Good morning again. Thank you for joining us to the Analysts and Investor Day 2018. This is the formal part of the AI Day. Previously, as you know, this morning we had the sort of investors' conference call with regards to our third quarter result, and this is the beginning of the actual introduction of the Analyst Day. Okay, so the backdrop of today's Analyst Day is perhaps quite mixed, I presume. As you know, our share prices from the beginning of the year have gone down almost to the tune of 40%, unfortunately, which is very sad for all of us in this room, especially for us from Axiata. Despite the fact that if you look at our core results, the revenue has indeed, constant currency has increased, indeed increased by 5.2%. The EBITDA, even including the Celcom VSS or ELP, has increased by 2.4%.
Now, obviously, the profit has gone down by, on a normalized basis, 25%, but the reasons that have been discussed before, including the LP and the depreciation, but if you look at the actual performance of opco by opco, we have a pretty good year, actually. We have six opcos where we have controlling shareholders. That's not the norm, yeah? Sorry? What have I done? Okay, so if you look at the performance of all the six companies, so for the nine months 2018, all of them performed in terms of revenue the best in their respective markets. All the six of them. Now, in terms of EBITDA, five out of six of them performed better than the market. In terms of profit, half of them performed better than the market.
And in the case of revenue gain, all the six of them, not just performing better than the market, performing the best in their respective markets. And for EBITDA, five out of six performed the best in their respective markets. So somehow it doesn't jive with the market price. Let's talk about it. So before I do that, I'm going to talk about, in a bit of history, as I always do, to talk about how do we compare our performance the last 10 years. There's a bit of background here. Sorry, is it because of me? Oh, Tunia. Okay. Now, what I'm going to do to you is to show our performance over the last 10 years, revenue, EBITDA, and profit for all the operating companies and give you a perspective of our performance. Now, obviously, it's very hard to compare because everyone has different kind of assumptions.
These are all statutory numbers, so there are a lot of one-offs here and there, acquisitions, and a few other things that kind of distort the numbers. So it's not meant to be highly 100% accurate because of our ability to compare with our competitors year by year because of all those variations and numbers. But just get a feel of it so you can see there are messages within those numbers. Now, if you look at our performance for the 10-year, we have been a top performer in revenue and market share growth in all of our markets, except Celcom, which performed extremely well all the way to 2013 and also year to date 2018. As you can see, they took a short break the last few years.
In terms of profitability, except Celcom, which was a top performer in 2013 and excelled second best, the rest performed the best for the last 10 years and almost year by year. Now, this is a chart that shows over the last 10 years our market share growth. Celcom, not much, 1%. XL, about 3%. Robi, 10+% , and Dialog, 11%. And not shown here is Smart because we don't have official numbers from our competitors, but even that has grown significantly over the last 10 years. Now, that's over the last 10 years. But let me share with you, again, over the last 10 years, but look at revenue, EBITDA, and PATAMI, and see how we compare. Now, if you look at revenue, Celcom, second best in the industry over the last 10 years. In Indonesia, actually the best over the last 10 years.
Robi, by far, the best over the last 10 years. And Dialog, again, dominated the last 10 years. And if you have a smart number, based on unofficial numbers, by far, the best against all competitors. Not against market average, against all competitors. So, as you can see, EBITDA, not exactly the same story. Celcom, not the best in the industry for sure. XL, actually marginally better than other competitors. If you look at, in the case of Bangladesh, in Robi in Bangladesh, second best. And again, Dialog and Smart, by far, by far the best against our competitors. Now, PATAMI is not necessarily the best picture too. Celcom, not the best in the industry. XL, second best. Robi, the best. Dialog, the best. And Smart, the best. So that led to the conclusion again before. But let's look at year by year.
It's interesting to see that, again, you can see from a revenue perspective, Celcom performed the best in the market until 2013. Then we had a disastrous three years. We woke up again 2017 to be second best and the best again in 2018. Excel, kind of similar story. I don't know why they talk to each other, but when one gets bad, the other decides to do the same thing. But fortunately, Excel recovered very well the last two years. Roby started second best from then on, almost always the best in the industry year by year. Dialog, after two years during the civil war, after that, just dominated the industry, and Excel, again, the best, and Smart, the best in the industry. You can see a pretty good picture of our performance from a revenue perspective. From EBITDA, similar, not exactly, but similar story.
Pretty good in most of the opcos over the last 10 years. Now, PATAMI, not very good. Now, therefore, what is happening below EBITDA is our concern, right? We are the best in revenue and arguably the best in EBITDA, but not PATAMI. So one conclusion that we have when we really study in a lot more detail than the chart that I show you is that our CapEx management perhaps is not the best in the industry. We have to do a lot—we have a lot of work to do in that area. So, hence, when you look at our focus on CapEx has increased. Although we increased our CapEx intensity 2016 and 2017, and therefore the moratorium on our dividend, well, we decided that we will, from now on, improve our CapEx intensity on the core mobile.
I repeat, on the core mobile, because our intention is to reduce that every year from now on, right? Bearing certain other circumstances, if there are opportunities, of course, we will increase again, but directionally, we want to reduce. And even despite the forex. FCF, of course, continued to be strong, but operating FCF has reduced this year because of interest hike and taxes. ROIC, unfortunately, below WACC in 2015 and 2016, but as you can see, it is improving 2017 and looks like it's improving 2018, and we hope to improve also 2019. Debt EBITDA continued to be very strong, except 2018, 2008, when we acquired Idea, and 2016, when we acquired Ncell. As you can see from this chart, remained pretty strong, below 2.5.
Finally, our dividend yield remained moderate, as planned, reduced in 2016 and 2017 by design, but to support the capital investment, and we hope to improve upon that as we go along from 2018 onwards. So that's the past. Okay, that's the past 10 years. Now, I want to reiterate. You have seen the next chart, last investors' conference, so I'm going to just bring it up just to put into context. No change in vision. We want to be a next-generation digital company. Now, what we have done is that we are now scrutinizing four areas of our businesses to really understand what does it really mean by digital champion.
We are scrutinizing the external touchpoints, exactly what are the touchpoints, how do we serve our customers directly and indirectly, and how do we serve the distributors, dealers, and our partners directly using the digital concepts and technologies. Similarly, from a products and services standpoint, we are now trying to even measure what are the products and services that derive from the digitization of the company. Obviously, internal processes is about cost. How do we reduce cost from digitizing our company, our processes, and our network and IT architecture? How do we modernize that to become more digital? And lastly, the culture and mindset of the company. Our strategy, no change from last year. Our proposition to the investors to provide, to give moderate growth and moderate dividend.
These are the three pillars, or we call it a triple-core strategy: digital telco, digital businesses, and infrastructure. Today, from a valuation perspective, this is just a guide. We are very much 85, 0, and 15, respectively. What we want to do is to move towards 75, 10, and 15, respectively, plus minus 5%. That's our vision by 2022. New growth areas in each of the respective pillars in the area of digital is about home, broadband, enterprise, and of course, hopefully we can consolidate and grow. In the area of digital businesses, fintech, edtech, digital platform, and IoT, enterprise IoT to be more precise. And of course, in the case of edotco, we have a lot of opportunities still to grow organically through higher tenancy ratio and new builds inorganically. So we are still very keen on consolidation and new footprint.
And last but not least, we presented to all of you the eight strategic initiatives. The most important one is the operational performance. We got to perform better and better and continue to perform. Hence, you have seen our 2017 numbers, the underlying operational performance, and you have seen the performance of the 2018 so far has been from good to very good. New growth areas, digitization, network superiority, industry restructuring. We have to say that, and I'm going to elaborate a bit later on, that unfortunately, operational performance is not going to be enough. We got to do more than that from an industry perspective and also from portfolio rebalancing perspective. So that's the eight initiatives that we have presented last year. This time around, we are not going to talk about it.
We're going to focus more on very specifics on what we're going to do given the external environment. So, okay, this is where I want to spend a bit more time. Bear with me for a while. All right. So the long-term strategy that I've shown earlier, the vision and the plan and strategic initiatives will not change. Generally speaking and directionally, will not change. However, there are eight specific changes that we're going to adapt given the external environment. All right? I'm going to talk about them, but let me give you a bit of context what we're trying to do. Now, we had talked about perfect storm. I don't know what is more perfect than a perfect storm. The global macro, national macro, local industry, and regulatory factors across the group has affected us quite significantly over the last few months.
Now, you all know what happened, what we call the three-day syndrome or three-day problem we had during a lot of things happened during the New Dawn and the Khazanah Megatrend Forum. A lot of things have been said that rattled the market. Unfortunately, not necessarily reflecting our fundamentals. It affected our share price for sure, as I said earlier. But to be fair also, a lot of things are affecting our financial performance. The forex itself has affected us. You saw the presentation by Vivek, whereby 9% impact on just forex translation. In other words, if we had grown 9% outside Malaysia, when we translate back to ringgit, we are back to square one at zero growth. The interest hikes in Indonesia, Bangladesh, and even corporate center has affected us. As you know, Indonesia, our own interest rate on average increased by 150 basis points.
Robi, 200 basis points, and corporate center, 75 basis points. That itself has increased our absolute cost and went down all the way to the bottom line of about $140 million. Those items in themselves are very, very painful. The tax impact that we mentioned primarily Ncell, where the increase of the direct and indirect, the 5% increase in corporate tax and the increase of the tax on the data and voice, especially data from 0%- 10%, although it was announced recently, so it has affected only $8 million so far, but for the full year, we're expecting $20 million. The forex, of course, has the biggest impact. The translation itself, $1.7 billion, if you're from an absolute standpoint. Again, those are non-cash, unfortunately. Mark-to-market, non-cash, $8 million. Transactional, those are the actual impact to our CapEx.
But again, some are mostly still unrealized loss, some are realized loss of MYR 160 million. So if you combine, just take away the non-P&L impact, although it's still non-cash, if you just take away the P&L impact, year to date has been to the tune of about MYR 300 million. That's quite a lot. MYR 300 million P&L impact, right? But fortunately, a huge chunk are not realized loss, non-cash. But still, it affected our P&L. In a way, that has again, apart from this overshadowing affecting our share price, those are unfortunately real, affecting our P&L. Although from a cash perspective, as I want to stress, it has zero impact.
The challenge for us is how do we bring up the short-term profit again, given all this tailwind, or rather headwinds, rather, headwinds that affected us and the funding cost that is material, no longer marginal, materially affecting us. As much as the overall direction of strategy remains unchanged, we have to relook at our implementation and priorities. The theme for this whole event is shifting gear. We have to face the reality of the next one year or even two years, all the macros that are affecting us that are unfortunately mostly maybe a matter of good field, but some affecting us directly in terms of profit. We are going to move at least the next one year, 2019, towards profit and cash. Profit and cash. Now, there are a lot of adjustments required.
We might have to refocus certain areas, reprioritize, relook at partnerships, portfolio, consolidation, balance sheet, and even our KPIs. The following chart will show the eight areas that will make major adjustments, some are significant adjustments. But bear in mind, I want to not leave you a message that while we are very focused on 2019 numbers, there might be a little bit of impact to the long term, but it will not to the extent that it will hurt our long term dramatically. Yeah? Let me be very sure again. As yet for the last many, many years, we are very much into long term. Our investment into all our goals are long term.
We have been into many of the operating companies to the tune of 20 plus years, and we are not rattled by quarter by quarter or even year by year to some extent, but we do recognize that some adjustment really required to show good numbers in 2019 with minimum some, but minimum impact to the long term. Yeah? I have to stress that. Okay, what are they? So the headline will be we are going to be more focused on profit and profit growth relatively. I want to underline that relatively more than revenue and market share growth for the first time. Now, you see the context of my earlier presentation. We have been number one in every market in terms of revenue and market share growth. Number one, not even average, better than average, better than everybody else in the whole 10 years, right?
Except for Celcom, but even Celcom for until 2013, for six years, they have been number one in their market, and we have been growing and growing and growing even 2017, even 2018. We will continue to grow revenue share and market growth, market share growth. I want to underline that, but if we have to balance between additional 1% or 0.1% market share growth and 1% profit, we will choose profit for next year, so we will lower a bit our growth revenue price expectation, but we will increase our expectation on profit. In our plans for next year for the core mobile, just for the core mobile, we expect every Opco to have their revenue growth, sorry, their EBITDA growth faster than revenue growth. That's the plan, right? Now, that's a plan.
You cannot call it aspiration, but that was the plan that we're going to embed in every OpCo for the core mobile. Now, obviously, when you see the total numbers, there are a couple of investments in home, broadband, enterprise that might change the numbers slightly, but not much, but for the core mobile, EBITDA growth must grow, EBITDA must grow faster than revenue in all OpCos and therefore, for the group, we want to make sure that the group EBITDA growing the group EBITDA grow faster than revenue. Just for you to give a feel, a lot of profit targets for the group and OpCos are very much based on that premise and we hope to improve our ROIC, right? So this is the first time that I have to say this that we're going to focus on short-term 2019 profit more than the revenue growth.
Normally, we want to do all of that with a long-term focus, but this is a temporary adjustment. Therefore, we will put heavy spotlight on CapEx and especially CapEx efficiency. Now, when I say that for the core mobile minus new investment in other areas, we hope to reduce the CapEx intensity slightly, but year by year. Now, that does not necessarily mean that we're going to dampen our future, but we hope to be a lot more efficient. We're going to find one, two, not three, or not five ways of being more efficient in our CapEx efficiency, our spectral efficiency, our rollout efficiency, our site operating cost efficiency, our people efficiency to ensure that when we implement CapEx, we get the best CapEx.
