 good afternoon you know the sessions after lunch soon after lunch are always a little a little precarious right lots of people are you know had a good meal and it's not a great time but what we're hoping to do over the next 45 minutes is just take you through a perspective on on personalization increasingly how brands and marketers will look at personalization but more importantly trying to get the whole customer perspective first what is it about personalization that the customer seeks and wants from brands and what are those little eerie bits that customers feel a little uncomfortable with right when personalization gets on to you and you know kind of invades your privacy and there's a very fine line and we're gonna talk about that because we've got a you know pretty diverse panel there is a Bipol from Ghana so he'll talk about the whole idea of you know music streaming and how they use personalization we have Varun from zoom car which is like the ever since and you know of India so effectively how you can have your car experience personalized and how they are in using data in a big way and of course Piyosh who heads the entire data platform at make my trip and the whole travel space so that's really how we gonna play this out I'm gonna ask Deepak to just play a couple of slides to set the tone for what we're about to discuss we have a clicker which okay here we go yeah so really essentially what you're gonna tackle is really what is personalization right what we're saying is personalization is all about how relevant I am across all interactions to my customer right how is it that I can present myself in a manner that's relevant to him so it really enhances his experience whether it's his buying experience whether it's his shopping experience it's really that that's one aspect and there was a survey done in terms of what do customers value the most two big things one is convenience and the other is personalization right everyone wants it convenient every wants it personalized so it's really designed to take into account the user's own perspective just a quick overview on what shoppers really want from personalized marketing I think the first thing really want is relevant recommendations things that they wouldn't have thought of for themselves right some of them some of us experience it in our online shopping experiences but effectively we have to recognize an omnichannel world that's the same expectation we have even while we are shopping in the real world when you walk into a store you've got your smartphone on you it's a something within the store ambience that can direct you into areas of interest to interest talk to me when I'm in shopping mode right you don't want to get those shopping messages when you're in the middle of work but there are times when you're out with the family on the weekend that's what you want to be spoken with and that's moments which are critical remind me of things I want to know but might not be keeping track of right how many guys here forget their adversaries right it's good to get a reminder a week advance make a reservation at a restaurant of your choice by flowers for your your partner and that's important no me no matter where I interact with you right we don't want this seamless you want a complete seamless experience right I might be in my laptop sometime on my phone sometimes without a device how do you reach out to me in a relevant manner and share the value in a way that's meaningful to me I think that's really what customers seek is really the whole meaningful interaction so I'm going to summarize this in a kind of you know value equation saying that the value that you drive to the consumer is the relevance and the timeliness of your action but keeping in mind the fact that you don't want him to feel that he's losing privacy right so never get in a sense you know make a customer uncomfortable so the whole thing is enveloped in a box of trust I think that's the that's the key part that that customers are seeking so have this you know a market one is the wonderful cartoon it says it pretty much right so here's this lady who says no I don't give consent to my personal data no cookies no tracking no browsing history the less the site knows about me the better right this is what she wants hey but why isn't the shopping experience more personalized so you'll always have this little dichotomy with customers right they want all the privacy at the same time they want that experience personalized so really that's how you know that's that's really the dilemma that marketers face I'll just throw it up into my co-panelists right now I've introduced them this Piyosh was going to go first he's the head of data platform engineering at Make My Trip we'll have Varun going next from Zunkar and then there's Vipul from Ghana who's gonna talk a bit right so head over to you thanks for giving us quick bites on personalization I'm going to quickly talk about a couple of things which we have done at Make My Trip hopefully some some of the learnings which we can share it with the audience here so thanks for joining me here at Matic so just to quickly talk about like so Make My Trip has different line of businesses so I'm sure like so most of the audiences would know like flight, hotels, holiday packages, ground transportation and then like so my experiences which is a new line of business which is added so for our customers like we want to personalize the whole experience when they are onto the platform and when they are off-final as well and the journey to personalization started with like personalizing the experience when the person is there on the app means like so when they are there in doing any activity on the funnel but like so we earlier had this joint systems which were like so any off-final engagement is handled by a marketer talking about