TRANSCRIPT
Tommy: [00:00:00] CVM is basically a mix between data science and then the marketing itself. And that’s where all the game is.
Tommy: [00:00:07] We keep expanding our CVM to make sure that our CVM is everywhere. It’s in the all channels. So our customer will find that, okay, I can get the good offers by entering whatever channel that we have.
Tommy: [00:00:19] We have been trying to tap into a lot of events actually, which will make the offers become contextual and then without disturbing them.
Tommy: [00:00:27] One month revenue in the previous four years is basically now only one hour revenue at midnight.
Exacaster: [00:00:34] Welcome to CVM stories, the podcast on customer value management. Together, we explore how companies can be more successful and the customers happier through the use of the latest customer value management techniques. Learn key commercial and analytical insights from telecoms, retail, finance and other industries that drive CVM forward.
Egidijus: [00:00:53] Hi, I’m your host, Egidijus. Today we are joined by Tommy, a Head of CVM Data Growth Strategy at Indosat Ooredoo. Tommy is an expert in mixing data science with marketing to help reduce customer churn, while also growing companies revenue. Join us to hear how Tommy employed 24 different communication channels for his successful CVM initiatives.
Exacaster: [00:01:19] CVM Stories is produced by Exacaster. We help companies take their customer value management to the next level. To stay updated on our latest episodes, subscribe to the podcast or sign up for an email newsletter at exacaster.com/cvmstories.
Egidijus: [00:01:33] Welcome, Tommy, thank you for joining us today.
Tommy: [00:01:38] Thank you, thank you, Egi.
Egidijus: [00:01:39] Could you briefly introduce yourself to our listeners?
Tommy: [00:01:42] Okay. Uh, hello. Hi. Uh, my name is, uh, Tommy Wahyudi. I am from Indonesia. So currently I’m working at the second largest operator, we can say, in Indonesia, where I work as a VP of base monetization. Uh, basically this one is implementing CVM and then trying to get, uh, increasing the revenue by implementing a CVM technique, actually.
Egidijus: [00:02:16] Um, so, uh, Tommy, I know that they’re very little people who, uh, you know, uh, start their careers and become CVM’s, uh, immediately, you know. Yeah. Could you describe a bit, uh, your path? How did you, uh, join telco industry, and how did you become a professional in this area?
Tommy: [00:02:42] Uh, before the, uh. My career in telecommunication. Actually, I, my background of uh, uh, education is, uh, information, uh, technology, information technology. So basically I’m familiar with, uh, programming and then also coding and all those, uh, technical stuff. Once I, uh, graduated, I. I went into a software developer house and then, uh, spent my time two years over there, then I think. Uh, I feel it’s not my patient. So then I. I try to join a telco company. Uh, at that time, it’s a CDMA, uh, company. And then from that one, I. Keep jumping around into the telecommunication company, I think around, uh, 4 or 5, uh, all in the telco company, then I. Over there. I started my career as a. Reporting and then almost operation in sales. And then I jump into the analytical work, and then I switched into the marketing after that. And then after that I get into this, uh, CVM, uh, kind of thing. So it’s at that time it’s very, very new and it’s quite interesting at that time.
Egidijus: [00:04:04] Uh, so, uh, so you, you basically did the a best practice, you know, for customer value management. It’s an extremely cross-functional role. And you basically touched, uh, almost everything, maybe some base stations, you know.
Tommy: [00:04:17] Correct. Correct. Correct. Correct. Correct. It’s it’s helps me actually, uh, from all the background that I have. So help me to, uh, to make this more, more, more richer.
Egidijus: [00:04:30] Uh, what would you say is your, let’s say, uh, how how development or, uh, programming skills are helping you on a daily life?
Tommy: [00:04:41] Oh, it’s helped me a lot, actually, because because, uh, CVM is basically is almost, uh, touch a technical things also. Right. So I think by understanding how the thing is, uh, working at the back end is actually helping you to build, uh. A good logic that can be delivered into the platform because yeah, when you do the programming. So basically you understand the what kind of algorithm it’s supposed to be and then where the data is being fetched and then how it gets delivered to the customer. So I think it’s quite, quite helped me a lot actually understanding understanding how it works.
Egidijus: [00:05:24] Okay. Uh, what would you recommend the marketing people to at least study programming, you know?
Tommy: [00:05:30] Uh, no, I think I think with the AI is coming, I think it’s very now it’s it’s more on how you develop your mindset and then how you you create a concept. And then I think a lot of tools is currently is available to, to to do that actually.
Egidijus: [00:05:45] Mhm. Uh, you mentioned that, uh, the CVM function itself in Indonesia is relatively new. Um, maybe with uh, compared to other, uh, let’s say, disciplines. Yeah. How would you describe what, uh, kind of, uh, your role in organization currently?