We are already, for example, deploying machine learning to do CapEx planning to supplement, of course, our people's capability to be more efficient in terms of planning and efficient CapEx. So that will be a major focus area. There will be a presentation on this today. We'll give you a lot more detail. Now, this can be misconstrued. I have to say this very carefully that we are looking at some of the investment where we can prioritize, rescope to make sure those, especially longer payback, we might not do as much as we have been doing in the past. We are still very much into home broadband, but relatively speaking, again, underline, we'll be more focused on fixed wireless rather than all fiber. So the shift will be more fixed wireless.
Enterprise segment, which again, I must emphasize, we will be focusing and investing in fixed enterprise segment and solutions, but we will moderate our investment in some of these areas to focus more on working with prime system integrators to minimize our investment. For digital, as you know, we have spent to date about $200 million in digital. And to date, our losses, or rather investment costs, is about $130 million. Last year, about $250 million. For next year, our plan will be a bit different. We will limit. We have a ring-fenced. Our limit of losses will be MYR 200 million. The plan for digital businesses will be based on that. So rather than from a top-down of what we can achieve all the way down to the impact of profit, we will ensure that it will not be more than MYR 200 million.
We ring-fence that. Not more than MYR 200 million losses. Now, of course, it can be even less. I'm not saying it will be exactly MYR 200 million, but that will be the limit of our investment. So going back in the past, if you remember, all our profit the last three, four years has been affected to the tune of MYR 200 million-MYR 300 million because of digital. We're going to limit that. But again, that does not mean that we're not going to still do anything about it to defocus on digital. Indeed, three of the businesses are doing extremely from well to extremely well. There'll be a presentation later on, but you will see later on. And we do not want to curtail too much. But at the same time, we're trying to balance between that and our profit for next year.
We will ramp up value-added services, which used to be almost double absolute wise from where we are today. Those are the content aggregation for all-time contents. We cannot defocus it. Unfortunately, the last one or two years, we're going to ramp it up. The beauty of this is that the incremental cost is very minimal. CapEx, very minimal, but the value is quite significant, right? So that will help in terms of our investment. The return is short-term. For the core mobile, we will, of course, remain positive in our investment, but we will moderate in certain areas. Customer experience is core to all of the OpCos, especially Celcom and XL and Robi and Dialog, for sure. We will moderate that to the point that we do not want to have the problem of low reducing or rather the problem of low marginal returns, right?
So we will look at that. And some of the investment that is long-term, we will KIV. We will remain nimble, though. If there are changes in the macro elements or we get alternative funding and support, we might just change what I've just said above. But until then, our focus on many of these projects will be reprioritized or rescoped. Now, on the positive side, if we can fund our investment in the new growth areas through partnership or financial investors directly or indirectly, we will, of course, take advantage of that. Not only that. Well, I have to stress here, we can fund many of these easily within our capability. Most of these, maybe not all, most of these, right? Then back to number three, we could have funded that.
The problem is the funding cost is so expensive at this point in time, especially given the interest hike, right? So if we can fund it through strategic partnership, a financial investor, and also moderate our risk and also reduce the impact to the P&L, which is the biggest point, not so much our funding or cost, so that we can reduce the impact to P&L, we will be working with them instead. And we are open to co-investment in home, broadband, enterprise, and especially in the digital businesses. And even to some extent, the core business. We will monetize our existing investment, our assets for cash and validation, right? Now, we have to be careful again. Some of my words might be misconstrued. We will not do those irresponsibly. We will not do it at wrong timing.
We will be still opportunistic, but we will balance between, again, long-term optimal situation versus maybe medium term, right? We are not in a rush in any of our investment. I have to be very clear with that. I have to repeat, we are not in a rush, but if we have great opportunity, we will do it, however, as soon as possible. Not just because of cash, obviously, it's important, but some of our investment, especially digital, we want to validate. Now, to some extent, when we brought in Sumitomo in our investment in ADA, our digital advertising, we kind of validated that it is not worth zero, but worth to the tune of $110 million. Small, but at least it has been almost given almost zero value by everyone. It does indeed being validated. We want to do more of that.
We want to validate all the digital businesses. We want to validate our investment in all the new growth areas to show that, yes, indeed, there are inherent intrinsic value of those investments, and of course, at the same time, we get cash. Now, this is tough. To some extent, my credibility has affected because two years ago, I was a lot more optimistic that by end of 2017, at least one would happen. Unfortunately, none happened, right, and I got one month and a half to make it happen this year. I still hope at least I can announce one, if not in January, to happen, but in any of this industry consolidation or network sharing. Now, this is very tough to be too forthcoming. I give a bit of hint here and there, but hopefully something will happen.
Maybe not substantial yet, but at least to show that we are very keen to accelerate this, and as I mentioned earlier, while our operating performance in all our markets has been from good to very good, not recognized by the market, unfortunately, but never mind, we are very much into fundamentals. The problem is in the long term, it's not enough. We got to do consolidation. We got to do network sharing in many of our markets. A big time rather than small. Last but not least, productivity improvement. We do recognize that maybe in some OpCos, we are not as productive as we should be. As already, we are looking at at least two companies where we want to do a major one, a major productivity improvement, quote unquote. In the case of Celcom, we can now mention publicly. We have done the first one.
It costs us to the tune of MYR 50-60 million, and the savings the following year in 2019 onwards is to the tune of MYR 50 million or so, year- by- year, each year. The second program was announced last Friday. It will cost us, again, I can't mention the cost because we don't know what's the acceptance. Maybe same, maybe slightly less, or maybe half of that, but hopefully the benefits will also be to the tune of MYR 0-50 million or so, depending on the take-up. But bare minimum, we should have some savings of MYR 50 million next year onwards. So that, to me, is structural from a people organization perspective. Now, not to mention the qualitative changes we are making across the whole group on how we reorganize, restructure the organization in almost every OpCo to make it more agile, to make it more digital.
We intend to relook at all our non-productive end-of-life assets, and if we can impair them, we will impair them. Of course, within the very strict rule and to be approved by the auditors, for sure. We won't do something which is wrong, so it will be done very much within that context. Now, the angle is that we have done a lot of modernization in our network and IT over the last two, three years. A lot. For example, we have moved a lot of our investment or rather transition into 3G and mostly 4G now in many of our operations. We have modernized the component network, the core network, IP core. We have modernized the base station. On IT, we have put more and more stuff in the cloud already.
We are going to sort of sunset some of the assets and many other things we want to do. As a result of which, we are looking at if we can impair some of the assets significantly, again, to benefit the following year. Now, I have to, again, be very careful. What I said is not because we want to artificially inflate next year, but because those are within the rules can be impaired. Last but not least, all the KPIs for Axiata Group, Axiata Management, and Opco Senior Management will reflect all the seven above, especially number one. Profit will be looked at even more. Our weightage has been sort of half-half between revenue and revenue-related measurements and profit and ROIC-related measurements.
But next year, it varies Opco by Opco, of course, but there will be a huge skew toward profit and cash so that we can implement and motivate our people to do all the seven. So these are the eight initiatives. Now, I do hope there are questions later on so that we can qualify and explain better some of these eight questions in case there are doubts because some of these read wrongly can be misconstrued easily, right? We are in no way wanting to jeopardize our long term, but yes, we want to skew it a bit to 2019 and 2020 profit and cash growth. With that, I'll end my presentation. Assalamualaikum and good morning again, everyone.
Jen and I are going to take the next 20 minutes just to give you an idea or give a feel about what are the things that we're doing within Celcom, what are some of the new priorities that we do, especially inside around the cost management, as well as how do we monetize our network better. Okay. We have a new CEO in Celcom because this is a new Celcom. It's not just, well, it's not just about we have this year a new government, but for Celcom, we also have a new HQ. So we're moving to a new HQ in Petaling Jaya. So this is Celcom. It's the picture of Celcom's new HQ. Well, it's not just a physical change for Celcom. We're also changing the way we work.
There's a lot of transformation going on internally as well as on top of that, they just got a new CEO as well. So Jen and I will take the next 20 minutes to just take you some of the changes that we're doing, some of the key focus areas that we're looking at. So I hope to cover these four areas today. How have we been doing in the past couple of years in terms of current strategy? How is it performing? I'll touch a bit in terms of what are the new operational priorities. For sure, I can't give you the details, but at least through this, you have an idea what are the areas that we want to focus on. We'll deep dive a little bit as much as we can on cost management and network monetization. I think this is something that's quite interesting for everyone.
We're looking at in the next three years, if we get to manage this right, we'll probably save on the run rate at about MYR 900 million over the three years. And then last but not least, to touch a little bit in terms of the new organization, we just restructured a little bit just to be a bit more focused on the organization. As Tan Sri Jamal mentioned earlier, Celcom has been performing very, very well until about 2013. We had a saying at that time is 31 quarters of growth, right? And then from 2013, 2014, 2015, until 2016, we decided to take a break. This is in terms of revenue, if you can see, and also in terms of PATAMI. But I must say, in the last two years, what the management has done is quite significant, where today we have stabilized in terms of the company.
We have stabilized the performance. I think now our revenue engine is working. If you look at on the right-hand side, in terms of service revenue from 2016 to 2018, we have increased in terms of our revenue market share among the top three by 1.5 percentage points. If you look at in terms of total revenue at the bottom right-hand side, you're talking about 1.4 percentage points in terms of revenue share. So we can say that over the past two years, in terms of revenue engine, this is the area where Tan Sri mentioned earlier that we've filled our distribution, our sales team, our product and services, and it's starting to show the results in the past two years.
Besides that, if you look at what is our strategy, our strategy was always to go after the high-value customers and how do we use customer experience, providing the best customer experience to go after the high-value customers. If you look at on the left-hand side of the slide, if you see how are we doing in the high-value customers segment, when we say high-value customers, we mean customers with MYR 80 and above RPU. You see the revenue has grown from 27 in the past 12 months from Q3 2017, a total revenue of MYR 592 million to about MYR 670 million. There's a growth of 13.3%. But you look at also at the bottom left-hand side in terms of RPU, we have grown the RPU from MYR 123 to MYR 130. So whatever that we're doing in terms of customer experience and focusing on the high-value segment is starting to show results.
The other initiatives that we're doing, if you look on the right-hand side, the bottom part, our digitization effort for those that have been using our app and for those that have gone to our website, we'll probably clearly see the development that we have done in the past two years. But more importantly, if you look at the number of transactions that we're seeing on our digital app has grown tremendously. And today, as of Q3, we are hitting about 850,000 transactions per quarter, so if you've seen this, the digital transaction is increasing significantly while the physical transaction over the counter, over the call center, all that is reducing, so this is still early days, but overall, in the longer term, we should see savings in terms of cost to serve to this customer.
And we're getting very, very focused in terms of increasing our unassisted channel to support our customers. So it's not just about cost also, it's also about giving the customer the best digital experience. If you look at the bottom right-hand corner, you'll see that even our customer impact in terms of our NPS, we have now, this is done by a third party. Over the past few quarters, we are now number one in terms of NPS in the market and with a gap of the next about 6 percentage points or 4 percentage points to the next best competition. So we're seeing some of this indication where our strategy that we said where we're going to do is currently working.
Besides this, Jen will talk a bit later what we have spent on the network and how our network, actually, the network experience has actually significantly improved than where we were about two years ago. Now, if I touch a little bit in terms of what are the priorities, key initiatives that we're looking at in 2019 onwards, I categorize it in four categories, in four areas. The first area is about winning in the market. I think this is something that we started to experience in the past couple of years, but this is an area we're going to continue to focus on. A couple of areas. Number one is we're evolving from traditional segmentation, for example, to a more granular and dynamic segmentation. We're going to be a lot more granular, a lot more precise in terms of our go-to market.
We have adopted and incorporated a lot of analytics inside our planning so that we can be more precise in terms of going after these high-value customers. The second area is about cost transformation and network monetization. We're taking this very, very seriously. Jen will talk a little bit more about this, what we're doing in terms of cost transformation. And when I say cost transformation, we're not talking about cost cutting. We're not talking about a 10% cut from the budget, but we're talking about really, really looking at all our cost structure in three key areas. Number one is the network area, network and IT, in sales and marketing as well, in people and productivity costs. We're also looking at in terms of how do we monetize our network better. Now we know exactly in the market where we're doing well, where we're not doing well, which cluster.
We go down to our 965 clusters around the country. And I'll show it to you a little bit later in terms of how are we doing to be very precise in terms of how we're going to target these markets. The third area is about executional and operational excellence. We're talking about prioritization. We're going to be a lot more focused on execution. Even if we talk about digitization, we're going to be very selective and focused on digitization. We're exploring a lot of robotics, artificial intelligence also inside the company. At this point in time, I can safely say that we have more than 20 different initiatives that are using analytics, that are using robotics, that are using artificial intelligence across the company. Another big area that we're looking at is also around regionalization and also empowerment to the region.
We know that the battlefront is actually in the region. That's where we go and hit the competition. And we're going to relook at how we operate in terms of how do we empower more to the region for speedier and better decision-making. Last but not least area, the fourth area is about rejuvenating the organization. Now, Celcom is a 30-year-old organization. I was saying jokingly, if you cut the hands of any of the Celcom employees, you'll see the blood is blue and the blood type is C, right? So it is an organization that has a lot of history and legacy. So this is, even though it's number four, but I think it's one of the most important initiatives that we're doing is how do we rejuvenate the organization.
