EP #8: How Hungama Drives User Growth with AI-led Content Recommendations

Pradyut Hande: Hi guys, welcome to a very special episode of Mastering Martech, a specially curated video series powered by Netcore Smart Tech, where we speak to leading growth practitioners who are working at the cutting edge of marketing technology. I’m your host, Pradyut Hande, Senior growth marketer and Product evangelist, and today I have with me a very very special guest, Mr. Siddhartha Roy, COO of Hungama. Thank you so much for being part of this video. 

Siddhartha Roy: Thank you for having me.

Pradyut Hande: Absolutely, it’s fabulous. So, you know, further ado, I would just like to first tell our viewers what this entire episode is all about. Now, the media-OTT space in India has witnessed a huge boom over the last few years itself; in fact, an EY report says that it’s growing at a CAGR of almost 22 to 23 percent, and set in this backdrop, how is Hungama placed, what is its genesis and growth story over the last few years? If you could just walk us through that.

Siddhartha Roy: Sure. With regards to the overall- the overall opportunity at large, what we’ve done over the last 4-5 years is- now largely transition itself to a complete consumer-facing business.

Pradyut Hande: Right.

Siddhartha Roy: Right, with our music business which is largely into the streaming space, there’s Hungama Music, and our video play, which is driven by Hungama play, we have a large focus across these two categories-

Pradyut Hande: Right.

Siddhartha Roy: And you have seen obviously exponential growth over the last 2.5-3 years, on the back of the larger change in the ecosystem that was driven largely by Device, and Data.

Pradyut Hande: Correct. Absolutely.

Siddhartha Roy: There was-there was very clearly platforms and services and destinations such as ours, which already had content-

Pradyut Hande: Absolutely, Correct

Siddhartha Roy: Users got an opportunity on the back of the change in the larger ecosystem, in massive consumption boom, catering to our services

Pradyut Hande: Right.

Siddhartha Roy: We’ve had exponential growth one per month, year on year-

Pradyut Hande: Right.

Siddhartha Roy: And these have doubled, trebled, quadrupled on the back of great consumer engagement.

Pradyut Hande: Okay, Okay.

Siddhartha Roy: If you look at 3 years back – if I was to look at basic data which was – which was published data- that time India was consuming roughly about 700 MB-per month of-of data. In this dataset which has come in, literally last week we are talking roughly about between 12 to 13 Gigs (Gigawatts) of data used.

Pradyut Hande: Right.

Siddhartha Roy: And that itself is roughly 15X growth.

Pradyut Hande: Correct.

Siddhartha Roy: Over the last 2.5-3 years. And I believe that as entertainment entities which are on this vertical – big, single biggest are beneficiaries of that.

Pradyut Hande: Absolutely.

Siddhartha Roy: So what does this consumer do? Are these consumers which are currently between our music and video property of roughly 65 million active users on a monthly basis – is-are consumers coming in, doing deeper and deeper consumption-

Pradyut Hande: Right.

Siddhartha Roy: -per day. For us, we’ve – we look at India, not only as India which is Bollywood driven, India which is linked to the larger regional content play.

Pradyut Hande: Correct.

Siddhartha Roy: We are, as a brand, very very conscious about the middle India content play and the fact that today, we service the Tamil users, the Telugu users, the Kannada user, the Bengali user-

Pradyut Hande: Correct.

Siddhartha Roy: – in the same level as I am-as I would go about serving Hindi-speaking regional users.

Pradyut Hande: Correct. So in a scenario like that, is it fair to say that, you know, Hungama is one of the only few players in the market that caters to the English, the Hindi category, and the vernacular languages at the same level as you mentioned; customer experience is so important, and the fact that each customer across socio-economic strata and various language backgrounds are catered the same?

Siddhartha Roy: I-I singularly believe that Hungama’s focus on the region has been there as a company over the last 10-12 years. This started when we still had network-based services, when- I still remember my founders said that when India crossed the 600 million consumer-based connectivity, in terms of global, that resulted in itself going to a set of audiences which never knew what reached.

Pradyut Hande: Right.

Siddhartha Roy: And the fact that regional content, which is already getting to consumers in big ways, we just kept about more in-depth. On Hungama Play, we started investing in regional, original space-

Pradyut Hande: Right.

Siddhartha Roy: -Quite early on. So- to just give you a perspective on this, last one year itself, we’ve done four originals, in the Marathi spaces, we’ve currently got projects that we are working on in Tamil, in Telugu, in Kannada, and we want to take local stories, regional stories out there. And we are able to bring these audiences closer to the kind of entertainment they want to look(?)at.