Tommy: [00:06:08] Yeah, yeah. Uh, CVM, I think, uh, some is, uh, referring this one as a CLM, uh, customer lifecycle. And then it got involved with the evolve to CVM. Uh, from what I see, this one is start to being introduced in telco in Indonesia around 2000, 2010, 2011. With at that time. It’s very, I think, uh, traditional. I can see where you do a batch, uh, campaign. And then where you uploading the white list? Then it got evolve and then, uh, more data science and and modeling and then, uh. Real time campaign engine. And where then after that it got evolved into the current state. So. So currently, I think, uh, xvm itself in where I work is, uh, now become a significant, uh, revenue contribution, where it was just a very, I think, a small, uh, small contributor of the revenue. But as the, uh, we implemented this, uh, CVM uh, more and more, uh, I think the acceptance in the market is very good. And then we see the traction over there, and then people actually, uh, like it how we do the approach of this, uh, CVM so I think I think we are we are at the pretty much at the mature stage in the implementation of CVM.
Egidijus: [00:07:44] Um, and, uh, as you are kind of, uh, seeing the transition. Yes, in the organization from, let’s say, uh, focusing mainly on acquisition, growing, etc., uh, like the customer base to now managing the value of, of the customer base. What kind of challenges have have you seen in the organization like, uh, how do you educate that organization through the process, etc.?
Tommy: [00:08:12] Yeah. Interesting question actually. Yeah. Uh, because, uh, you know, uh, CVM is basically, uh. A mix between, uh, data science and then the marketing itself. So. I think it meets in the CVM because I think people in the organization always one thing. They become a Google or or the technology company actually, where they utilize a lot of data actually, uh, without. I think not so many. Uh, not so many. I think, uh, understanding that actually in the CVM itself, I think it meets over there the data science and also the marketing. I think, uh, it meets in the CVM. And then, uh. That’s where all the game is, I think I think is is being played, but they only, uh, I think like you say, uh. What we see. I think the, uh, because this one’s EVM is is new, right? I think it’s a new, uh, kind of, uh, uh, branch. So I think a lot of CVM guys is, uh, being addressed here. You are just selling a cheap, uh, selling a cheap, uh, product to your customer. So that’s why you are growing. Well. Well, actually, that is not the case. Uh, yeah. We we we trade off the the yield with the with the. With the subscriber to make them spend more, right? I think so, I think eventually then, uh, after 2 or 3 times of, uh. Question. And then also we even do some experimenting to check whether you are really increasing the revenue so we can now at the sea level, I think our sea level is quite, uh, educated on how the CVM is, is helping the, the organization. I think so even. Yeah, I think it’s it’s a long journey to into that one because yeah, it takes time to people to understand the CVM how it works.
Egidijus: [00:10:18] Mhm. And uh, when you talk about let’s say uh, when you align, um, the strategy with your sea level, what kind of KPIs are actually delegated for you or, uh, what kind of, uh, primary objectives is delegated for CVM area?
Tommy: [00:10:36] Uh. CVM is started as a main KPI usually is, uh, reducing churn. So what we see is, uh, that’s how usually CVM is, uh, being, uh, given a KPI. But, uh, then, uh, after some time, it’s, it’s, uh, it’s evolved into the revenue also because. We also, uh, trying to reduce this churn by implementing a, uh, by giving a good offer to the subscribers so they not going into the other operators like which now? It’s one of the CVM. Uh, KPI for us is also a revenue, but the way it’s, uh, being measured is not only that KPI, because I think we on the CVM side, you need to see, uh, I think two, three level, uh, below, such as, uh, even the revenue is, is there, but it’s supposed to be. Are you giving, uh, incremental revenue, uh, or not? And then what kind of, uh, gap between the offering that in the ETL and then the, in the CVM. So we are not trying to make it too low. And then. Yeah, basically arguably the is the one that is we are trying always uh, constantly checking.
Egidijus: [00:12:04] Um, and how do you solve the handshake between, let’s say, acquisition function and customer value management? You know, this is kind of, uh, usually, uh, a typically a challenge, you know, because. Yeah.
Tommy: [00:12:20] So, so on the our current, uh, function, uh, we are splitting our base into the less than 90, uh, less than 90 days. We call it a baby care. So anyone who’s just. Joining the network. They will handle by that function. And then after that, uh, after it is more than 90 days is usually fall into the we call it base management. So where over there. I think we are, uh, having a pretty major, uh. Development on the our subscriber. So it mostly they will stay after after 3 or 6 months. But on the the handshake is between uh new acquisition is the this baby care function actually where we trying to give them a lot of good stuff. So at least you stay first. And then after that you can go into the base management.
Egidijus: [00:13:17] Yeah. And, uh, is, is this baby care period handled, uh, by your function or is it actually handled by acquisition?
Tommy: [00:13:27] This one is best. Uh, this one is part of CVM, but, uh. They specifically handling this, uh, new acquisition. So the acquisition function is, is, uh, is in sales. Uh, so, so they just trying to get as many as, uh, acquisition they do. And then our, our, uh, CVM, CVM function is trying to hold all, uh, most of the most of the new acquisition, which is sometimes is tough.
Egidijus: [00:14:00] So, so you are basically responsible for, uh, sales team, uh, KPIs, you know?
Tommy: [00:14:08] Yeah, yeah, it’s what I can say every time we see that the it’s the survivability is very low. What are doing then. Yeah. Yeah.
