that local awareness of weather, weather events, weather catastrophes, so many such things which happen in hills. So yeah, okay.
00:47:35
Speaker
So why did you call this way to news? Because this name is not vernacular language friendly. It's an English name. Why not choose a universally understood name?
00:47:50
Speaker
I just continued my old, the way to thing. I did way to SMS earlier. So it becomes a natural extension to that is what I thought initially. So I said, the domain is also available to me during those days. And as a way to use is a catchy, generally is a catchy word. Okay. Easy to recollect. Although like, yeah, the news, the way to use is little English in nature. Oh, you know, especially for a local language platform.
00:48:20
Speaker
But no, I can definitely take the advantage of my way to SMS is what I thought initially. Tell me the whole journey. You started your entrepreneurship. How old were you when you started? And so way to SMS was the first business you started? Well, yes, I started off my entrepreneurial journey in the year 2004. And how old were you? Soon after finishing my MCA, I came to Hyderabad to start up a company.
00:48:49
Speaker
Inspired by my brother, my brother is into trading. He used to do a lot of stock trading and he used to talk more about CEOs, founders, their journey, their achievements, all that. And all that inspired, inspired me. And then I started dreaming about creating company. Where are you from? I'm from Vijayavada, which is a capital of Andhra Pradesh, AP, separated AP state.
00:49:19
Speaker
And what did your parents do? What was their background? Well, I come from an agri-dependent family background. I was born in coastal Andhra Pradesh and all my schooling happened in Telangana region in a government school and then migrated back to Andhra to Vijayavada. My high schooling and college happened in Vijayavada. I did my MC in Neera to Vijayavada.
00:49:46
Speaker
My life, it all changed in my first year of MCA, where I have decided to start off my company. And then I started tracking all the companies, their business models, all that from the first year of MCA. And soon after finishing my MCA in the college, I came to Hyderabad to start a company. I did a lot of research on various business models. I used to visit Smart Industries Service Institute every day to read business models. They used to publish, they used to have a long library of business models.
00:50:17
Speaker
I've read so many business models and the investment required, the kind of effort that is required, resources requirement, all that. And I realized that I don't have any money, by the way. Okay. I have a strong intent of starting up, but I don't have money during those days.
00:50:32
Speaker
I realized after all studying all the business models, I realized, okay, I'm not fit for starting any of those models. So I went back and I started thinking about internet-based business models, where knowledge-based business models are something that I started thinking. And that's when I started SMS Alerts business initially in 2004.
00:50:54
Speaker
I used to provide bulk messaging service to enterprises, especially real estate companies, stock market guys, they used to send tips and alerts to their clients, all that.
00:51:10
Speaker
And marketing became a lifestyle business after one, one and a half year, it became profitable. It became a lifestyle business, but that has not given you. How does that business run? You buy messages in bulk? We buy messages in bulk from a telecom operator and then we sell it. We buy at X amount and we sell at X plus Y amount to these guys. What is that value addition that you are doing?
00:51:40
Speaker
Nothing much in it. So, so that's what okay, after one, one and a half year, I understood that I started making decent money on it. And that's when I started again, my consumer tech, I started thinking about innovative business models. And that's where I started off a venture called V2 SMS. And V2 SMS used to provide free SMS service to consumers. SMS used to cost almost a rupee during those days.
00:52:06
Speaker
And it all triggered from emailing. Yahoo Mail used to be the popular email service during those days, which used to be a free service. And I thought, when emailing is free, why not I can make SMS be free? And that statement triggered me to start B2SMS. And I have seen all sorts of troubles being a startup or a bootstrap startup company.
00:52:28
Speaker
Tell me that whole roller coaster Janina, I'd love to hear that. Like what went wrong? What did you learn? I would love to hear that. Superb learnings, you know, it's very, very tough to create a consumer tech business with your own money always. It's like you always think something and something else will happen. Initially, I thought I have started off this free SMS service. I thought I can embed an advertisement in every outgoing SMS.