some of the campaign related stuff which he wanted to communicate to the users to unify the whole piece right so we built in-house our own CDP and the key pieces so just talking about like so what what are the key things while doing personalization so having a customer data platform really helps you do the personalization in the right manner the relevancy is really important right so like understanding the user intent earlier if you have seen ads etc those were retargeting based on the user recent or last interaction itself but like understanding the various session chains like so the activities done by the user etc it's really important for anybody to come onto a platform and like interact with a good customer experience itself then like so there are good amount of data insights which any organization generate in-house right those insight need to be utilized in form of persuasions etc and need to integrate it in the right manner on app and while communicating to the user via notifications etc so I'm going to quickly talk about these things maybe skipping a couple of slides but yeah so talking about the customer data platform right so the very key stuff for us was instead of looking at bunch of rows and columns you need to really treat every customer has its own database with you to understand like what he's doing his customer journey etc so we have built in-house called Boulder DB which integrates like per user databases where I can store like user journeys and like his profile data device details and many other pieces related to it so and hence like having it in as part of your CDP really helps you nail down to a one-on-one level rather than a hot level or a segment level itself the other key pieces the quality of data is really important so the key is like when you are generating that data or while you are capturing the data do it in the right manner and hence like so having the right schema or common data model at the back to capture those event of users really helps you do a good stuff while you are like getting insights out of it or building a customer experience on top of it for the user the other key pieces like so most of the stuff is like so you only do gather from the client events right so but like there are a lot of information out there at the back with the server side where you can do a lot about like for example if you are taking a pricing decision based on if you want to just personalize the price point to a particular person you have the information about like how it is getting computed at the back and the other key pieces you need to really do a good job and stream processing to get to a like millisecond performance on to stream processing itself. The other key pieces you need to decouple your customer data platform from your use cases or the various data products you build on top of it be it like if you are building a orchestration layer on top of your CDP for campaign management etc. This doesn't need to coexist like really like a monolith sitting on top of your customer data platform and hence like so that helps you do a better job at doing the personalization across various channels itself so that's another key point. Relevancy as I said like so you need to really know the user from his past sessions what he is doing what he is looking for and do a right job of communicating with the right deep links when you are communicating with the user itself so that that is one of the key stuff. Persuations like so where and like so the insights which you are generating like so 80% of the hotels book XYZ location etc. You want to use all those insights to do a better persuasions while communicating to the user. So this is one of the other key piece other key learnings on to working at like so building some of the stuff the last thing probably is like so do there is no magic wand right there right so which is you need to do a good amount of experimentation and while doing experimentation multi-armed vended help really help you allocate traffic to the right winning experiment which you have configured so utilize MABs as much as you can while doing experimentation that really helps you and the contextual banded like so it's not just the experiment which is winning you are also passing on the user context or the search history of the user or the actions are you performing right now to your machine learning models to take out the right decisions yeah that's about it. Thanks Piyush so I think what Piyush attempted to do was really take you through a little bit of the back end technology involved with how to make the customer experience more personalized and in a sense he's the he's the only geek in our panel so logically we're gonna move to the market here but I was talking to Piyush just before we got on he was talking about how a lot of his time and make my trip he actually spends with the marketing folks in terms of you know how do you how do you drive usage how do you drive you know increase loyalty and you know kind of cross-sell and upsell so really it's the interface it's really a lot of the back end technologies that powers great customer experiences and great marketing and therefore it's a logical segue to actually have Varun come on next and share with us his experiences in how he's using personalization in Zoom car. Great thanks Lloyd and thanks Piyush lovely audience full pack today quick you know like intro and Zoom car those who are not aware that Zoom car today is the market leader in the cell drive space and we have like more than 15,000 cars in a fleet and we give you know cell drive cars to for rental you know to the end customers and we have a user base of to the north of 20 million and we've been operating in there for last six years now so if I get the device so yeah I mean like I will quickly of course like Lloyd touched upon personalization