Egidijus: [00:14:24] Yeah. Uh, well, um, okay. Now, um, during our, our initial discussion, you mentioned that, uh, uh, now you feel pretty confident that, uh, your CVM activities are in, uh, pretty advanced state in the market. Yeah, correct. Could you, uh, could you describe, uh, kind of your current state of the CVM? Uh, how do you see it?
Tommy: [00:14:52] Yeah. Yeah. Uh. Before, uh, our CVM is basically, uh, beds, uh, beds campaign. So we start our journey in the 2020 where we move our beds campaign into the real time, uh, real time engine, uh, together with our partner, which I think at that time is quite challenging, I think because any, any migration of the system will be, uh, it will be tough. So. But after a successful, uh, successful, uh, migration, then we see this one is, uh, uh, start to take off. So. We start, uh, with two channels at the time, and then along the way at, uh, after, I think four years, I think we are almost everywhere, uh, connecting with, uh, 20, almost 24, uh, partners. Or the channel? Uh, yeah, it’s it’s quite big. And then we even even I think, uh. Tied up with, uh, online shopping. And also the latest is banking. We, we keep, uh, expanding our CVM to make sure that our CVM is over is being everywhere. And then, uh, yeah, it’s in the all channel that so basically our customer will find that. Okay, I can get the good offers by entering whatever channel that we have. So. Yeah. Uh, that our state is currently is. Yeah. At that, at that, at that big actually. Yeah. I think from the two channel now we are already at 23 or 24 channel.
Egidijus: [00:16:35] Oh, my God, this, uh, this sounds like an insane number of channels.
Tommy: [00:16:41] Actually, somehow is COVID-19 is somehow is, uh, accelerating this one in. Yeah. Because even even we work in the, uh, digital, so but somehow people become, uh, feel the need that to integrate. So at that time, I think we, we integrate a lot of, uh, a lot of channels, even though we are, uh, working from home. Yeah. That’s.
Egidijus: [00:17:05] Uh, but, uh, okay, so as we touched this part and, uh, let’s go deeper in that because, uh, for me, it’s pretty inspiring. How many different communication channels do you use? So, um, for example, um. Could you mention some of the. So you are saying like 24 different channels. So I know sms USSD yeah. Ivr okay. Uh telemarketing. Now what are the 20 more you know.
Tommy: [00:17:46] Yeah, yeah. So uh, we have uh. Uh, our apps. Telco apps. That is, uh, that is also our biggest, uh, revenue contributor. So and then we have, uh, Google and then we have a meta product, actually Facebook. And then we have a lot of partner of, uh, e-commerce. Uh, I think we have. Think or seven, uh, over there that is, uh, connecting to us. So whenever their subscriber. Typing their numbers. It will be, uh, trying to get what kind of offer it’s being asked first to to us. And then a traditional catalog. Okay, this is the kind of, uh, catalog. All you you just choose whatever you want and then switch into this kind of, uh, personalized. So I think that one is, is, uh, acceptance is very good. And then we also, uh, we have also banking, uh, banking channel. We, we just integrate. We have also, uh, positive SSD. Uh, so it’s not just a, um, also we, we make. Some certain, uh, triggers when whenever you hit the trigger, then this use SSD will will come into you. And then. Yeah, I think, uh. That’s that’s pretty pretty much, I think. I think pretty much so, yeah.
Egidijus: [00:19:23] Okay. Can we dig a one level deeper into, uh, every of those because, uh, it’s like, um, uh, for me, I see at least 3 or 4 channels that are not so commonly used. So first is, uh e-shop. Yeah. So you mentioned there that you have kind of 8 or 10 different partners there. So could you describe, uh, what kind of partners are those and how how does this kind of how is the user experience, uh, happening for the consumer so that we would understand it better?
Tommy: [00:20:01] Yeah, yeah. Oh, before I answer that one, I think I forgot also to mention that we also, uh, expanding to retail. So I think which kind of new actually. So yeah. Uh, and then going back into your question, it’s a lot of, uh, e-commerce actually in Indonesia, it’s just like, uh, Amazon, Amazon, uh, of the US, but this one is we have a lot of, uh, e-commerce actually, uh, that is selling, uh, data pack, actually, even, uh, I don’t know whether, you know, uh, some, uh, transportation like Uber of Indonesia is like a Gojek and grab. Uh, they also selling, uh, data pack actually over there. So what we’re trying to tap is actually, uh. In that side where they are selling a data pack. We integrate at that part when they put their number. So basically it will be just creating a separate box, uh, where it will only give you 3 or 4 offers. So you don’t need uh, to confuse actually, uh, what is uh, kind of offer that I need to choose because that four is already curated. So. Basically that one is based on your personal, personalized, uh, spending and then, uh, all the historical. So it’s supposed to be, uh, matching with whatever you, you need. So that’s why while at the at the same time, we’re trying to upselling them, uh, with a good offer over there with the they are they being not physically.
Egidijus: [00:21:26] Uh, okay. So, uh, I will, uh, will reiterate if I understand correctly. So, for example, in Indonesia, you have a grab, which is kind of you can get a ride sharing, you can purchase anything basically there similar as Amazon. And one of the things is that they actually sell top ups and uh, packages for uh, for uh for the telco. Yeah. Correct. Correct. And here you personalize that okay.