A lot more initiatives around cross-functional, around solving problems, and also around adopting how to work as an agile organization. A lot of programs that we have put in place to be able to so that Celcom as an organization can transform itself to become what I would say a 30-year-old, 3,000-strong startup, right? This is an initiative that we're doing internally on people and organization. Next, I'm going to pass to Jen to talk a bit about our cost program.
Yeah, that's it. You know, in the last, in 2016, when we actually have dropped quite a lot, as what Tan Sri has actually said earlier, you know, that was hopefully the end of the worst time for us was mainly because we have also underspent a little bit in terms of CapEx for network.
Having said that, what has actually happened then in 2016 is that we have actually ramped up the investment for capital expenditure for us in 2016, 2017, and also in 2018. When that actually happens, the LTE population has also increased, and we are actually catching up in terms of the LTE population coverage with the competition. In 2016, we were actually lagging behind as compared to the competition. In 2016, we were even lagging behind compared to DG for the LTE coverage. In the last two years, what we have actually done is that not only we have actually improved in terms of the LTE population coverage to 90% in Q3 2018, where we now match the Maxis LTE coverage, and we have actually exceeded the LTE population coverage compared to DG. Having said that, LTE population coverage is not something that is our target.
Well, that is one of the things that we actually look out for. But on the other hand, what we actually look out for is the network experience for our customers. In the past, if let's say you refer to the OpenSignal report, it's always emphasizing in terms of download speed. But for us, I think the download speed is one of the elements that we actually look out for. But in essence, what really matters most for the customer is actually video score. And that the ministry and MCMC has already realized that the video score is actually more important. And hence, that's why the YouTube measurement was actually being considered by MCMC.
In the last report from MCMC, our rating for network quality is actually at par with our competition, Maxis, because most of the main area is either they are leading or we are actually leading in those key areas, then I think in the last two years, we have actually done quite a lot of catch-up in terms of CapEx, and our network experience has now been competitive and very matching to our next competition. So hence, that's why during the last two years, we have also spent quite a fair bit in terms of OpEx for network, because when we were doing catching up, the main thing that we need to do then is to make sure that our network quality is at par with the competition.
Now that we are actually at par, our next focus is actually to make sure that the network OpEx, which is on the right-hand side, will start to rationalize. In the last two years, we have actually ramped up the network OpEx to up to about close to 20% in 2016, and the number is starting to decrease as we go along. That is mainly driven by the cost optimization program that is actually ongoing. There are two ways for us to actually look at the cost optimization. One is the structural, the other one is the tactical. Currently, we are only seeing the impact of the tactical initiatives, and for the structural initiative, we should be able to see it in the longer run.
And hence, that's why our objective for our network cost is to actually trim down by close to 3%-4% in the next couple of years. Hopefully, in the next three years, the number will be much closer to the competition. Cost optimization, our EBITDA margin, I think we will be very open about it. Our EBITDA margin is the worst as compared to the competition, and we are very aware of that. Hence, that's why in the next three years, we have actually put a lot more, I mean, in this year at least, we actually put a lot more emphasis in terms of cost optimization. We are hoping that in the next three years, we will get a savings of accumulated savings of MYR 900 million in the next three years.
and that will result in the EBITDA margin to improve by 25% compared to the current level that we are actually experiencing. The three key areas that we are actually going to focus on is in terms of network and IT, which is the technical technology side of things. And then, of course, the sales and marketing, and also the engagement with customers. As what Idham mentioned just now, we will go a lot more in terms of digitization, the transaction that we actually have with the customers, and also the way that we actually engage with the customers. That hopefully we will be able to shed some cost in terms of sales and marketing and also the cost to serve the customers in the longer run. The third pillar that we will be looking at is in terms of the organization and also the simplification side of things.
In the last two years, what we have actually done is that we have delayered the organization chart. Once we have actually delayered that, the process is much more simplified now. So the first step of the simplification is already done when we actually reduce the number of layering in terms of organization structure. Currently, what we are actually looking at is to actually simplify the processes that we have so that when we actually move on to digitization, it's already a more streamlined process. So that's where we are actually heading to for the next three years. Hopefully, in the next three years, again, to recap, we are hoping to actually improve our EBITDA margin by about 25%. With that, I'll hand over to Idham.
Thanks, Jen.
Next area I'm going to talk about is just to give you a feel in terms of what do we plan to do in terms of how do we monetize our network better. So we have about 965 clusters throughout the country, and we have done a study, and we have looked at each of these clusters. Where are the clusters that we have a very good network, good customer experience we're providing to our customers, and which areas we have a strong market share and which we are not strong in terms of market share. We have looked at this into four separate segments. If you can see this, just as an example, one area that we call it, we have a very good network experience, but we still have a very low market share. We call it it's an attack opportunity. We identified about 360 plus clusters.
That actually is an attack opportunity for us where our market share today is still below 30%, but we are providing one of the best network experiences in the clusters. Similarly, in the areas where we are our stronghold, where we have a strong market share as well as a very good network, is another area where we can build on our strength and continue to optimize our network as the demand grows. And the area where our network experience is not that good, where we have good market share as well as not such a good market share, we have, exactly, we know what to do in each of these sectors. So the whole 965 sectors that we have throughout the country, there's already been identified, and we're going to be very granular in attacking each of these sectors.
I think this is one area of one example in terms of how we're going to be a lot more granular in terms of how we're going to attack the market and how we're going to win in the market. Last but not least, just to touch a little bit in terms of the organization, we have reorganized the organization a little bit. We're looking at now it's a little bit flatter organization broken into three core groups. The first is being the customer-facing organization where we start looking at from a segment point of view as well as from product and customer support. These are four different groups inside. We have a consumer and home business units. So this is, as you can see, we are prioritizing a little bit more in terms of what we're going to do to the home segment.
Now, the home segment, it reports directly to me. We have enterprise business units as well as a product and innovation unit and the customer experience and customer service unit in the customer-facing and revenue-generating organization. Then we have a technology and digitization organization, which now have the digitization team, IT team, and the network team. All three reports directly to me and transformation and enabling organization with the strategy, transformation, and governance team, including ethics and integrity inside this organization, human capital, and as well as finance, so it's three core groups: a flatter organization, also a flatter management organization, a lot more focus, and with clearer accountabilities and we are going to introduce a lot of initiatives that are going to be cross-functional basis where each of the senior management will have to work in a cross-functional manner.
Now, just to recap, in the last 30 seconds, what are the key messages that I would like to leave to the room today? Number one is we have stabilized the company. The revenue engine is working. We started to gain back the market share despite the challenging market conditions. However, we know that the cost and profitability is still challenging. We know also some of our strategy is working. We are growing in terms of the high-value customers. In terms of digitization, we're starting to see some of the impact in terms of customer adoption of our digital services. And also, customer experience is starting to pay off. Now, organization and execution plan is in place. We have focused on the four key themes as I presented earlier, very high level. The cost transformation is in place.
The program is in place, and we're going to put a lot of focus on cost transformation and network monetization. Last but not least, now we have a new team with a flatter structure. Hopefully, we are going to execute a lot more efficiently. With that, thank you.
Our presentation for XL Axiata, as well as Celcom, is divided into two parts. I'll start talking about our strategy, which was set two and a half years, three years ago, and the link result to that. Maybe some of you remember last year we were on the stage. We talked about our partnership with Google YouTube. We talked about some of the very disruptive offerings that we had on the market with the fist in the air, talking about free voice and free SMS, which was pretty bad for our competitors at that time. And we talked about our network rollout.
I'll somehow touch upon the same topics, but at the same time, put a little bit of flavor regarding our SIM registration process that happened in 2018. Now, Adlan, our CFO, will later on cover most of the questions that we got in the last two quarters regarding the yield. I think some of you have asked that question, right? I can see some of you are doing like this. So we will try, at least try to cover that and where are we right now on the yield and give some explanations. Now, this is a strategy. I think all of you remember that. First of all, it's a dual-brand strategy. We have XL, we have Axis. XL is mainly targeting white-collar and blue-collar. Axis is for the young people of Indonesia, and of course, there's somehow a spillover effect between the two brands.
Secondly, it's about the network rollout. And as you heard on the Q&A, we still believe there's a lot of potential outside Java. It's not only about putting a lot of sites now. It's doing it in a smart way, and it's doing it by CapEx planning as well. And the last one is, of course, the mobile data leadership. We strongly believe in data-centric going to be the way forward. And we also believe that the legacy product, the voice and SMS, is a sunset business. But at the same time, we also believe, and we can see that the total voice minutes is not going down. It's just shifting service. So it's about doing the data and monetizing the data in the right way. Now, looking at XL and looking at Axis.
Last year, we talked about this one on the right, top left side, the partnership with YouTube, and that gave us some momentum in Indonesia. If you talk to the consumer or customers of Indonesia today, they're saying that XL is kind of a link to YouTube. We are always red, even though it's not a nice color for us, but we are red in the market. Now, we want to accelerate this data usage, this data penetration in 2018, so what we did, and it sounds very easy when I'm saying it, but it took a long, long time, a lot of negotiation, we were actually able to introduce a handset, a YouTube handset. The reason for doing this was, and they still are in Indonesia, only 50% of the blue-collar guys have a smartphone.
We also know that the white-collar, less than 30%, has a 4G handset. We introduced a very low handset with a flavor of YouTube, XL YouTube brand on the back of the handset with some YouTube Go, etc. That got a very good momentum in the market. Today, we are not only known as a YouTube brand, but also as a handset brand. On the Axis part, it was basically continuing what we were doing, saying that we knew that it was about gaming for the young people. It was about music. It was about watching TV, watching whatever they are watching. We just needed to elaborate on that one. We also knew that people wanted, because Axis was mainly a 3G brand, we wanted them to move to 4G. We introduced what we see here is an awesome product.
This is a mainly 4G offering for our young people of Indonesia. We are still trying to differentiate the two products by the way we are offering and the proposition we are doing in the market at the same time when it comes to our distribution. Now, the good part here on your screen on your right side, that's the NPS score. We were very concerned in the beginning of the year regarding the NPS. As you can see, it was pretty low. But what happened during the year was actually we more than doubled the NPS score. And I think most of you know the net promoter score, right? How many promoters do you have minus how many detractors do you have? The best part, we are closing the gap to our competitor.
So today, we are only 5 percentage points away from Telkomsel, even though that'd be much higher than what we have seen in the market so far. Now, this is for us a very interesting slide. Hopefully, you'll see it as well. As you remember, in May 2018, the SIM registration process was introduced in Indonesia. We were very, very prepared for that one. Just a couple of examples. First of all, for Axis, we had something called IGO. IGO is a very old traditional way of doing it. The SIM card industry is about a scratch card. It's about a voucher. So people in Indonesia still want to go down and buy something. They want to purchase at a mom-and-pop shop. So instead of buying a SIM, they went down to buy a scratch card. We were prepared. So they do not need to do registration.
They just buy a scratch card. For Excel, we did almost the same. We introduced a hybrid product where you have a package. You both have a SIM card and a voucher. Sounds crazy, I know. But you can use the SIM card. You need to do registration, or you can throw the SIM card away, and you can use the voucher and transfer whatever proofs that you have brought into your old handset, meaning you do not need to do registration. And the last example is that we hired a lot of people in our call center. So every time a customer called somebody and we could see in the system that they have not done the registration, we diverted the call into our call center, and we made sure that this customer, whoever it was, did the registration.
In that way, we were actually able to secure 97% of all our customers, which hasn't aged more than three months. Now, this one, a lot of concern from us and from many of you guys, what happened post-registration when it comes to transaction in the market where people go down and do something, whatever scratch card, SIM, whatever, and the revenue linked to that, and as you can see on the bottom, it's a transaction. How many transactions do you do in the shops, and you can see for XL, when it was introduced, it went down, but we slowly recovered in the distribution regarding number of transactions. In Axis, we were better prepared because we already had this voucher, scratch card. As you can see, it didn't went down. It's actually higher than before the registration. Now, the revenue linked to these transactions is actually increasing.
There was no dent in the revenue for the new customer after the SIM registration, meaning that number one, we took the prices up. Number two, the customer buys a higher package for a higher price with more GB. It was very positive for us post the SIM registration. Network, that's a hot topic. We were very happy what happened, and we got a lot of help from our analytics team in XL Axiata. What we actually did, we introduced one single KPI, which was covering both the coverage and the experience of our network for our customer. We call it NQI. With that NQI, we were actually able to determine where to exactly put each of the sites in a lat-long.
And we can see today it's working because our competitors are actually copying what every time we put a site, we can see that our competitors are always putting a site beside that one. So that was very successful because it's not just to create a lot of sites in Indonesia. It's about doing the CapEx planning in the right way. So as you can see here, yes, we have done a lot of new sites. We are now covering more than 420 sites at the end of the year. As you can see, it's mainly on the west and on the north. At the same time, you can also see that the revenue has increased 38% outside Java. So I've got this big discussion. Why not continue that journey, but do it in a smart way with the right planning?