00:52:58
Speaker
And I thought I can make revenue through this embedded ads. On one hand, users started consuming SMSs heavily, but the monetization is a challenge. It's very nascent, right? Mobile advertising is very beginning during those days. I'm talking about 2004, 2005 days. Nobody understands mobile advertising users those days. Nobody wanted to invest on it.
00:53:25
Speaker
And then I have to change my business model. I have to control SMS usage on the platform. And then after some time I have moved from SMS advertising to online advertisement. I started monetizing through Google ads. Okay. So people who would come to the platform to send SMSs, they would see ads being displayed while they're composing their message and sending their message. Very true. Very true. Okay.
00:53:52
Speaker
Okay. Okay. And, and, uh, and I used to create a split experience instead of one page. Okay. I used to think, how do I make the user to navigate three or four pages so that I can make more money and all these learnings, you know, you, you can't understand, uh, you know, even breaking the experience to multiple pages. You'll, you'll, you'll get this idea after like going through a certain pain. It won't happen immediately.
00:54:19
Speaker
Right, because you were making losses, so you needed revenue, so the only way to increase revenue is Chopra. I used to pay around $27,500 on every outgoing SMS, and I used to make nothing. And it continued for almost three years from 2016. How much were you losing each year?
00:54:37
Speaker
I think 33.5 crore is what I have spent during those days. My own personal money secured through friends, family members, through various sources.
00:54:51
Speaker
So, and God knows, okay, without, on one side, on a consumer side, it's going to the masses, it's growing. I'm getting 15 to 20,000 users every day. You probably didn't need to spend anything on user acquisition. The user acquisition would be organic. User acquisition is zero. My user acquisition cost is absolutely zero. Because the product is so powerful. The product, it's a communication product. It has virality inbuilt.
00:55:16
Speaker
So that's where I have learned so many things. There is that immediate cost saving and the Indian mentality is to look for cost saving. So I've also used the way to SMS back then. And the moment you receive one message, okay, maybe you'll see, okay, what is this? And second, third, fourth, okay, you'll start thinking about what is this? Why is it, you know, he's sending through this platform? You'll also come and you'll also register, right? So that's how I used to get. Okay.
00:55:45
Speaker
It's a network effect, purely like a network effect. And at one point of time, I started adding more than a million new users every month and it's totally organic.
00:55:56
Speaker
So this is where I have learned managing the tech resources, handling the tech resources, hiring the right talent to handle that scale. Imagine if it is a Diwali day or if it is a January 31st or December 31st, you get millions of messages. Millions of people will land on your platform. They send millions of messages. Everybody sends around 200 to 500 messages to their contact group.
00:56:22
Speaker
We get a course of messages in just one hour time, handing that scale, all these things I have learned during my various days.
00:56:32
Speaker
At the same time, I have seen the financial challenges, managing those financial challenges. At some point of time, I have actually controlled the growth also, in order to manage my month-on-month financials. How did you control it? Instead of buying my sales, you can say... I used to stop surveys during the night hours to a non-metro audience.
00:56:54
Speaker
to survive the business. And I used to have a good set of advertisers for metro, but for non-metro, I don't have any advertiser during those days. So obviously, so I had to stop their activity a little and I have to encourage more metro activity because I have advertisers to fund their activity.
00:57:20
Speaker
Okay. Okay. Interesting. What was the peak number of users? It's like, you know, you're fighting with the growth there. Okay. And, you know, growth versus survival. So survival has taken a, you know, a front seat than growth at some point of time. And then in 2009 or 2010, we became profitable. And then thereafter, I never looked back. I mean, what was your peak user base?
00:57:49
Speaker
Over on BTS mess has a 60 million registered audience and it used to 60 million registered audience. What we got around 35 to 40 lakh kind of daily audience during our peak days. And that's a very big number during those days. I'm talking about 2017. Trust me at one point of time, I have a,
00:58:17
Speaker
what one maybe one fourth of the entire internet basis with way to way to SMS during those days. Wow. And we are a most popular keyword in Google for almost three consecutive years. Wow. Global biggies like Yahoo's and Google's during those days. Wow. How is it that you didn't have VCs to take money? Okay.