Tommy: [00:21:55] They personalize that one.
Egidijus: [00:21:56] Okay. That’s a very interesting, uh, use case. Uh, really not so common. But uh, for user it should be a really good experience, you know, because you only see what, uh, is relevant for you.
Tommy: [00:22:08] Yeah, yeah. And also the, uh, that partner is giving a good discount, so sometimes the offer is already good. And then the partner itself is giving a cash back. They say. So I think it’s, uh, a good combination over there, actually.
Egidijus: [00:22:22] Okay. Then, uh, other very interesting channel is banking. You wrote that you launched a new product, uh, with, uh, local bank. So could you, uh, tell a bit more the experience there?
Tommy: [00:22:37] Okay. A banking is actually the hardest channel that we try to integrate because, you know, Bing, Bing is, uh, having a lot of, uh. Yeah, they have their own rules that you must, uh, cannot give this, uh, you cannot, uh. The integration process must follow their own development time and all those things. So basically. After trying to approach them two years and then convince them that, okay, it’s better you use the CPA because people is actually migrating to this kind of channel because they see this one is giving you a better offer. And then then after that, one bank is okay, let’s try to integrate. And then that’s what what we have I think on the January uh, or February, then this one is get, get launched. And then after that we see another bank is trying to okay, let’s do it then. Yeah. I think once you get one and it’s, it’s uh, the other start to, to coming in. But what we see is actually the implementation is quite, uh, quite the same. It’s just basically you just need, uh, create, uh, 2 or 3, uh, relevant offers and then asking the bank team also to, okay, a bit differentiate the how it looks and then put it on top and then yeah, I just give some, some level, uh, like, okay, this is the exclusive offer for you or what promo for you or something like that. Then basically people is catching the box and then they see the attention into that box, and then they, they kind of, uh, start to moving, migrating into that one, because I think that one is giving better value actually for the offers. So yeah, that’s that’s how we implement on the bank.
Egidijus: [00:24:24] Uh, so you say two years? Yeah. Project.
Tommy: [00:24:29] It’s tough. It is, I think the longest. Are we trying to. Yeah. We constantly meet them one month and then they say, no, you miss our development time. Okay. We may. Okay. We we will, uh, wait for the next development time, which is when and. Okay, before that, we try to come, uh.
Egidijus: [00:24:51] How do you calculate calculate the business case for that? You know, it’s like.
Tommy: [00:24:58] Yeah, it’s just, uh, I think while we developing the big one, I think we. Okay, let’s let’s circle, let’s follow whatever they want. And at least we landed one first. And then just like we expected, once one is, uh, falling, then basically, uh, the other will will will continue.
Egidijus: [00:25:19] Uh, then let’s move to the methods. Uh, product that you are using and, uh, the channel. How, uh, um, how do you use that? Because for me, this is a totally unknown territory.
Tommy: [00:25:35] Uh, can you can you repeat again? Sorry.
Egidijus: [00:25:36] Uh, you said that, uh, uh, one of your channels is, uh, metas, uh, specific product, and then Google’s. Yeah. So.
Tommy: [00:25:45] Yeah, yeah.
Egidijus: [00:25:46] You describe, uh, uh. What’s that?
Tommy: [00:25:50] Yeah. On the meta is basically within Facebook itself. Uh, you will found that, uh, you can, uh, connect with the, uh, with the. Uh, with the operator, uh, where we can sell, uh, data packages over there. So, uh. Using that one, actually. We start to, uh, integrate our CVM because I think after we, uh, after we, uh, integrate with the ETL product. We start to integrate with the CVM one where we. Again, this mechanism is the same. It’s basically, uh, API, API based. And then. We ask also the same box where, uh, where it can show the CVM offers over there. It’s like we ask more. I think on that side, I think we we get 5 or 6, uh, offering over there. So that’s where, uh, CVM is get into, uh, meta product. And then for, uh, Google, it’s basically in the Android phone. There is a Google, uh, they call it a mobile data plan. You can browse it in the settings, uh, basically over there. Then you will see, uh, a lot of, uh. Uh, feature. Uh, when you integrate with operator, you can see what is the quote, uh, over there and then what kind of, uh, package that you can buy. So yes, another one is, uh, this one is just another, uh, integration with our API. So, uh, where once that one is established is basically, uh, CVM over is, is is open into the for the all the subscriber.
Egidijus: [00:27:41] Hmm. So, Tommy, it seems that, uh, your mission is to hunt subscribers everywhere. It’s like.
Tommy: [00:27:50] Yeah, it’s our it’s our mission. We we want to be everywhere. And then, uh, ours is supposed to be contextual, and then we must be a personalized one. So I think that’s okay. Our mission.
Egidijus: [00:28:05] And now if you, uh, kind of, uh. So you said that you started your journey in 2020. Uh, yeah. And, uh, yeah.
Tommy: [00:28:17] Correct. Correct.