Of course, look at the return that you're getting from these sites. Now, this one, which is linked to the data-centric strategy. I could talk a long time about what we're doing about product, etc., what happened in the market. I just want to show this one slide. What are we doing? When you can see this mainly talks about that we are ahead of our competitor. Look at the smartphone penetration. We are number one in the market. We have more than 78% smartphone penetration in our network. If you look at the acquisition happened in Indonesia today, 93% of any new acquisition we are taking has a smartphone. That's the highest number in Indonesia at the moment, and you can see our competitors are lagging behind. When it comes to revenue, almost 70% of our total revenue in XL Axiata is data-driven.
And you know what happened with our competitor? Very difficult to monetize the legacy product. We are not in that discussion right now. We need to monetize our data. Especially when you look at the data traffic, we are much higher per subs than our competitor. Very good, but at the same time, we need to monetize. We need to make sure that we're getting the right return from these data customers. And the last one, which is on the bottom left side, for me, the most interesting part is the Facebook. Now we are getting daily analysis from Facebook and stats from Facebook. And we have it today. When you look at the Facebook, we have a market share of around 22-23% in the market. But when we look at acquisition, we can see that we are taking the highest share.
Every time Facebook sees a new customer on Facebook, they can see that they are coming from our network. So we are taking almost 30% of any acquisition in Indonesia going into Facebook. This is a very good indication that we are on the right journey when it comes to data centricity. Now I'll hand over to Adlan to at least show us how does that look in the whole industry when it comes to market share. You have the clicker?
Yeah, yeah. So very quickly, I think if you look at the strategy that we have done for the last two years, two and a half years, I think we are starting to see results, right?
If you look at the year-to-date September, I think you probably can see from the numbers itself that we are not only growing our revenue market share, but also our EBITDA market share. While the rest of the industry, I think for the nine months of 2018, declined by 8.9%, I think we are probably the only one that has actually remained flat, and as a result, I think we have probably gained market share by about 1.6%. On the EBITDA side as well, I think we have actually managed to keep EBITDA pretty flat, and I think growing EBITDA share to about 13.8%. While the entire industry, as a result of the implementation of prepaid registration, actually declined by 19.4%. Yeah.
However, despite these strong results that we have shown outperforming the market, I think we have seen in our call, in Q3 call, there were a lot of concerns from investors, shareholders in terms of the declining yield. Maybe I think let me take this opportunity probably to elaborate a little bit in terms of the declining yield that what we're seeing. And I think technically the declining yield itself, yield is a function of a lot of factors. And declining yield does not translate to reducing price, right? I think we'll try and articulate this probably a little bit better in some of the analysis that we've probably done. I think firstly, I think if you probably seen that one of the things that we have probably done, there are probably four main factors I can attribute for the declining yield in our numbers.
One is since beginning of the year, we have been upselling to our existing customers bigger packets, data packets. And as a result, we are collecting more from these existing customers, even though, right, by collecting more, it's also giving us a lower yield, right? We'll probably take a little bit more detail to substantiate this with some data in the subsequent slide. Second is we are the leader in 4G, right? The migration from 3G and 4G, we are probably the fastest in the market today, right? And I think, as you know, that 4G yields are probably lower than 3G. But as we accelerate over the last 12- 18 months into 4G, our yields are probably declining faster than everyone else. But is that something that bad or good or bad? I think it's something that is in line with our strategy.
We'll talk a little bit about numbers as well in subsequent slides. Thirdly is better utilization of our network, right? We are trying to promote a lot more off-peak traffic as well because we saw last year the gap between peak and off-peak traffic, there was a significant gap. So to better utilize our network, we have introduced some cheap off-peak traffic product in the market earlier this year, and it has actually borne some fruits, which we probably would share as well. Last but not least, as we continue to expand our network outside Java, we are also going quite aggressive in terms of acquisition, right? We've built this new network, and I think we still have a very low market share outside of Java. And typically, I think the aspiration is to increase our market share.
And you probably would also know that the acquisition yield today for all operators in Indonesia, the acquisition yields are probably lower than where existing customers are recurring their package every month, right? So, and as a result of us being aggressive in expanding our market share outside Java and actually growing market share, that has also an impact on our yield, right? So let's take this analysis one by one to see how does it all impact our yield. Just very quickly, I think if you look at in quarter one, 70% of our customers have actually packaged less than 5G.
However, because of our success in upselling bigger packets to the customers, I think throughout the year, when I think beginning of this year, we actually had the, what we call it, the channel, the MCCM, where we consolidate all our offerings of the various channels that we upsell to existing customers. And the result was actually quite positive. And if you can see in Q3, we have managed to upsell a lot of our customers to bigger data packets. Yeah. And as a result, we are collecting more from these customers as well. So from an ARPU perspective, a cohort basis, we actually see increasing ARPU from this group of customers, right? So we are collecting more and increasing share of wallet. However, as we move and sell a bigger package to this customer, it has an impact on yield, right?
As you can probably see as a result of that as well, we see customers are probably using more. On an index basis, you can see that our GB per subs have probably increased by about 49%, right? As a result of us selling more bigger packets to customers, and they also tend to use more. And if you look at the index utilization as well, from every level of the packet that we are selling, customers are actually using more. Whilst this has got an impact on yield, however, I think it's in line with our strategy in terms of monetizing data because we are collecting more from customers as well. And as a result, we are also seeing an increase in ARPU from this set of customers, right? So, and that's one of the biggest factors that's probably impacting yield. Is that good or bad for us, right?
It's in line with our strategy, and it's part of our data monetization. Second point, I think if you look at 4G, we are the leader in 4G. If you look at our traffic today, more than 75% of our traffic today are already on 4G. If you look at our 2G traffic today, it's less than 5%, right? And I think if you look at our 4G traffic today, it has grown close to about eight times, right? And you know traffic yield on 4G is actually much lower than in 2G or 3G, right? As a result of the fast migration from 3G to 4G, you are actually seeing yield is actually impacted. The good news is while yield is declining, cost per MB is also declining even much faster because 4G is actually more profitable in terms of serving the data customers.
So we are migrating people to 4G, and we are probably ahead of the industry. As a result, this has got an impact on yield, right? Remember last year, I think Axis was clearly more of a 3G product, right? So, and I think late last year, we started introducing 4G on Axis. And as a result of that, we also see a significant acceleration in Axis customers moving into 4G, right? And that has probably got an impact as well on the migration of 3G- 4G as well as the data yield. But does that mean bad for us, right? If you look, as people evolve from 3G- 4G, you probably see that ARPU is increasing for these customers. You also see that traffic is also increasing as they use more of a, they upgrade to bigger packet and they use more of the data traffic.
It impacts the data yield. So from our perspective, with regards to this, we will continue to accelerate the migration from 3G- 4G, right? And I think our focus, especially in Java as well, is purely about 4G. And I think we are probably ahead in the game, right? And as a result of that, that's got probably impacting yield. This has got nothing to do with reducing price in the market. Third point, I think you've seen that to better utilize our network, we have actually introduced a data packet that promotes off-peak traffic, right? So, and I think we introduced that late last year, and it has got significant traction in the market, yeah? And today, obviously this off-peak traffic carries a lower yield than probably the peak traffic, right? The data traffic.
Today, I think you probably have seen from beginning of the year to today, the off-peak traffic has actually increased from mid-single digit as percentage of total traffic to mid-teens in terms of the total traffic. To an extent that we are better utilizing our network, but also this has got a lower yield as well, and that has probably impacted the average yield, our average yield in the market. So, all in all, I think if you look at the three, four factors that I've actually probably shared with all of you today, that none of this actually translates to a declining price in the market, right? We are upselling more bigger packets of our customers, we are collecting more, we are migrating faster to 4G, which unfortunately has got a lower yield than 3G, but essentially it gives us a higher ARPU from these customers.
Lastly is to better utilize our network as a result of promoting the off-peak traffic. Yeah, so I think the impact, the declining yield that you, what everybody is seeing in the quarter three numbers, and I think are a function of this. It's not necessarily a result of a declining price that we are probably reducing price. In fact, on the acquisition product, we have actually increased prices in the market, right? Since prepaid registration in May, I think we have actually continuously increased the acquisition product in the market. I have to say that if you look from an acquisition product perspective, the gap between us and the market leader has actually narrowed down if you compare 12- 18 months ago. Today, probably the acquisition prices are probably the gap is probably around 10% or so, right?
So we are definitely closing the gap from an acquisition product versus the market leader. Okay, having said all that, right? I think again in line with the team, I think that focusing on profitability, right? I think in 2019, we are probably trying to look more in terms of how we achieve a sustainable profit moving forward, right? And I think it's not going to run away from our core strategy, but this is the area that we are probably going to focus on. Data monetization. Data monetization is going to be a key part of our strategy, especially within Java, right? In the Java area, I think our strategy is to defend our market share, monetize data, collect more from customers. And at the end of the day, I think that will lead to probably trying to extract more from existing customers, right?
So Java is very clear in terms of how we're going to grow our revenue coming from data monetization from existing customer base. Outside Java, outside Java is we build our network, we need to achieve scale, right? We need to continue to be aggressive in acquiring market share outside Java. Today, 20% Telkom, the market leader has got about 80% share. 20% share is shared between the three of us, and we probably have the lion's share of that. However, we have not achieved scale at this point in time. And with our expansion outside Java today, we've seen that the quality of our network, the perceived perception of a consumer on our network quality has also improved quite substantially, as indicated by the increasing net promoter score for both XL and Axis brand.
We have seen so far that the trajectory of growth outside Java has been quite significant, right? We have seen the double-digit growth, and I think for the next three years, we are probably looking at the same trend as well as we continue to gain market share outside Java. Third pillar of our strategy as well is on cost optimization. Nothing new here. I think as part of our transformation exercise, costs have been the key in terms of us moving forward. And I think one of the key challenges that we probably have is how do we keep our costs flat, right? And I think that's the challenge that all of us in the industry have today, right? And I think so far in 2018, we have managed to keep our costs flat, even though we are probably building more new sites, especially expanding outside Java.
In 2019 and beyond, I think we probably need to do more, but focus would probably be more on structural changes, right? We are already changing the landscape, for example, in the tower business that we are enjoying probably rental at 50% discount when we renew these sites, right? There are a lot of other areas, for example, in terms of transforming our cost structure. For example, one of the things that we are probably doing, we are moving to fiberize a lot of our end sites today, right? Instead of building microwaves, we are fiberizing the sites. Because I think when we look at the two, three, five years business case, for example, it makes a lot of sense to move straight into fiber rather than microwave and continue to replace them over a three-year period.
At the same time as well, we also see that on microwave as well, the frequency fees have been increased as well by the regulatory, right? And as a result, it makes you serving customers with microwaves even more expensive. A lot more other structural changes that we are probably doing on the channel side and in terms of how we do our network architecture and design to probably drive this cost optimization. The fourth pillar, and it's also very important, is on analytics, right? Analytics is going to be the key pillar in terms of us driving all our initiative, be it in terms of upselling to customers, in terms of growth to market, as well as our network planning, right? And this is going to be the center of all our initiatives moving forward.
Last but not least, we started digitalizing our processes, not just internally, but also all our customers' touchpoints as well as the channel side. While I think we are starting to invest, I think in 2019, you would expect some benefits coming from this digital initiative that we are probably doing. All this initiative, I think, will probably give us a lot of muscles in terms of trying to drive into a more sustainable profitability starting from 2019 onwards.
Okay, that's it. Thank you for making sure.
Good afternoon. Look, I've been attending this investor conference for the last eight years, and I hardly have seen any interest from investors about Robi. But I think today was one of the exceptions.
A lot of questions were asked about Robi, and I'm sure I'm going to pass this information to my team, and it will definitely motivate them to do even better. Anyway, I have Roni Tohme. He's our CFO. He's with me. Any numbers, he's going to explain. But in general, let me take you through whatever work that we have done. What we are going to cover today is primarily the dual-brand strategy that works so well for us, and you have seen already about XL and also about 4G leadership that Robi has taken in 2018. At a glance like how Robi is performing, I think you know all about this, but still quickly to take you through. In terms of RGM, which is revenue growth multiplier, 2016 was a year of pre-merger.
We had a difficult time, but in 2017, despite operational integration and network merger, we managed to achieve growth, and not only growth, also increased revenue market share. In 2018, as of now, it looks quite good. Hopefully, we'll be achieving around 2-2.5x of RGM. In terms of service revenue in the last two years, we are achieving around 17%-18% growth. In terms of EBITDA, it is almost double. Even if you look at it in terms of PAT, it has improved from minus 8.4 to almost minus 1, 1.5. Just to inform you, these are underlying numbers. Reported numbers are in general better than the number that what we are sharing here.
In terms of CapEx intensity, it has marginally declined, but a clarification here in case of 2016 and 2017, that includes spectrum 4G renewal, not for the technology royalty fees, and also the merger network integration CapEx. So all these are included in 2016 and 2017 numbers. And finally, ROIC, it was negative 8%. Now it has come to a positive level, and that trend is continuing for the last seven quarters. Moving on, just to give a market perspective where we stand in terms of market. In terms of service revenue market share, if you see that we have gained 2.3 percentage points as against GP of 1.5, and most of this has come from our third strong player in the market, which is Banglalink, which belongs to VEON.
In terms of share of net adds in the last two years, Robi's growth is as much as number one player in the market, which is Grameenphone. So you can see that Grameenphone, with a huge base, has grown relatively less than Robi. And if you look at the NPS, if you look at the bottom slide, that is there you see Robi and Airtel used to be the bottom-ranked third and fourth player in terms of net promoter score. But in such a short time, now Robi and Airtel have turned out to be number one and number two player in net promoter score. And not only that, the biggest weakness of Robi used to be network NPS, and that's where also Airtel has become now number one player, even exceeded Grameenphone. So overall, it sounds to be a good story for Robi as of now.