00:58:45
Speaker
There's not much VC ecosystem during those days. VC ecosystem evolved after 2009. Right. Before that, not much activity from a VC side. And by the time I started thinking about raising funds, the business became irrelevant. And because of WhatsApp, not because of WhatsApp, because of dry regulations, we became irrelevant. The cost of SMS has gone up. Okay.
00:59:16
Speaker
And some of those technical policies that they have laid down has killed our interactivity in usage. So overall, it became irrelevant. And then WhatsApp thing happened at a macro level. People started migrating heavily to WhatsApp than using SMS.
00:59:35
Speaker
So overall, I have a great learning experience from my way to SMS days in building and scaling a consumer tech company and understanding the consumer behavior, especially on consumer platforms. So always my first choice will be building another. That's my way to use also. One thing is very clear. I wanted to be in a consumer tech. And then I have arrived at a language as an option.
01:00:04
Speaker
The thought thing is the short news thing happened. But my first choice of pick always will be a consumer tech company.
Transition from Way2SMS to Way2News
01:00:15
Speaker
You must have gone through like that, you know, that valley of despair where high to SMS was becoming irrelevant and you didn't have something else to replace it with. When did you start going down where, you know, you realize that? It all started from 2011 onwards and then it... And what was your revenue? It took almost two years. We were doing around 35 CR annually during those days. Wow.
01:00:45
Speaker
Okay. Amazing. Yeah. Around three CR per month on a way to SMS. And my cost used to be not more than two and a half CR. Okay. We are a profitable company with more than 80 odd people working for a way to SMS during those days. But the problem is that I can get the natural extension for way to SMS would be getting into chat.
01:01:13
Speaker
But by the time the big guys have entered into the category, like hype is a well-funded company, started doing some good work and then WeChat entered into the country and then WhatsApp, everybody knows, has got a great traction in India straight away.
01:01:34
Speaker
Yeah, there was a whole bunch of... Yeah, yeah, yeah. All of them, there was a Korean company also. Yeah, Korean company has started off. I realized I cannot fight, you know, with these big geese, with my bootstrap resources, all that. So then I started thinking completely differently. I started thinking more about creating something else, something very big. It's always pain, right? You know, whatever you have built passionately over the years, and you want to take it globally.
01:02:03
Speaker
later on, but you know, it's going down the drain. And one day because of regulations, it's always be painting, it's not that easy to know, accept such kind of a change. And I used to work almost what 12, 13 hours every day. Keep the service alive. And you know, we always work for competing with mobile to mobile SMS.
01:02:29
Speaker
during those days. So, you know, great experience overall in benchmarking certain things and, you know, achieving the process of achieving it, aligning all the people towards one objective. I want to understand how you manage the downfall when your revenue started falling. Did you have to lay off people? How did you survive from 2011 to 2014-15? That's where, you know,
01:02:55
Speaker
I realized the strength is okay, not a consumer destination, but I realized the data that I'm sitting on. So I started doing a lot of marketing on the data. 60 million signups. Yeah, 60 million signups. And I have a highly profile audience during those days and we started doing a lot of marketing work on it. How did you have profiles of them beyond this email idea?
01:03:20
Speaker
Like you would ask them some information. We used to take his gender, the location, and age, all that. Okay. Okay. So we used to do a lot of lead generation activity, performance marketing activity on this database. And eventually, email campaigns, SMS campaigns, and then we used to apply these consumer insights and used to target users on Google and Facebook as well. Okay. Okay.
01:03:48
Speaker
So this was like a services business? Yeah, it's a services business that continued for several years in order not to disrupt the activity. And I have not laid off any team during those days. But my mind is always a consumer tech brain. Okay. I just survived through a services business for a while, but I always used to think about consumer tech and that's how our business has evolved later on.