Pradyut Hande: Right.

Siddhartha Roy: On the music side, we’ve been extensively providing to the regional space for the last 8-10 years and in the streaming space. Outside of Hindi, the regional languages put together contributes roughly 50+ percent of consumption, which means that this is a very important block –

Pradyut Hande: Correct.

Siddhartha Roy: -For services such as us, to be able to identify and go after. So there is a very very clear core strategy in place, and with today’s services, three languages across India – the total languages in Hungama’s services is roughly 17 languages. 

Pradyut Hande: Right. And I think I did a minor glance at the map itself and the fact that here, the user who is launching the app for the first time has to put in three preferences-

Siddhartha Roy: Absolutely.

Pradyut Hande: -and it’s based on that, then you can kick start the entire personalization journey. So, in a scenario like that, I just wanted to pick your brains on what the key challenges would be for user engagement, or conversion point of view, especially in the OTT segment where different players are grappling with different challenges; it would be viewer increasing, an average session like, it would be converting freemium listeners or viewers into paid subscribers – so what would be some of the key challenges that Hungama grapples with, especially OTT?

Siddhartha Roy: So let’s first talk about the fact that – so there are two sets of the question, so I think it is a very apt question to ask. One is the consumption journey, the second is going from a free to paid journey.

Pradyut Hande: Right.

Siddhartha Roy: and the consumption journey itself comes with a whole host of vectors in play, with your on-board users; so apart from you going out and spending money, to be able to acquire the user, once you have acquired the user, you need to go through the entire acquisition funnel, where you move the user from an acquired user to an active user, to an active consuming user, and then to building out a larger retention funnel with him, where over a TV, in free, or play comes into play. Okay? So with regards to challenges, okay, you’re aware of the- of the kind of the Chinese devices that can break through India. So with them, there is constant work that we are doing, where we are serving them, to ensure that we are giving notifications, we are reaching out to our users, our products, our apps are optimized and opportunist for that.

Pradyut Hande: Correct. For consumption, because that is a huge challenge for the mobile marketers today-

Siddhartha Roy: -Absolutely. And the fact that we are dealing with sub-100-hour devices – over your runtime and the amount of stories that’s available to the user, is very very finite, which means that he needs to make a choice of keeping our products out there, which means that my products have to be hugely optimized to be able to handle that basic infrastructure that has been given, which is the device of entry in play.

Pradyut Hande: Because at the end of the day, you are competing for memory space, viewing space, and watch space.

Siddhartha Roy: You are also aware of our strategic partnership with Xiaomi where we are very deep in Europe, with Xiaomi stack. That itself shows our commitment towards wanting to go into middle India, wanting to go across the youth in India, is very strong. There, we work very closely with Xiaomi as a partner where we optimize our offerings directly on to the growing lot of us.

Pradyut Hande: Right.

Siddhartha Roy: Okay, and we bring in a lot of great consumer touch points, to ensure the customer has always a fun experience. And – he’s building his consumption habit, which helps us to then move him from a free to paid.

Pradyut Hande: Right.

Siddhartha Roy: So that’s the big piece that’s getting played up. With regards to retaining the customer, is another big challenge that’s happening. And the third is with regards to payments and how we can bring in seamless payment modules in place, to give an ease of use, and create value for the customer for him to be able to shell out. There, I believe that a lot of work that wallet companies, that payment companies are doing with OTT companies, such as us, from where cash backs to credit card programs, to price-up programs, are extensively built up. And we create very very strong, try-and-buy experiences, where we want users to try the premium offering, taste what is available in the premium bucket, then move into a paid experience. So that’s a big, big play that’s happening. The other is the fact that there is – you might mind the consumer’s fault for choice, so you create reasons to keep- to make them come back to the services. So there again, technologies, platforms are brought in, to be able to assist us in creating touchpoints and lifetime values in customers where both ML and AI play a very important role. So when today you come on to Hungama music, you see the entire onboarding journey; I Choose-I get you to choose a set of languages, basis which I program and curate your experience for you. As you start consuming my product, the entire homepage of that product starts changing – changing basis, your consumption happens. Here, there is a massive ML-AI layer at play, where we start understanding what you like, we back this with consumers consuming the same content, and we start then recommending in layers, which basically takes over the overall consumer experience – so your device, your home page of Hungama Music, is my place of —-, goes into individualization and the way the averages are continuing to get better, where we are pushing more programmatic layers on to these homepage layers, to bring in a sense of personalization and a greater consumption matrix.