Moving on, what are the drivers of this success as of now? I'll not say great success, but it will be a reasonable success post-merger and 4G launch. First one was dual-brand, and I must admit one fact, like XL started with dual-brand strategy in our group. Initially, they had some struggle, but then they came out very successful. But the good part of Robi, Robi has seen that experience of XL and picked up each of those learnings and applied here so that we don't make that mistake. And exactly that Robi team has done. And therefore, dual-brand has turned out to be a great success for Robi. And you may be knowing globally, there are very few success stories of dual-brand there. So from that perspective, I would say the team has done a good job. We'll talk about that in the next subsequent slides.
In terms of 4G leadership, you know the way I term, 2017 was the year of merger, 2018 was the year of 4G. The third point is operational excellence. If you look at Robi number, and there was one question about profitability improvement quarter on quarter or year- on- year basis, there also the reality is actually for the last seven quarters, we have been continuously improving on a number of things, including cost excellence, operational excellence, digitalization. Each and every area, we have made a significant improvement. Now, if you talk about dual-brand, just to give you a perspective, before merger, where we used to stand. If you look at brand perception level in terms of best network, which I just said, Grameenphone has been on the top. In terms of price, also they are on the top.
If you look at Robi, it was in the middle at par with Banglalink, and Airtel was with very poor network. They used to offer least price. So they used to have a price leadership. The geographic play was Robi was more of a regional play who used to play only in certain regions. The red dotted area, you can see that was Robi dominated area, where we were dominant number one player. But other areas were quite weak. On the other hand, if you look at Airtel, which is shaded one, there are Dhaka and Mymensingh; there are two areas Airtel was strong. In the rest of the other areas, we were quite weak in terms of network presence and also market share position. So that was the situation. Then dual-brand approach, we clearly identified brand positioning of the two brands.
Robi stands for innovative digital brand, and Airtel stands for price leadership and number one network for friends. This is a youth-centric brand. And those two propositions have come out to be pretty successful as it looks now. And if you talk about the result, in terms of how to evaluate the two brand success, if you see both the brands are growing faster than the competition, that is one of the criteria of success. And if you look at Airtel has grown by 24%, Robi five%, as against Grameenphone of four%, and Banglalink had a growth of nine%. And in terms of one of the great successes of this two-brand strategy, where Robi has been very, very strong and not too dependent on some regional places, we have been extremely successful in the areas where Grameenphone and Banglalink were strong.
If you look at that non-CCD, non-CCD is the area where Robi was not focusing in the past. And with that focus, more than 50% of the growth, EBITDA and PAT was coming from the new areas where we never focused in the past. So that is the way we evaluate that dual-brand was quite a success for Robi as of now. Now, the second one, that 4G leadership, I don't think in Bangladesh market, no one ever expected Robi would be coming out so strong in 4G. In 3G, we were behind Grameenphone. We were behind Banglalink. In 2G, we were significantly behind Grameenphone, but we were at par with Banglalink. In that kind of situation, actually, neither competition nor anyone else had expected that Robi could come out so big bang on 4G. And management and group proposition was very clear.
We want to go for a data play. Data play, if we want to really win in the long run, we must take it dominant in any new technology comes. That's what we did in Robi and very successfully. On the first day of launch, that evening, we were getting license, the next day we had covered 4G in 64 districts. Bangladesh has got 64 districts, and all these 64 districts, we introduced 4G. Not only that, 1,500 sites within one day, and it was quite unbelievable. No one competition was not there at all. As of now, though the number here is saying we have got 4.2 million 4G unique subscribers, as of last week, it was exceeded 5.5 million. Again, competition is nowhere in the picture as such.
As per Facebook, from Facebook, you can get to know how much is the market share, subscriber market share of Facebook base, where in Bangladesh, more than 67% data users use Facebook. So from that perspective, we have got almost 51% 4G market share. Okay. So that is the way to look at it. And if you look at the breakup in 4G, if you look at average data, Robi market share is 26% and Airtel is 8.2%. And in terms of 4G, it is 36% and 15%. So that kind of success we have got as of now is not only that, the flexibility we have around the spectrum and the kind of technology we're using, that is also giving us a bit of edge.
If you look on the right side, that 4G bands, if you look at, Robi is the only one who is using L900 and U900 bands to give a better experience to our customer. Competition has not been able to do that as of now. Sites, 7,500 against the second best player, Grameenphone, which is 5,000. Coverage, we are significantly ahead of competition, 75% versus 45% and 28% respectively of competitor number one and two. So in all aspects, Robi has outsmarted and outperformed the competition in terms of 4G, and we want to maintain that dominance in 4G leadership going forward. Then operational excellence, you have already seen the kind of EBITDA improvement year- on- year we have been delivering. It's only getting better. It was coming out of, of course, underlying performance in terms of revenue is contributing to in a big way.
The second big factor is margin synergy. All the margin synergy we targeted, we have over-delivered on that. And simultaneously, we are working on cost excellence program that is also delivering a good result. So all this together is giving us excellent improvement in profitability improvement. And if you look at the EBITDA market share, 9%- 16% from market perspective. But one thing you have to keep that in mind here. In case of Robi, we have split our tower assets. In case of Grameenphone and Banglalink, it is not the case. If you consider that tower, in that case, actually, our share will be more than 20%.
If you look at OpEx, last two years, 2017 and 2018, our absolute OpEx, despite significant increase in revenue, despite significant expansion of network, we have managed to keep cost in absolute terms same as or lower than 2016 level in 2017 and 2018. That's what we have achieved. CapEx intensity, gradually we are also reducing CapEx intensity. If you look at in quarter one, 2017, it was 26%. Now it has come down to 20%. Even going forward, we'll make an attempt to further bring discipline into CapEx spending. Now, while there are a lot of good things going on around Robi, but of course, one of the biggest challenges that we all know is regulatory environment. It has not been very good.
There are a lot of challenges around regulatory environment, but I'm happy to share that some of the good things have started happening. Already, you know that unified pricing happened a couple of months back, and there was a question around that as well. What unified pricing is doing in the phase one when that voice growth, degrowth is quite prominent, at that point of time, unified pricing is coming as a good support, and also, it is helping us to win against the number one player, which is Grameenphone. And initially, there is an upside of revenue, but I think in the second phase, hopefully from next year, we'll see significant improvement in terms of churn because of that unified pricing, and that's the way we believe Robi can take a stronger position in the market going forward.
Second one is I'm happy to announce though here, it is the significant market power. It is showing that it is coming. Actually, it is done. SMP has been implemented, not implemented, I would say it has been passed. The regulation around SMP has just been passed. It came out in the Gazette or printed only a couple of days back. So I think once the regulations will be further worked out, and I believe this will give another impetus to Robi to grow further. This is after a lot of hard work we have managed to get it. Data floor price is expected immediately after election. You know that Bangladesh election is coming towards the end of December. So after that, we are expecting that.
Mobile number portability in the subsequent slide I'll share, it has been implemented, and so far, it is very much in favor of Robi. The unified licensing structure being worked out, this is one of the major bottlenecks because we can't build our own tower, we can't build our own fiber. So there are many restrictions around that. So hopefully, with unified pricing, when it happens, we should be able to overcome some of our existing challenges. And active sharing, if you look at, this is something on principle we have agreed. We may be allowed RAN sharing, but not necessarily in the phase one, the spectrum sharing. So this is something also we are expecting within the next one or two quarters, it may come.
So while it is not up to the mark from regulatory perspective, but a lot of positive things have started happening over the last couple of quarters, and we see some good hope even in the coming days. Now, this is just a snapshot of that NPS number one, number two. It doesn't make any sense if it doesn't get translated into some numbers. And this MNP is a good indication by which we can say our network is really good all across the country, and it is helping us to grow. The number that we are sharing here, it is more of an average number as of today. But if you look at current ratios, it is not in case of Robi, it is 3:1. In case of Airtel, it is 2:1 roughly.
But if you look at yesterday's number or last one week's number, it is far more than that. It is around 5:1 versus and 3:1 in case of Airtel. So the gap is widening every day we are passing. And you can see the newspaper, they're declaring that Robi is a clear winner in terms of MNP. Robi is a clear winner in terms of 4G. So it's giving a new image to Robi as a premium brand and as an aggressive brand in the country. This is the last slide from me. Saying everything looks good here, but bottom line is, as of now, we are not that positive as such. Yes, there's significant improvement happened, but because of that taxation regime, various other challenges, plus two years of integration and everything, we could not achieve PAT positive.
So up to 2019, our philosophy and our strategy is to drive growth, growth, growth. And simultaneously, we have to start working on PAT positive as well. So EBITDA, you can see that high double-digit growth in 2017 as well as in 2018. And that kind of growth momentum, we will try to continue even in 2019. But that will not be good enough to have a positive PAT. So positive PAT, hopefully, we should start getting from exit 2019 or early 2020. And how it is going to happen? It is going to happen through, I have already shared that geo-expansion. Earlier, we were concentrated in few regions. Now, we are all across the country, but investment is happening in a very selective way. Wherever there is money, we are putting investment there, not everywhere.
Voice revenue is a challenge going forward because of 4G and also OTT influences, but we are playing different tactics to minimize impact. Data monetization with data floor price, hopefully, whatever challenges or value destroy is happening in the country, I think through data floor price, we should be able to tackle that very effectively. But as of now, Robi is quite competitive compared to the number one player. Enterprise and digital services will be two critical areas which will be the critical driving growth for Robi going forward. Tan Sri has already talked about it, and Robi is not an exception to drive these two agendas. And MAD organization like Modern, Agile, and Digital, we are progressing quite well, and Robi is ahead of the curve in many ways. And cost excellence, this is a continuous drive that we are going to have.
So as far as Robi is concerned, a lot of things are going in the right direction. We have got a plan to make it profitable portfolio for Axiata. And as a team, we are quite confident that we should be able to take this to that desired level. So that's it from my side. Happy to take any questions after this session during lunchtime or so. So thank you very muc h.
Before I introduce the CEO of Digital, Khairil, let me talk about, let me give a preamble. I mentioned earlier during my presentation, during the introduction, that the single biggest, I mean, we have a biggest frustration we have is that we believe, we can argue with that, that our market cap today has almost zero reflection of our fundamentals.
In fact, on the contrary, as I mentioned earlier, and it's very frustrating that the fact that all our opcos are doing well or very well. But there's one of our opcos, the market has given zero value all this while, understandably. In this case, understandably so. So when I say the valuation right now between the three pillars of digital telco, digital business, and infrastructure is 85, 0, 15, it's not so much about what we believe the valuation is, or it's more because what is being recognized by the market, which is almost zero. Indeed, however, we have through actual acquisition of some of our businesses, but more towards more external validation, is indeed easily more than 5% of our valuation, easily. Now, I know we have a very tough time trying to convince all of you that it is worth much more than what it is today.
So zero is not a reflection of what it is today. It's a reflection of what we believe the market is treating our digital value. So the next presentation, therefore, is very important to convince you that there's huge value that has been accrued from digital, and the investment they have done so far is much, much more than worth it. Like I said, the external validation plus the external valuation of the exercise that we did recently has proven that it's a lot much more than what it is today. Now, we have a lot of demonstrations to show the actual products or products that we have across the digital business. The next presentation will therefore be able to at least enlighten you the value of digital. If not, we have canceled all your flights to make sure until then you believe there's some value. Okay.
Now, let me welcome you, the CEO of Digital.
Thank you. Thank you, Tan Sri. No pressure, right? That was a tough introduction. So it's tough. This slot is tough for several reasons. One is it's the last slot before lunch, so I'm holding you back from your wonderful lunch. The other thing is that setup from Tan Sri Jamal just now just sort of makes it very difficult. I have 45 minutes to convince you all that there is some value in terms of what we've built. I'm going to start with a very busy slide because the reason why I do this is in the past, whenever we do these kind of presentations, by the time it comes to us, we have very little time, so it's very important that we get out these key messages. So I actually prepared an executive summary this time around.
The first thing to note about what's happened is in 2018, we actually completed our pivot to become very focused on three main businesses, right? If you like our financial business under the brand name of Boost, you'll see it over there, our advertising business, our ad tech business called ADA, and Apigate, which is our large platform business. And I think the premise behind all of this is that we had gone through a big part of our history trying to figure out what are business models, what are businesses that us as a telco with all the telco assets can actually leverage upon to build and to actually grow. And we actually figured out that these were the right three to build. 2018 was a big year for these three businesses, and I'll describe a lot more on that.
Obviously, we're now on track with our ambition, and I think we've stated this ambition before that for each one of these businesses, we want to be able to contribute at least one billion in EV to the group within the next five years or so. And I actually, hopefully from the data that we'll share with you, you'll be convinced that there is actually some value to be accrued here. Let me start with these businesses. Number one, Boost. It is now the number one e-wallet in Malaysia. Hopefully, some of you are Boost users here. If you're not, then you're kind of a subset of the outside of the 3.3 million users that are now using Boost in Malaysia. We do have 54,000 merchants now in Malaysia. That is the largest. I think the closest next to us is the Alipay guys.