Egidijus: [00:28:17] At that time, do you understand correctly that your primary channels were like USSD, SMS and, uh, uh, your own app? Something like that. Uh, maybe point of sales, um, customer service. So these channels and during those years you transitioned to 24 different channels. So how is number of, uh, kind of purchases or revenue part splitted between the channels that you own as a telco and the channels that you introduced? Is it kind of, uh, well, maybe not exact numbers, but uh, just to understand, is it like a significant part, is it, uh, a lowish part, how to understand, uh, uh, how to understand the business case of basically going everywhere.
Tommy: [00:29:14] Yeah. Uh, so, uh. Basically uh, our although our, we have, uh, 24 channels integrated, I think. But the top one, uh, the top uh, revenue is coming basically coming from the still, um. Um, and then also our own apps. That one is the two top, top in the, uh, top revenue contributor. And then. It follows with uh, by retail, uh, where we uh, also part of our channel and then uh, online and then uh, followed by, uh, we also, uh, I forget to mention there is, uh, WhatsApp. So basically the top, top 4 or 5 is, is, is the channel the the rest I think is just, uh, small, small one, but it’s there actually. Um.
Egidijus: [00:30:11] So they are still, uh, basically growing and, um, kind of being adopted. Yeah.
Tommy: [00:30:18] Yes. Correct.
Egidijus: [00:30:19] Correct. But, uh, from the customer experience point of view, I think this is kind of a true omnichannel experience. Yeah. It’s like, yeah.
Tommy: [00:30:27] Yeah, wherever you go, uh, you will find find us actually.
Egidijus: [00:30:33] And, uh, if, if we move, uh, back. So, so you have this really broad channel adoption strategy, and you said that you I want that, uh, you know, uh, offers would be contextual for every customer. Could you dig deeper, uh, on this part, what is your definition of being a contextual, you know.
Tommy: [00:31:00] Yeah, yeah. Uh. We, we try to to tap in into the right moment of the. When subscriber is trying to purchasing a pack. So at the beginning, uh, we do a lot of SMS broadcast, right? I think I think that is common to all CVMs sending the SMS broadcast, where at that time is we just send, uh, at 9 a.m. and then all the subscriber will receive a you buy this one, you can get this one. And but then after that we try to move into the. Contextual by. Tapping in into what kind of event that, uh, currently they they are, they are having. So what? What? Event that actually we see it’s quite successful. It’s whenever the subscriber is doing the, uh, reload or recharge. Uh, then we didn’t. Second, we sent the UST post to to them, and then where they can see all the offices over there, and then that’s where we see the acceptance. We start to grow, uh, quite big on the CVM side. Then people start to looking again when they miss the USD. Then where I can find this one, then, uh, basically with that one, our journey to the contextual is, is, is begin with that one. Because we see when you tap in into the right moment, uh, then basically they, they, they will, they will, they will take it. So then we go into, okay, what what if, uh, your quota is almost over, so. And we send you a good offer. And then at that time it also, uh, give a good, uh, increase. And then we see again whenever, uh, your quota is almost expired, will it, uh, give a good one. So basically then, yeah, we see, uh, we have been trying to tap into a lot of events, actually, which will make the offers become contextual and then without disturbing them because I think SMS is, uh, yeah, it’s SMS is quite. Intrusive and then people not reading it anymore, I think in Indonesia. So that’s why we we need a new and then a. Uh, innovative way to. Yeah. To connect with our subscriber.
Egidijus: [00:33:22] Okay. So you basically kind of, uh, uh, gotcha. And then, uh, so you define the context within to deliver, uh, the message and you link it to some specific customer event? Basically, yes. To make it, uh, contextually relevant. Now, um, you still need to deliver, uh, the good offer itself, you know. Yes yes yes yes, yes. Yeah.
Tommy: [00:33:49] Even even, uh, you know, uh, you know, even even we don’t send any SMS broadcast anymore to our supplier. We stop, uh, doing that. So because we know that subscriber will. Trying to find the offers by themselves because we are already exist almost in all channels. So I think we are at the stage where the subscriber is actually know how to find the offers. Then sometimes we, we, we experimenting with, okay, let’s give good offer in this channel and then what will happen. So basically people is just migrating into that one. So that’s kind of thing is interesting. Things like that is happening.
Egidijus: [00:34:30] But I, you know, it’s uh, it’s uh, it’s uh, really good that you have, uh, really engaged, uh, people, you know, it’s like, okay, if you if you want to move your customers to your own channel, like to mobile app, and you know that they will come, you know.
Tommy: [00:34:46] Yeah, yeah, yeah, yeah. I think we already at that stage, actually.
Egidijus: [00:34:51] So, uh, this is, uh, like a superpower.
Tommy: [00:34:56] Yeah. Yeah, yeah, we can direct. Where, which channel is, uh, where they need to go and then what kind of channel that we want to promote this month. So yeah, we have that kind of, uh, power.
Egidijus: [00:35:06] And, uh, do channel managers bribe you, you know, to, uh, so.
Tommy: [00:35:11] They, they come, they come some sometime when they see why my channel is dropping down. What are you doing? So, okay, we need to push this one. So that’s why we kind of steal it from you a bit. And then we will come back. We’ll give it back to you later.