but as a marketer you know like what exactly if I have to you know like talk in a very at a very fundamental note like what is it like does does it and nicely done relevant ad copy is it about creating bunch of hundreds of taglines have a lot of creative variants have very relevant in-app banners or you know display banners I mean this was quite shocking and troubling especially for all marketeers and this was this survey was little survey was done by e-markety earlier last year where you know most of the CMOs and you know the marketing heads they really struggle okay like personalization is the best word and a lot of tech is at play but how do you how do you actually attack the customer and get the business implications out of it right so just wanted to share this like you know like I this is actually not Photoshop this I got from one of my colleagues at work at zoom car and it's a personalized can from coke so the moment coke had their campaign share you know share a coke campaign and you know that this really drifted the sales by over 5% in UK when it was introduced in UK so what exactly you know like you can see a brand like an FMCG brand doing a lot of personalization stuff where there's not of tech play involved I mean before jumping let me let me skip this in the interest of time yeah so basically like I like you know I am fond of saying this that all of us know but just to summarize it actually marketing if you look at marketing marketing is actually the new finance I'm sure all of us realize this right it's all about numbers while Piyush was talking how do you get numbers from different segments different tables put it into you know one place and then obviously how you make meaningful actionable insights out of it so the holy grail let me be very make it very simple for you guys the holy grail of personalization today is that having segment of one so it's no more about you know target audiences this is sake this is sake be and especially for us at zoom car this is the holiday road tripper this is a weekend commuter or this is a weekday commuter nothing like that you actually if you actually have to make you know juice out maximum benefits from personalization you have to create a segment of one so I need to know who Varun is what he does what are his digital footprints and then what are the data available at my platform and then how I can obviously target him better so in order to create so let us look at this user Vinay so at zoom car what we look at this that how long he has been my customer what does his preferred city where he actually picks a self-drive car then what sort of customer he is champion customer would be someone who is use zoom car thrice or probably more than that then what is his preferred card hatchbacks had done you know like and what sort of brand he prefers what is his destination what is his preferred destination right where he actually goes what was his last booking with my platform and whether as I said whether it's a weekday weekend customer and then we arrive something called as driver's score interestingly we are one brand since we are into O2o space while most of the companies they do a lot of in app personalization we do a lot of in trip personalization because for us the average duration when a customer books a car with zoom car the average booking duration is more than 50 hours which means with for 50 hours he is inside a car offered by me and imagine the kind of experiences I can co-curate basis on the information or the single view of this customer I have right so I will quickly jump to that slide how do we achieve at that is this by using this in trip personalized recommendation example so basically you know we have an onboard diagnostic tool which is powered by IntelliCar that is placed inside within the car so basically tells me the GPS location of the car the latitude and the longitude it tells me that this user Vinay you know like how you know how many turns he is taking how well how he's using the car is he driving rush driving or not what is the clutch what is the break what is the acceleration speeds so we use lot of these onboard diagnostic signals and since I spoke about single view of the customer I also have Vinay's lot of marketing data a lot of user data a lot of digital footprints so I so for example I know for a fact that Vinay has been with zoom car for four years his preferred location is somewhere in say towards Manisar and he's starting from Gurgaon and he loves lot of KFC burgers so what exactly I will do is that moment this customer Vinay is actually going and you know he's like 20 minutes 15 minutes away you know like due to like he's away from this KFC restaurant I'll throw him a notification because I have done an in-trip I have done an in-app you know integration with setups like hybrid delight ease my trip and these setups they create co-curated experiences and offers when you're on the on the move so moments Vinay is approaching KFC he will get a push notification saying okay hey by the way you can get 10% off powered by zoom car powered by hybrid delight and I know for a fact that Vinay has been using me and loves burgers so then then obviously this user gets really excited by the kind of relevant you know in-trip experiences that I added the second example I want to give here is that we are actually in talks with you know music streaming player of course we have people from Ghana but we are in talks with Spotify so very soon we're going to launch custom playlist so I know for a fact that Floyd is a holiday road tripper loves go to you know going to Ladakh, Leh and Goa and can I have a Goa playlist can I have a custom playlist like that so I'll so this basically what I'm trying to accentuate here is that within the trip since you have a lot of information of the customer you can actually you know