Of course, Alipay is only available to Chinese wallet holders. They're at about 20,000. We are at 54,000. We're quite proud of that, and then the thing that scares us a lot is the fact that our transaction volume on a month-on-month basis is growing close to 40%, so it is a fantastic growth story for us right now, and then we're at the cusp of sort of monetizing it. In Indonesia, we've also got the Boost brand there. It's now built the largest retail base. If you compare it to kind of all our competitors in Indonesia, GoPay, OVO, some of these other wallets, we have about 400,000 merchants that are out in Indonesia that actually sport a Boost QR.
That number will probably hit 500,000 before the end of the year, and by sometime next year, we will probably be at the 2 million mark in Indonesia, which then creates a huge asset base for us to leverage on. That's on Boost. I'll go on to more in a short while. ADA, as it says there, it is now the largest independent digital agency. The word independence is important because we're not associated or affiliated to the WPPs, the media cons, and whatnot of the world. We are independent, and our size is now sort of the largest. We are winning pitches against all these large established players right now. So you'll see some of our clientele in a short while. They include some of the largest banks, some of the largest telcos.
Actually, some of the banks that you guys represent are actually clients of ours, and we've been able to win that mandate from your banks against some of the large advertising agencies in this region. We did close the funding round. I think Tan Sri mentioned that where Sumitomo Corporation injected MYR 20 million into ADA. That represented a pre-money valuation of 2.8 times our cost of investment. We've actually never announced what our pre-money valuation is until this morning. Tan Sri let it slip, but that's okay. I think we probably are in the process of sort of getting further rounds of funding, and it's probably useful to establish that as a lower threshold. On track to be PAT positive. In fact, I'll show you some data in terms of our EBITDA profile. That business is going to generate money, a positive PAT by next year. Next one, Apigate.
Very hard to describe this business. If you don't understand API, then come and see me at lunchtime. I can explain to you the concept of an open API. We built this now. It now has access to APIs from 110 telcos around the world. So it's no longer just confined to our regional operation. It's now connected to pretty much everywhere around the world. So Middle East, Africa, Latam, Southeast Asia, South Asia for sure. We don't normally kind of talk about this, but if you add all the consumers of all that 110 telcos, it amounts to about 3 billion users. So in actual fact, then our platform is now connected to about 3 billion users. We're in the process of monetizing that very aggressively.
You'll see some of the Northbound partners that we signed up with, Disney, ESPN, Amazon Prime, some of these guys that are now connected to us. What it is that then it creates opportunities for us to sort of make spreads, if you like, right? MDRs, if you can call that, rev shares from each and every one of these partners. The remaining businesses, for those of you who've attended previous sessions before, you'll know that we've actually built up a fairly large portfolio prior to the focus on these three businesses. The remaining businesses have now been parked into an entity called Digital Ventures. We are now looking at independent fund managers, third-party fund managers to actually take over and manage that for us so that our management team can focus on those three businesses. We are in the process.
I cannot say much about it right now, but we're going through an independent process. There are final bits that were submitted about two weeks ago from several parties to actually run this. Some of them are actually large, well-known American PE and VC funds likely to book some kind of profit when we complete this transaction for Digital Ventures. So that's that. I have another 40 minutes. I'll go through this very quickly. So this was what I was talking about. For those of you who've attended last year, you'll remember this. We had, over time, from 2013 to about 2016, built up quite a large portfolio, and it was largely with a view of trying to figure out what are winning business models for us to anchor around, for us to leverage, to actually build some value for.
And so in the peak of it, I think there were something like 29 businesses. If you like, a lot of them were actually internal incubations. Some of these incubations were actually done with external parties, JVs, and then, of course, there's some minority investments. The bit on your right, that's our innovation fund. Technically, it doesn't quite sit under the ADS portfolio. This was actually created such that we can actually help local entrepreneurs in Malaysia actually build digital businesses from it. You'll see some of our investments, some well-known brand names there, including Fave, Slurp, Katsana. Actually, one of the things that we do a lot is we find ways for these companies to actually work with our core telco. So if you think about one of the businesses that we invested in, Aerodyne, Aerodyne is a drone inspection service that we invested in about a year ago.
They now do a lot of services for edotco and actually monitoring their towers and looking at and assessing them. So that was Act One, right, all the way until 2016. So in 2017, we talked about the pivot, and we pivoted to these three businesses, right? Financial services, where the large part of it is micropayment. We have launched a microcredit business. It's called Boost Credit with a QR. Bima, obviously, is our microinsurance business within the portfolio. We do have plans for micro-remittance and micro-savings. Those are actually big opportunities for us. But for now, you'll hear a lot of talk, especially if you're in Malaysia, on the micropayments business. ADA, this was an amalgamation of several different ads businesses that we put together.
One of the coolest businesses in this portfolio is the fact that we're leveraging a lot of data that we've ingested from multiple sources to be able to work with clients on outcomes. So we don't charge the clients. So some of the banks, for instance, if we go out and we do credit card acquisition, we don't charge CPMs, CPIs, none of that. We don't charge you for how much inventory that we're booking for you. We charge you for every single revenue-generating credit card customer that we're bringing to you, right? So for instance, for a credit card company in Indonesia, every time we bring them a credit card customer, we charge them $50, right? And obviously, the bank or the credit card issuer is looking to monetize that over time, but we get actually a very nice fee for that.
We use a lot of our data, a lot of our analysis to be able to kind of get that. And then, of course, Apigate, this is the, what do you call it, the business where we allow all these Northbound content developers like the Disneys of the world to connect to our assets and then charging them sort of a spread every time they use our service. So those are the three businesses. Let me go on to each one of them very quickly. So Boost, I think there's a session later in the evening where you can actually kind of get a sample of what these businesses are. So the premise behind this is that in all our markets that we are at, there is actually a fairly large unbanked segment. And in fact, there's also a fairly large underserved segment.
In these businesses, we actually have a right to play, right? Obviously, smartphone penetration in our markets is already very high. Our base of consumers that we can leverage, our base of retailers that we can actually leverage is very, very high. But then, of course, the key here is that our cost to serve, right? We can give out, and I'll give you an example. We give out loans. Our average loan size in Indonesia is $61. The average bank clerk pay for an, I don't know, an hour, it's probably around that same range. So it's impossible for banks to be able to go out. We can leverage the technology. In fact, our unit economics on a $61 loan is very, very high, right? So this is something that we're leveraging considerably to build our business.
So Boost was essentially creating this payment wallet as effectively a master account that we then want to then lever up all other financial services. So if you look at the bottom, once you have this account, we can then sell you a whole bunch of different things, right? Whether it's micropayments, microinsurance, microlending, remittance. We call it the five micros, essentially, right? So the four micros down here plus micropayment. And I think in many regards, we've actually done a pretty good job of sort of penetrating this. It is, like I said before, the largest by far. If you're looking for market share information, I can share with you some anecdotal information that we get in Malaysia from retailers that have opened it up to all other wallets, right?
For those of you who are in Malaysia, you know that there are actually other wallets in the market right now. We work with some retailers that have actually opened it up. In most situations, they'll come back to us and say that we account for about 80% of all wallet transactions that are in their retail premise, right? In some cases, it's as close as 100% as well. With that, I'm actually going to skip this because we don't have that much time now, but this is actually a video. You can watch it when you walk by the booth later on. Let me just try to skip this. Okay. This is the range of things that we now offer within Boost, right? Obviously, the prepaid top-up. We want to leverage the fact that we are a telco.
Most of our telcos pay anything from 6%- 14% to the distributors for a top-up. We obviously charge them. We charge the telcos a lot lower. But in that sense, we also make a lot of money because our technology is a lot cheaper. We have our digital shop that includes things that you can buy on Fave, things that you can buy on 11street, on Zalora, etc. Through that digital shop, obviously, we make some money around that. Payment via QR. This is our largest and fastest-growing use case right now. Like I said, we have 54,000 retailers that are now accepting our QRs. We're hoping to hit 100. Sorry, we're hoping to hit 50,000 before the end of the year. Then, sorry, we're hoping to hit 100,000 by the middle of next year. Request money. This is our P2P function.
You can request money, you can send money. There's a very cool neat feature there that you can actually split your bill if you're going to a restaurant and you're paying. Sometimes restaurants don't like us doing that because now they have to entertain five different bills, and that's quite difficult for them. Bill payments, you'll see that most of the big billers like Telekom Malaysia, some of the other telcos, waterworks companies, those are some of the standard features that we've also built in. Parking, street parking, actually. So one of the coolest things for those of you in Malaysia, when you do street parking, you always have to fidget with your coins and so on and so forth, well, the way we do it is that you book that slot on your app. You go into a meeting. If time runs out, there's a little notification.
You can just top up without actually having to go down and slot new coins and get new tokens and whatnot. That is our second fastest-growing use case in Malaysia right now. Finally, of course, transportation. We're competing against some very entrenched transport players in this market, but we're making some headway. For those of you who are traveling on ERL, on KLIA Ekspres, use our feature. It's actually pretty cool. I use it all the time. It allows you to book your ticket well in advance, and all you have to do is just walk up to the counter, scan, and you'll go through. So those are the features. There's a lot more that's being built right now. This is a quick snapshot on where we make money. And as you look through it, there's a lot of places where we make money, right?
So obviously, on payments, there's a lot of margins on payments, whether you're a bill payment, digital shops, online payments, offline payments, MDRs, essentially, we're charging with the merchants. The float is starting to grow quite large now, and so we're starting to accrue quite a bit of float income because we parked a lot of our balances with a custodial bank. And then, obviously, financial products, right? So microlending, we talked a little bit about it already. As I said before, the unit economics of that business is really good. So far, because of our engine, our credit scoring engine and our smarts that we have in there, MPLs are so far zero in all the loans that we've given up, but it is actually looking quite interesting as a growth opportunity for us. I'll show you some.
I've been asked to take out all the exact numbers there because I don't want my competitors to know what exactly it is. But this is kind of the growth that we've had to experience, right? A huge toll on the team. So from when we last spoke, the last analyst day, which was sort of October, November of last year to October of this year, so 12-month period, we've grown 110 times. I don't know, Tan Sri, any other business that you have that's grown 110 times? Several, Celcom, okay. And then for the last six months, as I mentioned before, for the last six months, month on month, we've been growing about 38%, which obviously puts a huge burden on the team. The team's very excited by the growth. But what's also very exciting is the fact that when you look at usage, right?
So our monthly GTV, our monthly transaction volume, or sorry, transaction value per active user has shot up, right? So in the early days, we found it really hard. We had to educate the market. We had to keep telling them what QR payment was all about. We had to keep telling them, "Keep your cash. This is an easier and a cooler way to do." You'll notice that from about June- July onward, there was a little bit of a spike. That spike was actually caused by the other competitors coming into the market. When they came into the market, we were no longer the only ones communicating this new form of payment to the masses, right? And so when they came in, we kind of took off as well.
Everyone was looking when they Google, "Okay, what's the number one e-wallet in Malaysia?" Boost comes up, and therefore people are downloading and actively using. The best part is the acquisition cost for us is rapidly coming down. Obviously, we leverage a lot of ADA, ATEC Company, to help us with the acquisition, but the acquisition cost is actually coming down. So therefore, the unit economics is looking very, very attractive now. Some of the areas around financial services, Boost, I think I mentioned it to you now, Indonesia, huge spike. What that means is now we're actually building, if you like, a gateway that allows all these other wallets in Indonesia. And these wallets in Indonesia, by the way, are well-funded, right? Gojek has launched a wallet, GoPay. GoPay actually has billions in their coffers to actually push for a very aggressive penetration in their wallets.
So rather than kind of competing with them, because obviously, Axiata is not writing $1 billion checks for me, because of that, we're finding a way where we actually work with them, right? So we're building an API gateway that allows all these other wallets to get access to our 400,000, well, soon to be 500,000 merchant points. And then with that, we earn a very small spread every time they use us. But more importantly, we collect massive and massive amounts of information that can actually be used for our credit lending business, for our credit scoring engines. Then in terms of the Boost Credit, so we launched invoice financing, first and foremost with 11street, obviously, because that's our investee company. That went very well. Loan sizes typically between MYR 500-MYR 5,000. Then we launched a supply chain credit.
This is kind of working capital in Indonesia because, again, we have 400,000 merchants. One of the appeals for us is that now we have information on them, we can actually give them lending. So in Indonesia, we launched this business in July. Loan sizes of about IDR 500,000-IDR 5 million. So very, very small ticket sizes, very high velocity, right. And then in Malaysia, we recently launched a Shariah-compliant loan, MYR 2,000-MYR 20,000, up to 12 months in duration. That was launched about a month ago. So again, sort of looking at broadening this business given the base of merchants that we've accumulated. Unit economics, again, fantastic, making tons of money on this business. Okay, so that's Boost. Moving on to ADA, right? So this is our AdTech business. Their slides are better looking, obviously, because they're an AdTech company.
So this is the core businesses that they have within ADA, that we have essentially an agency practice that does a lot of the digital creatives, right? If anyone saw, I wonder if anyone saw the Singles Day campaign that was being launched by Malaysia Airlines. It was very targeted, probably not targeted at you guys. It was a very targeted campaign. That was all done by the folks at ADA. They had something like about a week to kind of prepare it, right? So they do things. One of their fundamental beliefs is that they do things at one-tenth of the time and therefore one-tenth of the cost, which is helping them win a lot of mandates from some of these large advertisers. So one is an agency practice. Number two is obviously media integration. They're integrating all sorts of media.