Egidijus: [00:35:31] Yeah. Well, uh, when you, uh, select the offers, you know, uh, do you build some, I don’t know, next best offer algorithms or do you kind of, uh, how do you solve this personalization part?
Tommy: [00:35:48] Yeah, that’s, uh, I think, uh, the next stage that we are trying to, to do the where, uh, basically we’ve gone in through this rule based things, uh, where. Uh. We come with a lot of model at the at the back end. And this one is actually start to, uh, creating, uh. Algorithm, uh, makes actually a lot of brands of this one. And then, uh, like, for example, okay, you are going to churn or not. Okay. And then what kind of, uh, state next? What are you buying a monthly or daily? Then it becomes a lot of, uh, a lot of brands on the, on the offering, which. If we try to do it, uh, more detailed. And we found that. It’s too complex to our operation. And also the engine itself is become. We see, uh, it’s sometimes become, uh. Stop at that time. So. Then we okay, let’s not going into too many brands. So let’s just pick 2 or 3 uh logical uh or logical uh offers to, to, to them and then. But what we see is. Okay, uh, sometimes we are tempted to do more, actually.
Tommy: [00:37:21] Okay. What if we see we add this kind of branch into this one, then, yeah, it will become big again. And then after it become too big again, then we. Okay, let’s remove, uh, reduce it again. So until now, uh, so we are thinking, okay, I think this time we are implementing a ncvo here, uh, where? Let the machine is, uh, thinking what kind of offer they should see. And then based on the whatever trend data that are coming into the into the the system itself, and then they decide what kind of, uh, let’s say we choose six, let the machine, uh, decide what kind of, uh, offer they, they need to see. So, yeah, we, we are developing into that one. It’s very, very, uh, early stage. We we we. It’s not even like we are planning to make it live, I think this month, but I think we are going into that direction because we see it’s become too messy because we put all the algorithm based and based on the rule based.
Egidijus: [00:38:23] Um, so, um, uh, rules, uh, when you kind of, uh, define, uh, let’s say the way you, uh, select the offers, uh, rule based algorithms, they have a lot of strengths. You know, it’s like you as Tommy, you can understand them, you can explain for yourself to, you know, to the team. Why does it, uh, kind of. Why did the user got something, you know, and. Correct. Correct. And, uh, what are the challenges, uh, when you move from, you know, the full control towards, uh, you know, somebody else decided and, uh, and for example, your machine learning algorithm, uh, decided who will get the bonus, uh, which channel will get the bonus at this time? You know, how how do you deal with those challenges?
Tommy: [00:39:22] We we we think about that one. Okay. So we are we are asking, okay, once we move into this enveo then what we are going to do to the to the. Yeah, to our partner. So and then how we will monitor this thing. Uh, because sometimes our, uh, we have a question, uh, from our uh. Sea level or higher ups is usually the one day they ask why I see this kind of offer. Why then at that time, like you say, a rule base we can understand, or because you are falling into this segment and this segment and this segment. So we can we can we can clearly ask, uh, answering their question. But I understand with this nvo, we will kind of, uh, don’t know. So that’s why on the our partner, we are asking, uh, we ask them, okay, you should give us a function where when we put this number and then we you should tell us, uh, like in the, uh, I mean, in the, like, in the e-commerce or in the Facebook or Google, right. It usually they telling you why you, you are getting this offer because you are uh, blah, blah, blah, blah, blah, like this. Like it’s not fully detailed, but we are asking that kind of the same functionality. So whenever the question is coming, we can tell them, uh, directly without being, uh, like losing control because I think once they ask, we don’t know how to answer, then they will start the question a lot. So yeah, that kind of thing. I think we, we trying to must have must have before we, uh, launching this one. It’s supposed to be like Google. Google is uh, you are receiving uh, you are seeing these ads because of this one. This one. Right. Something like that. So we are trying to follow that one.
Egidijus: [00:41:12] Uh, well, good luck, you know.
Tommy: [00:41:17] Okay, now I see, I see, yeah, yeah. Let’s see. Uh, okay.
Egidijus: [00:41:24] Uh, could could could we drill into some of the use cases when you, uh, kind of, uh, so you implemented, added the context to the user, added personalized, uh, offer, uh, using, uh, the business rules. And, uh, which of the use cases do you see that, uh, kind of, uh, actually worked immediately, you know, worked like magic and said, uh, you, as Tommy said, for the team, we need to implement this. And you will see it will work. And and it did, you know?
Tommy: [00:41:56] Yeah. Uh, I think for the use case, I think, uh. It’s. Uh. Not. I think at that our stage right now, I think, uh, it’s easy, uh, we can play with the offers is quite, quite, uh. But first and then see the result, because our base itself is already, uh, very understand how uh, very understand with the this this offers we have uh, we brand ourselves uh, offering with hot promo. So whenever a subscriber is seeing that one, they they kind of understand this is a personalized one. So. What we see is whenever, uh. Whenever we we, uh. Well, how I say the, uh. Whenever there is a changes on the offers, it’s basically they they, uh. They kind of know, uh, directly whether we are. We are. Make it, uh. Make it. Uh, this one is cheaper or or make it, uh. Make it. This thing is, uh. Uh, more expensive because of the business needs. So basically what we see is when when you do, the offer changes, uh, they, they, they kind of, uh. Start to. Yeah, start to taking it if that one is, uh, beneficial for, for them and when, when that that that side I think we do a lot of, uh, trying to do a flash sale also and then, uh, yeah. Special offer all, all, all for you. All all of those things actually is, is is working well actually.