co-curate a lot of experiences during the during when the customer is on the move and how we are actually doing this is actually we use an open source called Kepler so Kepler is an open source which is used by grab which is used by lift which is used by Uber so basically it's an open source and it has an input from Google Maps API and over and above the moment I input my latitude and longitude you know real-time signals that I get from the cars OBD system I can actually tell where this customer is so this is actually this map is actually you know all the trips originating from Bangalore real-time and I actually know where this customer is right now how many times he's stopping so basically as a market here this helps me in real-time taking a lot of optimization decisions that whether I have to throw him offer or whether I have to tell him hey by the way there's a washroom you know probably 10 minutes you know from where you are and if you want to have you know obviously go to the loo you can obviously do that so yeah more interesting stuff on that and this was the marketing stack that I was talking about so we have car data we have user data we have a lot of response data from the IoT device then we create a single view of the customer by amalgamating all this information and obviously as you can have a lot of analytics and recommendation engine you can create a lot of offers so this is an example of an app personalization so for example a customer like me I love a lot of hatchback I always take the top variant I always take without fuel I want the car to be home delivered I prefer leather seats I want the car to be scented so I we are launching a service called zoom premium but I know for sure that if these are the things this customer has always preferred on my platform I can actually monetize on this and he will love this experience right so that is an app personalization example and this is this is an example of driver score so so basically as I said initially that we create a driver score and this last slide of my presentation today that we create just like credit score in the credit card world we have a driver score where we know for the fact that how what was the behavior of this driver did he pick the car on time did he fill the checklist hopefully he did not do a rice driving right and he maintained the car he you know dropped the car back in the clean condition so if we create a dynamic score which gives a credit credibility which means if you have a high driver score with me I know for a fact that you're my you know superpower customer and I should leave no stone unturned in giving you delivering an amazing experience which means if I have to wave off something like a security deposit next time you are booking a car I will have a differential or a dynamic security deposit I might even wave it off so just to summarize this point the fact is that while all of us you know we talk about personalization we assume car actually you know like doing a lot of part breaking stuff as as far as you know entry personalization is concerned because without which you know we just can't sustain the business model because it's all about delighting the customer you know whether he's booking on on your app or whether he's interacting with the brand of the road thanks cheers thanks Varun so that is the interesting perspective I think that the thought that you know he leaves behind is that car rentals are a legacy business evice and hurts have been around for decades it's how you use personalization to make the driving context and the experience more seamless and interesting right while both make my trip and Ghana are digital first businesses car rentals are a whole business and that's really taking so I think a lot of the challenge of personalization is how we gonna use personalization with existing legacy businesses the more traditional categories like fast moving consumer goods and consumer durables and then make the whole customer experience more seamless I think that's one of the lessons I'd like to get Bipu now to kind of share his perspective music streaming is a new business it's it's changing the landscape of you know how people consume music right we've all seen the eras when we collected vinyls and then CDs and then we downloaded music I think now it's pretty much you know stream as and when you want but more important to what time I guess everyone in the room has experienced it how we see that the streaming sites really know our tastes sometimes better than our spouse right what kind of music you're gonna listen to what time of day you know what do you what do you listen to on your morning jog what do you listen to on your drive to work what do you listen to when you're having a low day and so on and so forth so would love to have we'll share his perspective we pulls from Ghana hi everyone I everyone's feeling relieved I'll just talk about what we do from a larger business standpoint so Ghana is India's largest and most favorite music streaming app as you all know by the way good luck with Spotify I'll talk about how we use personalization and on what metrics we try to do that I mean Pugh has covered a lot of it from an engineering standpoint and you know conceptually more or less everyone who is investing in data uses some or the other variant of what Pugh has talked about so he's covered it well what I'll talk about is that you know we use personalization primarily for three main business goals one is consumption what I mean by consumption is content consumption second is to marketing like Warren talked about and third is monetization when I say monetization so Ghana is an ad supported platform the market is who want to use Ghana as a publisher they need a lot of personalization like one wanted to target a specific customer with a specific behavior so if he wants to