So they're playing with what goes on Facebook, what goes on Google, YouTube, what goes on actually the telco inventory, and of course, more recently, what goes on AVOD, right? iflix, HOOQ, some of the AVOD players that are out here in the region. And then a core part of what we have is actually the data. They've actually amassed a huge amount of data, obviously to a certain extent from telcos, but also the other digital properties that are part of the Axiata digital family. We've amassed a lot of that data, and that data is now helping us figure out what's the right targeting for the different segments that we're going after. That's a quick intro on them. As I said before, they are now the largest independent player. They have about 250 specialists now across eight countries.
120 of them are in advertising services, 75 data engineers or data engineering support, and then we have about 50 folks in the ad ops business, right? So a very large business now. The exciting thing about this business, which I'll show in a short while, I think, yeah. So they're actually winning some of these clients. And these are not, these are won through bids. These are won through pitches, right? And we're going against some very, very established media players, right? Agency players. We not only have to break the mold in terms of them moving away from traditional to digital, we're also breaking the mold from moving away from, if you like, the rebate mechanism that a lot of agencies typically use to something which we can offer, right? Which is very data-oriented.
Some very big names that we've been winning over the last few months or so. Obviously, as you see right at the top, Axiata is a big client of ADA, and that's because we were founded to sort of support the Axiata family to start with. But then since then, it's actually kind of gone on to actually working with all these other big brands as well. Here's some information. Again, it's all disguised. Media spend from the start of this year to September has grown 2.1 times roughly. We expect that number to grow even more over the next few months. Net revenue, obviously, because media spend is, if you like, gross revenue. Net revenue has grown in the last few months at about 2.4x. You'll see some spikes because obviously the advertising business has seasonality.
There are certain cycles, certain periods where advertisers kind of push a little bit more, and you see it reflected in us as well. Now, the most exciting part is actually our EBITDA profile. So we had committed to the group that this business will be PAT positive for 2019. Very, very confident that actually this is achievable because when you look at out of the eight markets that we have, six markets are actually already EBITDA positive. And if you look at the profile for next year, given where we're projecting, given all the clients that we've been able to win in, this should be sort of accruing, this should be generating positive PAT for the group. Okay. And then we talked a little bit, I think Tan Sri mentioned we did get an external validation, if you like, from Sumitomo Corp.
The reason why we went with Sumitomo was because they were promising Japanese advertisers, which was something that we found it very, very hard to crack, and then they were also promising access to Japanese inventory, right? Folks like Line and so on and so forth, where we wanted to get into the inventory, so we did this deal with them. They pumped in $20 million. Tan Sri mentioned the valuation at $110 million. That's $110 million, by the way. We're seeing a lot of our assets being kind of built for potential monetization, potential validation in this regard. Final one, Apigate. This is our global API marketplace. I think this is. I think I've heard Tan Sri say that this is one of two businesses within the Axiata family that is actually going global.
It's going global because it's able to sell all of these assets, if you like, right? The best way to describe it is when you book a Grab. Sometimes your Grab driver is late and you want to call him. Well, the number you're calling is actually not his number. And when he sees the number that's being called, it's actually not your number that he's seeing. It actually goes through an API gateway because it's actually masked, right? Because after the transaction, you don't want that driver to have your phone number because he'll be calling him and hounding you. And likewise, you don't want to have his number because Grab will say that, "Hey, look, you can actually make a private booking," which is what Grab doesn't want you to do. So it goes through an API gateway, right?
Another example would be that in several of our markets, let's say if your kids are playing Brawl Stars, which you probably have no clue what game is, you should go talk to your teenage sons or daughters. But if they're playing Brawl Stars and they need to buy tokens and gems and whatever that is, they can actually go into our, they'll pay through Google Play or through the Android, sorry, through the iOS. And then they can actually link that payment to the phone bill, right? Your phone bill. That's actually being facilitated by our gateway. So those are some of the things that we have. So the one that I talked about, carrier billing, that's the example that I mentioned earlier on. The voice and the messaging API, that's the Grab example that we talked about.
So this business, by the way, is sort of tiny fractions of spreads that you're earning for every single transaction. So the key is to get large volumes of transaction, right? Sign up as many what we call northbound players, content providers in the northbound, such that they then flush through a lot of volume through you, right? If you like, the unit cost, once you've actually built the platform, the unit cost is actually relatively low. You do need to have BD people to kind of go up and sign, but the technology is there. It actually allows you to sort of leverage at very, very big scale, right? So if you look at it, this is, again, all the different ways that we make money. So if it's operator billing, it's the margin on the gain, it's margin on the content.
If it's wallet, obviously it's the margin on whatever that we're selling through. So very, very clear ideas how we're actually making money from all these transactions and how we need to get the scale such that we are actually able to achieve the kind of valuations that we want to achieve. Some of the clients that we have now signed in, Tencent, Tencent video, Tencent games, that's being pushed through our platform right now. Obviously, Apple, Twitter, Google Play, those examples that I talked about before. And then you'll see some telco groups as well within the setup here, right? So Zain, Zain is a very large, you guys know it, right? You're a telco analyst. So Zain, very large telco group in the Middle East.
We are now connected to their hub, which then allows, if you like, if you're going after Tencent, Tencent can now connect to the Zane portfolio via our hub. And then for Tencent, it's worth the while because I only need to go to one point of integration. That point of integration allows them to hub with our hub, either in India or in Singapore. That's where our technical hub is. And then it connects to all the Zane portfolio as well. Cartoon Network, some of those interesting labels that we've sort of been able to sign through. Then some information here on how we're doing. So gross transaction value, 1.4x between September and January. Again, quite lumpy. You'll notice that typically summer months when kids are at home and nothing better to do, that's when we see spikes in volumes.
Whenever there's a holiday period, we see spikes in volumes because they tend to spend on these games and whatnot, right? So not as high as we would have liked, but that's because we did a lot of cleanup when we decided to, when we linked up with all these global players, we realized that some parts of our technology was not sort of up to scratch. So we did some form of cleanup to kind of get it up to scratch. Volume of transactions, though, has gone up quite significantly, and so has our list of content partners, right? That's gone up by about 26%, 27%. So we're getting to that level now that it's actually going to be very possible that this business kind of turns on and contributes positive PAT in a short while.
In fact, the commitment we've made to the group is that by 20, not next year, but the year after that, 2020, Apigate will be a PAT positive business, right? And based on kind of our trajectory on the number of content providers that we're signing up, the cost, the unit cost that we have to sort of manage this business, we think that it's absolutely achievable. Okay. Final one. I want to talk about these two businesses, sorry, the whole digital ventures portfolio. Just to kind of let you know, Bima, obviously, this is one of our, this is our micro insurance business. There was a raise that was done pretty much at the end of last year. So after we spoke at the investor conference, soon after that, there was actually a raise.
It was a $97 million raise from Allianz, specifically Allianz X, which is their digital services unit. It's a combination of both primary and secondary shares within that. And basically yielded, if you like, on a cash basis, that basically revalued our portfolio in Allianz increased by 2.5x over our cost of investment, right? So did quite well in that little one. 11street, we closed, I think you have seen the reports, a MYR 40 million round with strategic investment from PUC Group. They're actually a local publicly listed entity here in Malaysia, obviously with quite strong ties with Chinese investors. They invested in 11street. One of the things that has happened since that investment is that it's allowed us to, if you like, kind of take better control over the asset.
One of the things we noticed within Axiata Digital is we're not very good as passive investors. We're better when we become, when we take operational control. So with this investment in 11street, we actually got into the nuts and bolts of the operations. And with that, we actually took a lot of costs out from the operations. When you're a startup, you don't need kind of high burn rates on a lot of expatriates sort of working at 11street. We got rid of all the expatriates. We got rid of all the burn rate. In fact, our burn rate went down by 44%, right? Which basically meant that our net revenue margins became positive for the very first time in 11street. So trending very well there. The last thing I want to mention is I think I gave the heads up before.
We're in the process of carving out all these non-core ventures into a separate entity. This is called our digital ventures portfolio. We are in a middle of the process, so I cannot talk to you too much about it. In fact, can you please ignore the fact that there are four PE and VC funds there? It says four. We should probably not let the whole world know that there are four guys out there because they are obviously competing against each other to try to get a hand on the assets. The transfer of assets we're targeting actually, we should probably be in good shape by the, in about two weeks' time. In fact, actually we're going to make the call on who wins the bid this Thursday or Friday.
Vivek, Tan Sri, myself, we sit on the committee to decide who of these four funds will actually win the bid. And then once they win the bid, we'll do the paperwork, we'll do the transfer by early January, hopefully. And then we're expected to book a pretty substantial one-off profit when we complete this transaction, right? It's a one-time profit. So that's pretty much it for me. Promised Claire that we were going to get you out by 1:00 P.M. for lunch. I don't know if I've done the job like Tan Sri gave me to convince you guys that there is some value in our digital assets. Happy to take questions later in the evening. Thank you.
The first one is on the cost excellence. Where are we so far?
I think there were a few questions during the initial call we had in the morning on where are we and how are we doing on the cost excellence program. I just step back a little bit on what we committed. We committed earlier in 2017. We said that we will deliver MYR 2.3 billion over a three-year period, which was MYR 800 million in 2017, and then another MYR 1.5 billion coming out in the rest of the two years. Then we revised our projection on the cost excellence middle of last year to say that we would deliver MYR 5 billion over five years, which is starting from 2017 until 2022. So I'll just basically give a perspective on where are we versus what we committed and how does that flow down into the P&L of the company.
And then also give you some glimpses of what are our focus areas on cost going forward. I think just going back to what Tan Sri said in the morning, a lot of focus around the profitability and cost is becoming an important agenda of our discussions in different board meetings, as you can see. So far, I think, let me just start. This is a long list of number of projects which we are running across the different opcos and different areas of operations. So cost excellence has been engraved into the operating organizations as one of the key elements run specifically by the CEO along with the CFO. We did not, while we focus on a large number of 80% of the benefits will come out of very few projects.
But just to create the culture of focus around cost, we are actually running many initiatives or projects across the organization. But the main one where we focus so far is really a cost on network. Let me just try and break it down on what we've been able to deliver so far. When I look at, I said 1.3 billion, so this is around a billion plus savings which we've delivered. A large part of that saving is actually coming from XL Axiata, followed by Robi Axiata. I think these are two markets which have been very important when it comes to investments for us going forward and returns on the invested capital in these two markets. So we are seeing a significant effect of cost coming into these two markets, but it's just not these two markets. The benefits have been spread across different markets.
As I said earlier in the morning, that the benefits have been pretty much equally between the OpEx and CapEx savings. We call this initiative as Arise. I mean, one can argue, why do you call Arise? But I won't get into that detail. But what you can see here is that the savings are coming into the P&L. However, these are getting offset by new investments which we are making to continue growing. So you would not see that kind of an impact on the EBITDA margins because we continue to expand our footprints in some of the markets. For example, in three key markets, Indonesia, Bangladesh, and even Celcom, we've expanded our footprint in terms of number of BTS during the last year, which does take away part of the benefits which we've been able to achieve through cost saving.
The other factor which is impacting us is really the data pricing, the data yield, which has been coming down by nearly 50% year- on- year, which has been offset by to protect the consumption by continuing investment, which is again not resulting in EBITDA margin improvement. But this cost excellence program is able to absorb that impact. And then new costs which are coming in some of the markets, which is again something which we have to incur, but still holding on to the EBITDA margin. So in summary, a lot of the cost saving, while I don't see impact onto the EBITDA margins, is actually flowing down into the P&L as we go along. This slide, I think, is what we have. We monitor as a dashboard to see how we are progressing versus each of the markets.
This is comparing one of the key enablers or drivers for cost. We are not looking at cost as an absolute value alone. We are also looking at what's the impact of cost on some of the drivers. For example, what is the cost impact on the subscriber acquisition cost per sub? What is the SAC per customer life value? These arrows, red and green, actually say compared to the last year, did the company do better or did the company do worse? Then across, of course, it basically gives a glimpse of where we are in each of these elements. I think here I would highlight that we've done fairly well on cost when it comes to network cost. You would actually see in most of the opcos, we are seeing the network cost per site coming down.
So that's been a positive development for us. However, two areas where we are now focusing actively is in the marketing cost and the manpower. So those are two areas where we still need to improve as we are where we stand. So those are areas of focus. From how we run the whole cost initiatives is driven through the KPIs which are delivered on the cost targets. So each CEO, along with the management, actually carries cost target as one of the key KPIs. And this is reviewed on a progressive monthly basis at the opco level, as well as what we do at the group level on how we are achieving that. And in addition, we are also looking at CapEx in a more dynamic basis.
There is a part of the CapEx, typically around 85% we allocate to the opco at the beginning of the year, and then we dynamically allocate depending on the performance and need of the ROI in that opco. We are managing through being very focused in terms of setting targets, KPIs, and allocating the amounts based on how dynamically the performance has been achieved. I think let me just go a little bit into details around the areas we've so far focused on and what are the structural and transformation levers we will go ahead and focus, and I'll get into a little bit more details around those areas as we go forward. I think we've done fairly well when it comes to network and IT spend optimization.
Our target has been to bring down IT spends, both CapEx and OpEx together across opcos as a percentage of revenue. So we've done well, and this has been an initiative which has been run horizontally across all opcos. A lot of these savings initially has come through better negotiations with vendors, and also we are seeing now impact on renegotiation on some of the contracts which are resulting in OpEx saving. Similarly, I think looking at sales and channel mix, digitization has been a focus. How do we move more and more transaction online, which is again something which we are monitoring closely, and marketing efficiencies focused between the traditional marketing spend as well as moving more into the digital marketing space. 40% of our spend now actually goes into digital marketing spends.