Egidijus: [00:43:47] Hmm. So everything where you give extra value to subscriber, uh, it basically works. Yeah.
Tommy: [00:43:54] Yeah yeah, yeah, yeah. That’s. So that’s why we need to, uh, calculate the uplift. So it’s not just giving, uh, more, more, uh, to the subscriber.
Egidijus: [00:44:06] So, uh, you’ve been going through this, uh, transformational journey the, the, the last four years. Uh, what was your biggest challenges? I would say going, you know, because you said that when you join the company you started, the kind of you changed the tool toolset, and you started kind of new, uh, new strategy for implementation. So what was your key challenges there?
Tommy: [00:44:35] Uh. I think when you trying to do this kind of changes. Uh. You will, uh, especially in the ETL driven company. I think they will, um. Start to see that, uh. The revenue is flowing into into the CVM side. I think that’s, that sometimes is, uh, causing uh, kind of, uh. A bit. Uh, question. Actually, uh. So are we doing the right thing? Uh, because people is, uh, let’s say shifting from ATL to BTL one. And then with the kind of mindset that they have, uh, that CVM is basically, uh, just giving a discount. So that’s why it’s people is moving over there. And then, uh, you give cheap offers a kind of thing, I think. So the biggest challenge is basically trying to educate the, the partners, uh, even, uh, in the works that basically moving to CVM, I think is. The best, uh, option to keep the subscriber. And then while we still, uh, growing the revenue because, uh, once they move to CVM. So basically, the CPU uplift, what we see, uh, it’s better. It’s almost two, three times better. And then. And then uh, also they stay longer with us. So basically that is the challenge. But once you. And convince them with data. So because I think, uh, CVM is is more on the data, I think we all good at that one. So we show them the proof. So basically after that they okay they not, uh, checking again, uh, on the revenue, whether this one is, uh, ethical or BTL. So what they say is now is the total, and then they convince that whatever they subscriber is moving to CVM, then the result will be good. So I think that’s, that’s I think the. Yeah. Had, uh, I think the biggest challenge here if we want to implement the CVM. Mhm.
Egidijus: [00:46:50] So, uh, basically in industries where, uh, let’s say the markets are maturing and the acquisition is, is kind of slows down. Yeah. Now the challenge is to educate the organization how to do the transition from acquisition. Uh, and uh, uh, ETL campaign, uh, ETL campaigns to ETL campaigns. Yeah. And uh, if, uh, let’s say a colleague in other company would be in your situation, you know, what advice would you give for for that colleague to succeed, you know?
Tommy: [00:47:28] Yeah. He. It must be slowly. Uh, process. Slow process. It’s. You shouldn’t just. Directly want to go to 24 or how many channels you should get your basis understand first with the CVM and then. Only from them. Then you start to adding one channel into another, and then also one channel into another. Then. Your maturity, and also the understanding on how the offer is being given is also. Increasing along with the additional of the channel. So I think. It’s not like whenever you want to implement this kind of thing is, uh, one years and then all is supposed to be, uh, good. It’s, uh. I think it will take several years to build this kind of, uh, state. But once you get into that one, I think you will see that there’s a. The good, uh, benefit to the company where you can. Give the offers without. You need to spend a lot of, uh, banner into the ATL one. Uh, and then they kind of, uh, just waiting for it. And then even what we see is they kind of, uh. Im immune to the, uh, other, uh, other telco uh, offers actually. So I think yeah, it’s it’s kind of you need to take it slowly on this one.
Egidijus: [00:49:08] So it seems like it was a tough journey with a very nice outcome. You know, uh, what is your future perspective? You know, what is your next big dream? Uh, Tommy, from the customer value management perspective. Yeah.
Tommy: [00:49:25] Yeah, I think from from our side, actually, that’s the kind of thing, uh. We are trying to do is we we trying to move to the AI or the machine learning, uh, things where we kind of, uh, invest uh, as an organization itself. Our company is going towards that one. So. We also trying to going into that direction also. So from like I say, we are trying to move from rule based to this NBA and then how it will gonna be. We don’t know yet, but I think we believe the AI will be the. Future, uh, development for the CVM. Because as a company, I think the implementation of. I that is directly, uh, connect to the customer is through CVM, right? I think because that’s where where the subscribers actually seeing it. Oh, now, now my offer is, is even even more, uh, relevant for for us while at the back end. Actually, this one is a I is, uh. Doing the doing the offers with then. Yeah. Then we hope with that one. Of course. Uh, the revenue still growing and still control, like you say, I think. And then we also can, uh. Yeah. Let’s see. We don’t know. Actually, it will be how, how it’ll, how will it will be on the NBA things here. But I think we are going into that direction.
Egidijus: [00:50:59] Uh, okay. So it’s like uh uh uh, I directly to consumer. Yeah. Uh, like.