use Ghana as a publisher he would need a lot of data around my audience and he would say hey I need to target only these many audiences so that I can get the maximum bang for my buck that I spend on Ghana right now I'll talk about the consumption bit first yes so I'll see how much we can cover because consumption is the most important bit for us because the whole business is about content if you get your user to spend more more time on the platform I think the other two goals get taken care of which is marketing and monetization now on the consumption bit so when you I mean traditionally it was like when you come on a platform like Ghana or any other Netflix I mean say six years ago most of the users would of four to five years about depending on which country you have been looking at your first thing is that you come browse and figure out what to do play then you search again and then you figure out what what to do next that behavior is changing users come on the platform they want me to tell them what they want right and they expect the platform to figure out what the user would like so today if my personalization used to drive X of my consumption say two years ago today at least drives 20 to 25 X of my content consumption on the platform so today if you come on Ghana and consume music let's say as a new user you have come on Ghana consume music for the first session which may last around 20 to 30 minutes which is about five to six songs depending on the songs that you're listening and how often you're skipping them I would know a lot about you in that session and the next session that you come you the first song that you listen I mean the song that you listen post the first song rather you wouldn't need to feel like browsing or searching we'll push you songs that you like and you don't you won't feel bad about it so what we do we look at so we look at metrics like the simpler metrics like the error of the song that you have listened to the playback artist of the song that you have listened to the actors in the song that have been involved the language of the song and then more complex attributes like the genre of the song genre is a very complex attribute by the way I'll talk about it if I get time and then the waveform attributes of the song right like the acoustics of the song the treble of the song the bass of the song and then we push you those songs list based on your listening pattern and personalization has generally two goals for a content consumption platform one is to you know let the user get what he wants and from a business standpoint I would let want the user to discover as much content as possible because you know browsing in general helps you discover content because you're browsing and figuring out what you want and stuff like that but as in when we move we have evolved over the past two years you don't want users don't want to browse but at the same time they want me to give them content which they haven't heard but they like so that is where the discoverability angle has to be taken care of in the personalization algorithms that we do so let's say if you have heard of a song let's say a party song which is from 2018 and 10 minutes later you may get a song which is from 2000s but it still has the same set of beats same set of waveform attributes and you wouldn't even feel like you've heard a song from 2000 so that is what determines the success of personalization on a music streaming platform from a content consumption standpoint now talking about waveform attributes and content you know we invest a lot in content now gone are the days when I say ha ye song I mean it cannot be done that way because the content is very expensive so we look at the waveform attributes of the song and figure to figure out what's going to be a hit in future like what works on my platform what does audience on Ghana like what kind of song do they like and then we figure out okay how much to invest in a particular piece of content that has come on our table right so these are the large broad parameters on which we drive consumption I mean I can go on and on and you know a few shows obviously talked about the last thing was that you know how it is more very important to keep experimenting and keep tweaking your algorithms so that you know you are in tune with your user you know user habits change very frequently you like one music today you don't like like I like I used to like mother they should ten years ago I don't like I like Katrina cap today so you know so those those patterns have to be taken care of and algorithms have to be figured out like say in Telugu and Tamil your data may tell you that you know people like to listen to song based on the actors in the song but in Hindi the data may tell you that people like to listen to songs based on the playback artists in the song these things are very very important when driving consumption on the platform through personalization algorithms one one algorithm need not work for every user on the platform or every language I quickly come to the monetization bit and then I handle the mic to light so on the monetization bit one of the so we all know about you know how Google has made dynamic display ads and you know how that ad gets ads creative get created on the fly based on the data of the user I mean data would be all sorts of data depending on the context you are looking at now from a music streaming platform given that most of the market is budget has been spent on radio if you have to attract those budgets and say you have to show them audio ads you know because a user on Ghana spends what I think 60 to 70 percent of his time in the background like on the headphone and maybe 30 to 35 percent of the time on the screen so so if I have to monetize the platform which is ad supported I have to show them audio ads in a very