01:04:16
Speaker
And did you raise funding after that? Was the way to news idea came? Not immediately. I continued by bootstrapping, applied all my consumer tech learnings on way to news. And again, like, you know, I have built it organically from zero to almost 25,000,000 hours.
01:04:34
Speaker
2.5 million doses where at that scale, I realized it's important for me to get some support from VC so that I can scale it faster. And I realized the platform has the capability to conquer the country. I mean, we can create a larger platform down the line because my numbers are category defining numbers. I mean, category dominating numbers in terms of retention, the organic acquisition,
01:05:02
Speaker
all numbers are looking very healthy compared to the competition. So that's where I felt it's important to have a venture backing so that I can build it faster. And I have raised money in the year 2022 from Best Bitch Capital. I have raised around 130 crores, close to 130 crores. And from then we are going very fast. We have almost, as I said right, like in last one year, okay, that 25 became 65. We added almost 4 million DAOs.
01:05:32
Speaker
in just one 14 months time. And we have proven multiple things that we can go deeper in a mature language like Telugu. And the same business model can be replicated in Tamil or any other language is what we have proven.
Role of AI and Future Prospects for Way2News
01:05:47
Speaker
Amazing, amazing. Initial journey of building 25 lakh DAOs was because of that database you had of 60 million that helped you or you also? No, not at all. It is purely through the content, capitalizing on the events, all that. Again, the product has helped me.
01:06:08
Speaker
I mean, imagine if there is an election result, right? Okay. And we celebrate such kind of events and we ensure that users will be sharing a lot of content on those days and we get more users. And every news event we celebrate heavily and see we have our own set of learnings when it comes to doing this news.
01:06:30
Speaker
And you're creating that engagement over the application and making one user to act like a brand ambassador to your product. All my way to SMS learnings, I have applied again on the news and it worked really well for us. How do you make users become brand ambassadors? How do you make them share? That's what, first, okay. I have not chosen the long, I mean, the longer version. What we have done is longer version, the share rate is lower compared to the short version.
01:07:00
Speaker
And we started publishing one news item in one screen. That means it becomes easier. You can take the screenshot and you can share. Even if you take the screenshot, you will see a logo on top of it. We started publishing logo on top of every news article that goes out. And that acts like a hook for us to get one more user.
01:07:23
Speaker
Right. This is how insurance also grew. I remember there was a time when I used to receive a lot of these screenshots of insurance. Right. And generally people are curious about after receiving multiple times, they'll go and search for what it is. Right. So you said that you celebrate like an election and that gives you more sharing. What do you mean by celebrate here?
01:07:50
Speaker
It's like, you know, we do a lot of shareable content compared to others. They just, you know, do the content, something that happened, but we encourage the users to share and we know how to create content that can be shared. So it will be like content with an opinion, like, like,
01:08:07
Speaker
you know, like our guy won. And so that will support those days. Yeah, the followers of that leader will share all his cards on that day. And we got so many users. Each story is a card, basically. Very true card. And imagine, you know, imagine like there is a 10th class result date announced by the government. We publish it in such a way that okay, every parent shares. Imagine like, you know, I have a parent, I have a
01:08:37
Speaker
My kid is like 10th class studying, right? Imagine I have seen this article, I will share to another parent. So we have built this sharing effect or like a network effect on the platform again. And we identified these events that can probably bring users to us.
01:09:01
Speaker
make this into a scalable process. It's a very human subjective judgment that this headline will be more shareable than that headline. You know, how do you scale that up? Very transparently, internally here, every writer here is equipped and understands what works and what not, and what kind of a content can get X number of shares or Y number of shares. How do they learn that? How do they get that? How do you equip them?