Tommy: [00:51:05] Yeah, yeah.
Egidijus: [00:51:08] And to wrap up our conversation, Tommy, we always ask our guests to share their one their, well, several things. Uh, thing number one is what was your proudest career moment?
Tommy: [00:51:31] Yeah, I think, for me, I think. Uh, for the progress of the in this CVM field is actually. Is transforming that, uh, that CVM into the major, significant, uh, revenue contributor from very small, I think, to, uh, I think our revenue contributor is contribution is, is become very significant, as maybe you can imagine. Our, uh, one month revenue in the previous four years is is basically just now is only one hour revenue at midnight. Even at midnight. So.
Egidijus: [00:52:22] Wow.
Tommy: [00:52:23] I think, I think, I think I think that’s kind of contribution is that big. So I think, yeah, it’s transforming from that one into the current state. I think we are very proud of, uh, even the whole team, I think is very proud with that, that that kind of moment even we try at that time, we. We capture? Uh, okay. This is our, uh, full day, full day revenue. And then this is our one hour revenue. So you can see even at midnight, people is actually sleeping.
Egidijus: [00:52:56] So I hope Tommy, you you got a really good bonus for for this achievement, you know?
Tommy: [00:53:02] Yeah, yeah, yeah, I cannot say it. Uh.
Egidijus: [00:53:09] Together with great achievements, I think comes some lessons from, uh, let’s say kind of, uh, uh, failure parts. Yeah. Could, could you share some of your, uh, uh, lessons as well?
Tommy: [00:53:26] The failure one. Yeah yeah yeah.
Egidijus: [00:53:28] Yeah.
Tommy: [00:53:30] Okay, okay. Okay. Now. Yeah. You know, uh. This CVM is basically is really rely on the data feeding. Right. So all the segmentation and then uh, all the offers being given is actually is rely on this kind of, uh, things. So if that things got stuck then basically all the subscriber will see a very cheap offer no matter what the segment they, they fall into. That happens once. It’s makes my heart is quite oh my God, how can we do this here. So because. Yeah. We one day we we found our avenue is become very, very, very high at that time. And then yeah we found out that okay there is something is not happening on the back end side is not being flown. So that’s why people is suddenly is buying all these all these things. Yeah. Then. But that one after that we recover quite, quite fast and then. Yeah, yeah. Then we set up a mechanism where this thing will never, ever, ever happen again. And that’s kind of a I think. Yeah. If we think about that one, I think it’s oh everyone is panicking. Yeah. So yeah.
Egidijus: [00:54:53] But we uh, so I’m uh, I assume that, uh, this revenue loss was deducted from your, uh, from your bonus. Yes.
Tommy: [00:55:05] But, uh, we see that it’s a good sign. Also, people is actually coming back.
Egidijus: [00:55:13] So this was kind of, uh, one of the best campaigns ever made. Yeah. Uh.
Tommy: [00:55:20] For one day in one month, we see is not really good.
Egidijus: [00:55:26] Uh, and, uh, for people who who wants to grow the career as, uh, you did Tommy. Moving through different, uh, ladders. Uh, could you share, uh, let’s say, your personal source of inspiration and learning.
Tommy: [00:55:45] Yeah, this one is in CVM is quite tough at the beginning when you start this thing because I think a lot of. The the practitioner is, is not really, uh, having a one channel, uh, or one source of book. So it’s basically. You, you’ll learn. This one is, uh, from the. Data science, actually. So. And then, uh, there is no single what I see. Uh. One source that is really. Talking about this one until I think Exacaster, I think this is one of, uh, good, uh, good channel where we’re actually sharing this, but I follow this one is quite I think I, I watched some of the stories, I think, and then we kind of get learning, but before this one is actually my source of, uh, information is basically just, uh, reading some, uh, data science book and then trying to imagine this one or this one can be implemented actually in, in, in our organization. And then more of the. Yeah, I think I think more of the that kind of book. Yeah. That data science and also. Yeah. Also algorithm is is more on that one. No specific book of.
Egidijus: [00:57:10] Yeah the this makes sense. You know it’s like, uh, this uh function. It’s not very well structured around the world. You know.
Tommy: [00:57:21] It’s actually a mix of, uh, many, uh, function, actually. Yes. So it’s not only one thing is marketing only because this one is, like I say, it’s mixed with the data science. So it’s actually interesting one.
Egidijus: [00:57:37] So thank you, Tommy, for sharing your experience today. It was very interesting to hear your, you know, insights and learnings and best campaigns ever.
Tommy: [00:57:49] Uh thank you. I hope that one will be, uh, can give a good insight also to other CVM practitioners in other other countries. I think I know, I think this one is. Early may be still, uh, small community, I think. But along the way, once, once the AI coming and then also the data science is become more, more common, I think that’s the CVM is the way, I think.
Exacaster: [00:58:22] Thank you for listening to CVM Stories. If you enjoyed the show, please leave us a review. You can also ask us a question about a particular customer value management challenge you have at work. We will happily ask our experts to tackle your challenge in a future episode.
#CustomerValueManagement #TelecomInnovations #CVMStrategy #DigitalMarketing #Personalization #MultiChannelIntegration