very targeted fashion if I have to divert the radio budgets from a marketer to a music streaming platform so the audio ads we have been working on highly dynamic audio ads where we look at you know various parameters of the user on the platform and then serve an audio ad to you which would where the creative gets created on the fly and it is a and and then it's you know the the success rate of that ad or the recall value of that is a lot higher so let's say a sweetie wants to advertise on the platform I would look at what time of the day it is what temperature it is outside where the user is I use his location parameters and then I'll serve him an ad talking about a specific beverage or a food which which is in line with these parameters that I just talked about so those are the dynamic audio ads that we work on then we are working on voice activated ads where we use a lot of artificial intelligence to talk to the user and it becomes almost conversational I mean we'll be releasing them soon we are already doing beta in-house for those voice-activated ads yes thanks people I think interesting you can sit down I think we'll just spend the next five odd minutes talking through a couple of key issues pertaining to personalization I think the first part which you know each of you can you know if you can answer in less than a minute is really a lot of personalization is really data that sits on your platform in the context of Ghana it's you know you know people's usage patterns you know what they listen to in the first few listening and over time you know kind of in a sense machine learning kind of refines this whole technique and comes good is there some degree of fertilizing data from other sources without kind of infringing on people's privacies in the context of Ghana you know you know every morning mostly on my way to work I listen to Coldplay and in the evening you know I want more soulful stuff so I listen to Mukesh or Muhammad Rafi so you kind of get a sense of my music patterns while my conversations on social media on Twitter I might be you know going ballistic or Mika how is it that you can use data from different sources say from social footprints to make the experience more enlightening and more seamless. So we look at a lot of trending data on alternative consumption platforms like YouTube and I cannot not talk about TikTok these days and then use these macro data to feed into our personalization algorithm so if I have to push a playlist which is a trending playlist to a user it won't just be based on his consumption on data to aid discovery for my business and I have to consume these data points around me on other platforms and then push a trending playlist to the user you know so the trending on my platform is very different from a trending on YouTube or a trending on TikTok. In fact just to add to what he's saying I mean the lines actually you have to be a little bit cautious because lines becomes little bit blurred because it might be a breach of privacy so for a fact I know if someone is logging to Zoom call or any app for that matter he can log in through through Facebook he can have a true caller login he can through have a Google login and for sure I know these apps do a lot of listening they have a lot of data available so I mean it's obviously a matter of ethics as well that how to what extent or what degree kind of information because user has not obtained for that data because it's written very miniscule somewhere on your app or on a website that when you log in through true caller of Facebook we'll pass your information right so I mean the biggest example is the face app you know that went viral right because they said that we'll have free royalty not to use your pick in whatsoever shape and form that we want so that's a biggest breach of privacy as well at any point of time we don't use the specific data to push them to clarify we don't reach any privacy those basics are always the other part and you know I want to get this to push is the whole issue of using data to actually direct consumers into slightly more profitable avenues right so in the context of let's say a travel booking site you know you know this guy is a regular booker of you want him over time to use hotels or to use you know self-drive cars and such like so to what extent does harnessing data help in behavior shaping right so if I'm an office guy I know one of my team colleagues you know always makes last bookings knowing fully well that finally booked out and he might get bumped up to business class so do you all enable a certain degree of behavior shaping patterns that that can be used absolutely and everybody does it like so the cross-sell upsells are the old keywords to do it but like so in our case primarily like so whenever somebody is coming on to our platform and searching for something like be it like like for flight travel or holiday package or anything right so we create a pseudo itinerary right that pseudo itinerary may have you may not have searched or done anything on to our platform with respect to other line of businesses or other offerings or other products which we have on to a platform but like so at the back to for me to do a recommendation right what is like is it going to be hotel with which star category and other pieces which we want to recommend it to you to get a more wallet share from you and those kind of things definitely that helps us not just on the business side of it but fulfilling the user requirement as well because on a trip when I am creating a pseudo itinerary I'm trying to predict okay what all your needs are so it's a win-win situation both from consumer and right where I'm trying to help him build the itinerary like based on the recommendations which I'm curating on to the platform the other is like so on the monetization side of it like so