01:09:31
Speaker
We have a comprehensive CMS built and that CMS gives all the insights and analytics about engagement of the audience on various articles and content. It provides a number of people who read that news article and how many number of minutes that article has been consumed and how many people have
01:09:50
Speaker
probably flipped off without reading, and it will give the engagement percentage on the article. It will also provide how many people have liked the article or disliked the article, how many people have commented on the article, how many people have shared it on their WhatsApp groups or Facebook or on Twitter. It will give all the analytics. So most of these writers are equipped to understand what works and what not.
01:10:18
Speaker
Okay. That is for the writers. How about your in-house team of editors? How do you train them to, so that they learn that judgment? What does that is for our internal editors? Okay. Yeah, this is for internal editors. Okay. Understood. So they are continuously receiving feedback on what was getting published, which helps them learn. Right. I'll give an example. The recent election results happened on December 3rd. We got around two and a half lakh users installed on that day and is
01:10:48
Speaker
everything happened organically. Imagine the scale two and a half lakh user installs in one day. That is the organic scale that we achieved on that day. Amazing. Amazing. I'm seriously impressed by the quality of your editorial team, you know, for them to be like you have, you said you have about 200 editors, right? 100 plus editors working on the content.
01:11:14
Speaker
Amazing, amazing. So to build a team of 100 people who are so good at creating viral content, like how did you do that?
01:11:25
Speaker
All this through, although we have a wide range of languages here, we all talk about only one language called number language here. It's purely because of the number language that we internally discuss here. And most of these guys, they don't understand the difference between an average and a percentage. But they know that 100 shares are more than 80 shares, the plain number that we provide. So they know that, OK, this 100 users have seen this particular article. And then what's this?
01:11:54
Speaker
20 shares happened, the percentage we provide the percentage, how many people have shared out of how many people have read, read versus share percentage. And you know, what kind of articles, what kind of a category is selling more. We understand all this. We provide all these analytics to them. And how do you hire such people? What do you look for when you're hiring an editor?
01:12:16
Speaker
We look for relatively three to five years of experience, guys. Experience is not much of a package. And then we train them here internally. What experience should they have? Probably worked in a previous media background, maybe from a print background or from a television media background. Okay. So guys from Sabaita's background, all that.
01:12:45
Speaker
Okay. Okay. Okay. Understood. I'm just wondering, eventually, do you think AI will replace this team? Because a lot of these things now, a chat GPT can probably do it faster, cheaper. Well, at least for the next several years, what I feel is that AI can help them to make better content. It will equip them better. It will make their work faster.
01:13:10
Speaker
But eventually, yes, AI will definitely take all their work, and they will be having what's writing a lot of content down the line. Yeah, that's true, that's true. Because AI can do very instantaneous A-B testing. I mean, you can just show that article to five people with one headline and five people with a second headline. Agreed. You know, that itself will tell you which headline is there. So that kind of instant A.
01:13:34
Speaker
The large language models have not been evolved yet. It takes at least another 18 to 24 months is what I believe. And later on, yeah, I think Watts will take over this job totally. So is there a challenge right now, the fact that this is vernacular language? Like most like chat GPT is like basically good at English content, right? It may not be that great at Telugu content. Is that the obstacle?
01:13:59
Speaker
And the quality of summaries also, I have, we have applied chat GPT internally here. Okay. The quality is not that great yet. It still works on Telugu, but not that great yet. And we tried even with English as well. Okay. Still, I believe it is a substandard. It is generating the substandard content in terms of sometimes we are losing the essence of the article, but it will evolve over the time.
01:14:26
Speaker
Yes, surely. How do you use AI then? What do you use for AI? AI for a language would not be so easily available of the self. How did you start incorporating AI in the product? Well, right now we are using AI in combating with fake news.
01:14:46
Speaker
discarding the wrong articles or like identifying culprits in the system, the bad actors in the system who wanted to cure system. Like we are trying to filter out a lot of repeated content, a lot of copied content. Most of these stringers, they copy content from other media and they publish to us. They want to make big bucks to us.
01:15:08
Speaker
So, AI does these basic checks and balances. And which AI model are you using? Which platform are you using? Or is this like in-house? It is totally built in-house. We are not using any third party. I have a great AI team internally here. They do all the great work.