Pradyut Hande: Correct, so that’s something that I was actually want to get to, thank you so much for approaching that before and the fact that the average consumer today, whether he is urban, semi-urban, rural; the fact that the average listener of here today demands greater context, greater elements. Not just in the content recommendations in the website or the app, but also, so much more across digital touchpoints, right, so that’s where your AI power of recommendation engine, that continues to deliver these relevant recommendations across the customer journey becomes so much more important.

Siddhartha Roy: Absolutely, so – I mean, I’ll just share with you, we use a state-of-the-art notification there, which is basically a smart layer. We see users- I start sending notifications, I have started to realize that between 9 and 11 in the morning, you are likely to open my notifications a lot more vis-à-vis any other time of the day. So the entire AIML layer takes over and generates those notifications based on your clickables in play.

Pradyut Hande: Right…when you are most engaged.

Siddhartha Roy: Segments get weighed on the floor. So today as a company, over 1800+ segments that we would be dealing with; now, this cannot be done through ?. It works through parameters that are built on basic consumption habits and people are put into buckets; that’s when technology and our recommendation layers; our AIML suites take over, which connect to third-party platforms, to be able to reach out and create the omnipresent demand of the user.  So same- often, if you were to email that this would move, instead of a notification layer, to an email there.

Pradyut Hande: Correct, it works  taking into account what channel the user would be most active on-

Siddhartha Roy: Absolutely-absolutely and all these channels are interconnected into a single port. So it is all about us having various consumer touchpoints in place; we have an intelligent engine which basically aggregates all consumer touchpoints and then decides which should be the right, optimized touchpoint for us to be able to reach out to the user and finally, the core KPI being, I need him to come and be an active consumer, in my experience. So that then becomes- and the way we measure our core KPI, based on a core-core matrix, which is time spent listening, or time spent watching. 

Pradyut Hande: Right, right. No, that makes so much sense and I’m sure other OTT players are also figuring out how to streamline and make it as seamless an experience as possible. Now, when it comes to personalization, it comes down to how much data a user is passing to you; now think of a user who comes to your platform for the very first time; one way to get him started on a personalized journey is, of course, getting them to select those requisite genres, or languages. But you require them to log in, or at least register, to you know, sort of kick start their personalization journey at a higher level; so how’s—how-how-

Siddhartha Roy: So it’s actually – the way we look at it is that we actually start the personalization journey without him registered, because we have got cookies in play and because extreme entities, our consumption matrix or our personalization drive at the back of costing consumers, okay? So language, genre, city, all that we get, all of that we get, okay; we’re able to bring that into a single structure. We then push users to create a reason to register; that reason is many; for example, you need a playlist, you need to register. As you register with me, I am able to bring your entire social agreements, I’m able to tell your users what you are listening to, what you are sharing – it is your choice if you want to make it public or private, we give you that privacy. Okay? Then, we go into a gamification engine; we have a gamification engine which works on- it’s a patented layer, works on consumers wanting for three actions; watch, listen, or share. For the share action, you need to register. 

Pradyut Hande: Okay, got it. Because you then get to their social-

Siddhartha Roy: Absolutely-this itself creates a great playout because gamification leads to rewards. Hungama has an amazing rewards program. We earlier had a digital rewards program, in December of last year we rolled out our physical rewards stores where today Hungama merchandises-all that’s available, and you can redeem that for technically free, if you’ve been spending time on the app, we reward you by giving you points-these points are-can be burned towards these good and services. We also link our gamification layer across music stat and our video stat. So you can cross play-you can – you’ve been listening to music, earning points and you can buy a movie, pay-per-view pass, a drama or play, and be able to consume it there. And vice versa. 

Pradyut Hande: Okay, okay. That makes sense. And a very interesting angle to all this because I don’t know of many OTT players who have ventured into the entire gamification angle, overall of this. Now, given the fact that Hungama has a video streaming service, you know, Hungama Play, and then you have Hungama Music as well; How do you ensure that a user is still receiving that seamless experience, without being bombarded with too many relevant recommendations, because at the end of the day, there is only a finite amount of communication that a brand would indulge in.

Siddhartha Roy: So, first of all, it’s a single sign-in experience, right? So single SSO experience, which means I know everything about you, depending on what you are consuming. We allow for our recommendation engines to understand user preferences, and based on his notification and reach habits, we decide how many notifications or how many responses need to be reached out. And that is completely based on segmentation and what we learn from it. We’ve realized that if you open a notification every three days, it would be a lot less for your vis-a-vis a person who’s wanting these notifications, these consumption habits based on Aaj naya kya hai. 