we get that particular point okay so it's brought basic this into a slightly broader question so for the various businesses and brands so therefore is there the like the equivalent of a war room right so if I have to get better customer experience do I need to get let's say a data scientist a technologist a marketer a consumer insights specialist together and then brief them so how would for example zoom car and and Ghana look at creating this kind of marketing agility I think that's a fantastic question basically as you rightly said it's a war room so like having a great data engineering is mission critical because you need to have clean data in one place then on top of data engineering you will have data decision sciences who will tell that while this is the data from different tables what are the actionable insights and they will give you actionable insights they will not own the business implications and there where you have consumer research specialist and marketing specialist they will sit okay with this customer has these many variables and attributes now I can have what layers of monetization keeping the customer experience intact so yeah like to your point it's it's kind of a very well oiled you know assembly line which are probably a war room if you have to say that everyone has an equally important role to play but like you just can't discount data as I said like marketing is a new finance right data is at the epicenter of it and what extent of this happen in in Ghana everything is very very data driven these days marketing I mean the best guy to hire is a full stack data scientist who can code as well so one of the key things I just to add like so apart from people and the skill set you need to build or invest on to the right platforms as well so the right tools and platforms which can self-serve or help those fellows like who are working on it helps okay so great I think we've covered up you know broad swath of things pertaining to personalization I think the one point that most speakers reinforced but I think we all need to remember is the whole fine line between customization and privacy to what extent you go to data mine customers data outside of your platforms and to layer it on to make more enlightened decisions to shape behavior that doesn't violate on individual customers privacy and increasingly that's gonna be you know a big watch out right while policy on this is still evolving and a macro framework the reality is as business decision makers we all need to watch out for that so that pretty much closes our panel discussion I know if you have time for just one or two audience questions if there are any and thereafter we'll wind up so anyone's got an audience question you can stand up raise your hand identify yourself and and say my name is Sunil Sharma as a marketer I have a question how do I invest because if I advertise on any platform you know I'm a roadblock to his entertainment so he feels very irritated it doesn't happen on a TV when we interact with the consumer this is TV but when we are on this thing and we see any advertisement is the irritant how do I invest more over here I think that's a great question so basically a vanilla advertising will not you know fit the bill here because as I was when I was speaking I mean you have to co-curate experiences if I have a custom playlist for that person right and I can monetize on over and above that second example I was talking about that you can have whether when you're using say something like zoom car and within the trip you have you know tied up with all the restaurants on the highway right and then you can throw offers so basically what where I'm coming from is that if you're going to show him a random ad about a biscuit right that will not fit the bill right you're just spamming your customers a very very bad customer experience right so point is that you have to go deeper into content and you have to co-curate the experience otherwise you will see that the native advertising basically the platform that you're using they will see high amount of uninstalls happening absolutely it has to be and that is why like before this discussion started I mean I was just having a discuss chat with quick chat with you know ripple and we are also like quite eager to you know launch something like Ghana playlist on zoom app so app and app integration otherwise you're just spamming your customers we've got severe time constraints we'll have to hold it to just one last question I think Zubin Zubin's waving his hand can you help me guys yeah I'm Zubin from Shamaru Entertainment we've just launched our OTT app Shamaru me the question to all three of you is where does video play a role in engagement and personalization for all your businesses I think less than 30 seconds because we really under time so 30 seconds we recently launched video on our platform two months ago so we do two formats of video one is short format and one is the whole video of the song so you know one of the objectives is to get the user to at least being a music OTT app one of the objectives is to get the user to spend more and more time on the foreground rather than on the headphone so that way video is getting more and more engagement on the video helps us meet that goal because you know why we do that we say because audio ads are yet to take off in India and you know rest of the ads formats which are in the foreground they do better so the more the user spends time in the foreground the better it is for us video can play a very important role there obviously there are video has a very different sort of licensing issues when compared to music but that is something how it is just to summarize you can't ignore video 80 percent of internet traffic by 2020 is all video going on yeah there you go okay that's pretty much it I'll add it back to