01:15:28
Speaker
Okay. Okay. So you would probably be among the best in the country in terms of like say a Telugu language AI tool. Maybe in language. Yes. Okay. We would be the best, okay. In adopting the AI and probably maybe delivering our learnings to the outside world, maybe in the next several years. And you would also be generating a large amount of training data, right? I mean, a lot of data. In fact, tomorrow we wanted to get a wise to text.
01:15:57
Speaker
because we wanted to encourage village level reporting. We wanted to publish stories to your village level as well. Imagine I have 40 users from that village. I wanted to show that news only to that 40 users. And a lot of these village level guys are not equipped in writing or framing your short summary as well. So I want the user to do a voice message to us on the unused news. And AI generates a beautiful summary out of it.
01:16:24
Speaker
And this is how I want to utilize AI down the line. Amazing. So you would actually have a second product, I think, which you can monetize, which is your vernacular language AI. Because you have so much training data.
01:16:39
Speaker
Load of data. AI would learn what works because your AI will also see that this got shared, this got read, this was ignored. So your AI is also learning what is the right way to frame content, what is the right kind of summary to do. My manual work has become very, very limited these days. Almost we get around 50,000 articles every day and just like 18 to 20 people are handling the remaining set of
01:17:08
Speaker
7,000 articles that are getting published. It is purely because of AI and over the years it became very, very accurate because of the large language set that we have built over the years. I would say I think we have more than 3 million published articles and then over 7 million odd rejected articles and AI learns a lot from rejected articles as well.
01:17:33
Speaker
Yeah, absolutely. Absolutely. So do you look at that as another product you will sell? Like a monetizable asset? No, at least immediately. But that would be an opportunity down the line to monetize and to release our learnings to the outside world. We really wanted to outsource these in the form of a SaaS products down the line because we are in an unique position to provide these in the country.
01:17:59
Speaker
Yeah, absolutely. Right now our focus is on our core business, adding more markets, getting deeper into the markets. That's what is our core focus at this moment. Okay. Do you intend to raise more funds? Do you need more money for opening more markets? Well, I still have a good decent amount of money in my bank account. I don't think I'll raise anytime soon.
01:18:26
Speaker
But yeah, obviously I will, I will hit the market at the right time to go and raise. And I, as I mentioned, right. Okay. My next target would be adding Karnataka and Marayalam markets and Maharashtra market. Probably I may need money for these markets. I don't have any timeline in the mind right now, but probably may hit the market anytime after six months.
01:18:58
Speaker
Karnataka and Maharashtra are like wealthier states. So probably your monetization would also be better there. Obviously Karnataka is yes. Okay. And even Malayalam has a very good educated audience base and a lot of advertisers want to reach out to Malayalam audience as well. So both are healthier markets and Maharashtra is a bigger market again.
01:19:23
Speaker
Overall, yeah. So every market has their own significance. And advertisers wanted to target these markets with highest per capita always. So Karnataka enjoys highest per capita. Malayalam even Kerala enjoys even highest per capita. Maharashtra enjoys highest per capita. So even our strategy is also based on per capita. Our strategy of entering into markets will resonate with the per capita of that state.
01:19:54
Speaker
Right, right, right. Okay. Understood. Okay. Amazing. So, you know, my last question to you, what's your advice to young aspiring founders, people who are listening to the show, thinking of building their own business? Well, go after your dream. Go start, see your own share of troubles and you will eventually evolve. You'll become better and you'll conquer.
01:20:20
Speaker
whatever the ambition that you are trying to chase. And that brings us to the end of this conversation. I want to ask you for a favor now. Did you like listening to the show? I'd love to hear your feedback about it. Do you have your own startup ideas? I'd love to hear them. Do you have questions for any of the guests that you heard about in the show? I'd love to get your questions and pass them on to the guests. Write to me at ad at the podium dot in. That's ad at T-H-E-P-O-D-I-U-M dot in.