Pradyut Hande:  Correct, correct. Definitely. Right.

Siddhartha Roy: So there are various ways – we also do a lot of regional content, both on music as well as on video on Play, that also leads to a lot of great consumers connect, because this sizable amount of work that we do, to channelize social communities of these actors, artists, celebrities. For this, engagement is there. And that plays for us pretty well. The other big play is that Hungama is very large in what we call short-form video, as a format and category. Because we are a mobile-first ecosystem, everyone knows that; we obviously have a great presence on TV containers, on various other containers- but Hungama is a very strong mobile-first ecosystem., because of the content that we put out we keep it in mind that content consumption – our sweet spot is between 7 minutes to 10 minutes of consumption. Because mobile devices are an opportunistic consumption happening. It is not an appointment-based consumption, which means the MLAI layers that we use, has a lot to do with the recency of content, type of content, if he is consuming a lot of Akshay Kumar, then he would right now be seeing a lot of Sooryavanshi.

Pradyut Hande: Okay, okay. So it is optimized at that level.

Siddhartha Roy: He-I’m not gonna send him a Hollywood Joker or Oscar pay-per-view. 

Pradyut Hande: Correct, because it is irrelevant to him. At the end of the day, everything is so interconnected, and consumption patterns can tell you so much about the individual user. So personalization and a solid AI-driven recommendation engine become so critical to any OTT player, long-term strategy.

Siddhartha Roy: And the size and scale of the community which we are dealing with is all big-little. If I’m gonna get millions and millions of daily active users, who are consuming hours and hours of content, I need to have platforms, technologies that come in and work towards creating a structure in which we are able to deal with. 

Pradyut Hande: Right, right – that makes so much sense. All right, so while we’re talking about personalization, you know, the OTT player has to provide as relevant recommendations as possible. And over a period of time, you realize that there are some recommendations that stick and there are some that don’t. So at the back end, the AIML layer is continuing to optimize, whether it is based on language, whether it is based on genre, can you shed some more light on how long it is.

Siddhartha Roy: So it’s an always-on structure within ourselves, to a team that looks at search and recommendation together. It’s not actually recommendations alone; I believe the entire structure works together. It works very closely with our programming grid because there’s a curation that has to work; so there’s what’s in it for the user, so there’s an effective ROI curation on measurement as well. When a container has a piece of content, if the CTR of that piece of content does not deliver, then programming teams are asked why that content has been placed.  So it is a technology that takes over that action, we have a large amount of human intervention, as they build the logic which the recommendation search engine uses to be able to outline a consumer offer, out there, in terms of the overall packaging that happens on the apps.

Pradyut Hande: Okay. So I think it is fair to say that complete reliance on AI is not the way forward; it is human intelligence at the end.

Siddhartha Roy:  It starts actually-if someone tells you that by putting an AIML layer, it does-it does work. Like you have a constant human intervention; AIML is part of your programming cycle, which allows for user consumption detail to go in, for it to generate a set of actions that makes the user consume your content a lot more or multiple times.

Pradyut Hande:  No, that makes a lot of sense. And on a final note, I would like to pick your brains on what you think is the way forward, when it comes to the application of the Martech or Martech technology, the evolving use of AIML, especially in the OTT space; how do you see it moving forward, what more can it do for players?

Siddhartha Roy: So I believe, each business is it’s own. We will treat it very differently to categories in our space, we’ll treat it very differently. So the fact is if you are in India, you’re dealing with hundreds and millions of users, which means you need to get technology to enable it. Okay? There are state-of-the-art offerings, of external third parties, sitting and building it yourselves, I believe that it’s a hybrid formula, that’ll deliver on what the objectives of these businesses are. I also believe that the core KPIs need to be extremely focused and these have to be based on short-term windows, which allows you to transition your business from the basis, on what your core requirement sets are. In a nutshell, the human capital is as important as the technology and the platform that you get to the table; you need to ensure that that works for you because these are largely Sass structures where you need to bring in your human capital, you need to homogenize the sass platforms according to your requirement; it should be able to give out an output which delivers for your consumers.

Pradyut Hande: Correct. No, that makes so much sense, and that’s precisely what Martech also does, enabling marketers and product managers to deliver highly customized, personalized, relevant, and contextual customer experience, across various digital touchpoints. You spoke about the critical significance of personalization across multiple touchpoints, through the entire customer journey, and that is something so valuable to businesses across the board. And on that note, I would like to thank you so much, it’s been an absolute pleasure. For more such content, you can subscribe and tune into Mastering Martech.