Similarly, customer service, we're looking at how do we optimize the call center costs across markets and process improvements or digitization, which is really looking at automation across different operations, and looking at the organization structure and rationalization. I think where we are now focusing really is around the structure and transformational areas, which is focused on simplification. I'll get into a little bit more details on all these areas where we are focusing on. Similarly, on digitization and redesigning the operating model, let me just go through details on what we are doing in each of those areas. I think simplification, all OpCos are revisiting their product portfolio. Some of the elements on the number of products has been rationalized. Some are ongoing processes. We're getting out a lot of legacy products by migrating customers progressively from those legacy products into new products.
Even on offerings on new acquisitions, the number of products across opcos have been lowered down. Similarly, I think on the legacy plans, we've been cutting down on those because these do bring a lot of complexity in our cost structure. Looking at some of these non-value-added add-ons, for example, I think voicemail and Celcom is something which does cost us money but does not have that kind of a significant revenue impact. So we are looking at some of those areas. Even, for example, end-to-end processes on dealer management, on collections and billing, those are areas we are looking at simplification. So simplification, process improvement is a key area which we are focusing around all opcos. IT stack simplification and cloudification is another area where we are focusing on across opcos.
I think one of our challenges has been on the change request given the legacy systems which we have in place. So combining the new architecture for IT as well as setting up plans around prioritization on how we deal with the IT changes is an area where we are focusing on, including consolidating and retiring old systems, adoption of open source. So some of these areas is where we are focusing on in terms of IT simplification. Shared service model, we moved to the Axiata Analytics Center, which is a combined resource of 150 data scientists which work not only on improving the customer uptake of different propositions, but are also looking at network analytics, how we can actually improve the cost position in some of the markets.
Similarly, on the digital lab, which is basically moving away from some of the outsource consultants for developing IT applications to internal. Axiata Procurement Center, which is basically focused around common procurement process across the entire organization, has been working. I mean, we've seen across the last couple of years, year- on- year, benefits coming out of the negotiation which we are doing collectively across all opcos and looking at centralized shared service support. Some of them we've started in a small way, but the areas which we will be expanding going forward. Customer care, I think we've talked about digitization of customer care. I think that's an area of focus. Digital sales and marketing, moving for around 25% of all the interactions through digital channels. We're looking at 10% of the customer acquisitions going forward, which would be digital. It's very early stage for us moving there.
It's not an easy one, we realize, but that's the focus area where we're moving towards, including the digitization of the interaction with dealers and retailers. The robotics process across markets, some of the good examples we have in Bangladesh, for example, even the backend processes, we are automating, moving into robotics, looking at some of the other areas of digitization. So these are areas where we are looking at going forward from a structural shift standpoint. Similarly, analytics-based, the network deployment. This is an area which we believe is going to have a significant impact on how we deploy our CapEx going forward. Some of the opcos have been leading in this, and we are trying to use the best practice across into the other opcos on how do we decide on more real-time basis decisions on how the network is going to be implemented.
Similarly, how we decide on the assets replacement. I think asset productivity is one area where we are looking at. We do realize our CapEx intensity has been generally higher. So we have really focused on how do we take out the returns on investments maximum for our invested capital. Some opcos, we are looking at possible shutdown of some of the technologies. We don't see immediate effect, but going forward, maybe in 2021, we would look at some of those effects. And Adlan did mention that, for example, in XL Axiata, we have only 5% of our customers on 2G. So we are looking at selectively moving out of 2G in those markets to bring that effect. Similarly, on the network side, I think looking at consolidation of network procurement through focus around more regional structure.
Standardization, I think we are looking at standardization of our configurations to deliver the benefits on buying capabilities, including consolidation of our procurement. Similarly, on IT, we are looking at standardization of applications, a lot more focus around regionalization and standardization of activities going forward. Similarly, looking at some of the benefits through aggregating the lease lines, looking at how do we maximize through our own wholesale business in terms of getting the best price for wholesale traffic and focusing on looking at media buying at a consolidated level within the company. So I think these are some of the areas which have been focused on. We have not put down the real numbers behind each of them, which is work in progress, but these are areas which we are focusing on going forward. Let me just say where we stand.
I think key areas to deliver five billion saving is going to be network, sales and marketing, and other functions. And these contribute around 40%, 35%, and 25%. So 25% other functions is more backend processes. I would say from a network standpoint, so far, we've had a very strong focus, and that's clearly visible on our network cost per site coming down. We are still on early stage on structural areas, which could mean how do we collect the whole demand management, how do we ensure we are able to move to, for example, managed services. So we are doing some work around the outsourcing of managed services on network, early stage, but we are working on those. I think from a transformational standpoint, network has been still a good progress, which we are seeing in terms of how we decide on network deployment in some of the markets.
And that's largely analytics-based, cluster by cluster, zone by zone. How do we actually have those deployments? I think there's still focus required on the sales and marketing side, which we believe 35% the costs are there, where we are looking at some of these opportunities going forward on how do we do sales-mix product rationalization and digitization in the area of sales and marketing. Other functions is lot-driven through end-to-end process improvement and rationalization of headcount in some of the markets, which we alluded upon earlier in our discussions. So I think we see this initiative taking us at least 300 basis points EBITDA improvement over the next three to four years. And that, to me, is coming after we reinvest most of the savings back into maintaining the effect of data yields, which are coming down across markets.
A lot of these savings are going to reinvest back into the networks, but despite that, we see around 300 basis points improvement in our EBITDA margin over the next three to four years, so I think that's brief on the cost initiatives we are driving, so we are fairly confident at this point in time, given that we've delivered what we said in the last two years on cost, that going forward, we will continue to focus and deliver the five billion target which we had set for ourselves. Let me go to the next section, which I was supposed to cover, which is on the optimum portfolio strategy. I think this is an area which I think a lot of the discussions I had around the table, there was a lot of interest on how Axiata looks at the capital allocation going forward.
And then I'll touch a little bit upon the M&A, how we have, what is the logic or the rationale behind the M&A activities which we carry out, how we've actually performed in the past on M&A activities, and what's the key focus area when we look at M&A going forward. Let me start with this slide. Our capital allocation priorities remain unchanged. As a company, our focus has been, our investor proposition has been moderate growth and moderate dividend. So it's a mix of dividend and a growth play. So it's neither dividend nor a pure growth play. We have three verticals, as you would have seen, telco, infrastructure, and digital businesses. While at a total level, we would look at still a moderate growth and moderate dividend as our investor proposition. However, digital businesses would be focused more around growth.
Tower or infra business, initially focused around growth, eventually would be a dividend yield business as we see, whereas mobile will remain a moderate dividend, moderate growth. With growth, we are looking at single-digit, mid-single-digit growth going forward. So that's the premise of the investor proposition which we have. We will ensure our capital allocation, we keep the prudence of 2.5x gross debt to EBITDA. We have seen that being managed reasonably well. This year has been slightly up, largely coming because of the forex implication. But on a sustainable basis, our capital allocation would be considering the premise of 2.5x gross debt to EBITDA. So we will remain within that risk appetite, as I may say. Dividend, we did take two years moratorium in 2016 and 2017.
2018, we would like to get back to the 2015 levels of dividend, and we would expect progressively dividend to be maintained or grown from where we are. So that's, again, a key consideration for us going forward on how we decide on our capital allocation. In terms of the CapEx for our three core strategies, CapEx we see coming down from the levels on telco from up to levels of 23%-24% at the moment, down to around 16% in some operations and 20%. So we would see a 4- 5 percentage points coming down. Why I say this is the way it would be? Because a lot of markets are at an early stage 4G evolution, be it Nepal, be it Sri Lanka, Bangladesh, even in XL Axiata. So over time, we would see that CapEx on telco coming down to the levels of 16%-20%.
As far as infrastructure is concerned, we look at infrastructure or tower business as self-sufficient. When I say self-sufficient, I think at the moment the CapEx is inflated because we've been acquiring new businesses. Over time, we would expect on an organic basis, CapEx to be at 20%-25%. However, if there are acquisition opportunities which we are actively looking, pursuing in tower business, we would look at those acquisition opportunities through a self-funded, whether it's through monetization or through different ways of financing the growth. But that would be essentially within the infrastructure business. So we, as Axiata as a group, would let edotco self-fund its growth. As far as convergence is concerned, I think we are fairly well developed on Dialog. So it's a Dialog fixed wireless access business has been growing very well. It would be self-sufficient. It will be like BAU for us.
As far as Malaysia and Indonesia is concerned, we will look at piecemeal organic fixed implementation. However, if it has got to be an accelerated investment in these markets, which is part of our strategy, we would look at more from a partnership, joint ventures, or an inorganic play in the converged play. As far as digital businesses are concerned, I think we talked about it earlier. We are looking at an exposure on digital businesses, which could be to the level of $50-$100 million over three years. Additional investments will come through partnerships, joint ventures, or investments if we continue to expand. It does not mean that we are going to slow down on our investment. It's just that we will ensure that those investments are coming through external sources of funding.
And as far as spectrum is concerned, we see spectrum requirement coming in a few markets, and then a few others would be 5G spectrum and some bit of small spectrum in some of the other markets. So those are the kind of principles we would follow on how we're going to be allocating capital going forward. Let me see how we look at funding the journey for us. Given the balance sheet has and the environment is something which we are very cognizant of the fact, I think we will prioritize operating free cash flow across markets. So we will balance growth and cash in all the markets. And we do see that improving over time as the CapEx requirements in some of those markets comes down over time. We do see EBITDA improvement.
That's going to come largely around cost optimization, which I talked about earlier, and also looking at some of the organization restructuring going forward. Cost, MYR 5 billion. I think we are focused on delivering that MYR 5 billion saving. As I said, the entire MYR 5 billion is not going to go down to the bottom line. A lot of it will get reinvested into expansion on network as well as taking care of the capacities for data consumption going forward. CapEx intensity, as I said, we will be coming down by around 3-4 percentage points from where we are. And monetization of non-core assets is an area where we are looking at. Obviously, that monetization will be determined based on the right timing. It's not that our balance sheet is stretched that we have to do those actions now.
It would be based on right opportunities and right time when we monetize, and then the partnerships to grow our new businesses. So we will continue to grow these new businesses, however, with external investment and partnerships. So how we look at ourselves, I mean, this is 2017. We were at MYR 24 billion, EBITDA MYR 9 billion, and PAT at MYR 1.4 billion. I think we will look at growth at mid-single digit taking us to 2022. We will look at EBITDA improvement around 3 percentage points between now and FY22. And consequently, we should see positive development and profits going forward. From a value matrix, I think Tan Sri did speak about it. What would drive our value matrix going forward? I think a significant part of that value will come out of operational performance improvement across six opcos and tower business we have.
We should see some effect of value coming from the fixed business where we will look at both organic and inorganic play. Infra, we see because tower business, we will continue to keep expanding. We will continue to invest into new markets, which should drive the valuations up for us and new businesses, which would be the other driver for value going forward. So eventually, from a 90- 8 kind of a value mix, we should look at around close to 80% coming from mobile and the remaining coming out of the tower business and the digital businesses. So I think that's pretty much how we look at the portfolio going forward. Now, let me touch upon M&A.
I think this is an extremely important link to the whole discussions around portfolio or capital allocation because we will look at actively M&A, but focus around some of the key principles of why we would do M&A going forward. But before that, before I get into those details, just to give you a perspective on where we were, where we started on our footprint. In 2008, we were present around 10 markets spread across a much larger geography than we are at this point in time. Within the last, we exited out of three markets between 2008 and now, and we have added around six consolidations, mostly focused on end market consolidation. So that's been the key focus for us as it comes to the consolidation activity. We have moved into one new market, which is Nepal.
And the rationale behind why we did that, I'll come to a little bit later. So I think the footprint now for us is largely these six operating markets, with edotco has a much wider footprint, but yet focused around South Asia and Southeast Asia as the footprint. And tower business is one area we will continue to look at opportunities outside. But whereas the mobile business is concerned, the primary target would be in-market consolidation. So let me just put down what does this entail when we decide on the M&A. I think, as I said, our key focus strategy is to be a strong number one or a number one or a strong number two player in each of the markets where we are present.
So we will look at M&A activities to be a number one or a strong number two player, which includes convergence as a play. We would look at majority control. And even if it's minority, it will be management control. And greenfield would be only selective opportunity. At this point in time, there's no intention of moving outside our existing footprint. And I think what is important is to look at what is the regulatory environment. Is it aligned with our long-term strategy? So this is what we look at as a qualitative measure when we decide on our M&A focus. However, there are quantitative measures we also look at when we decide on our M&A. I think it should be earnings accretive.
When we say earnings accretive, it may not be on year one, but it's clearly short-term earnings accretive for us, be it in the form of profitability or a dividend yield. So we are not looking at M&A, at least on the core side, which are earnings accretive on a longer period of time. We are looking at it should contribute positively to revenue and EBITDA in short to medium term. And it should be ensuring that we do stay within our risk appetite when it comes to the balance sheet, which means from a gross debt to EBITDA perspective. And it should be cash flow positive in an acceptable period of time. One of the key metrics for us when we look at the M&A activity is the return on invested capital and obviously the IRR, which is there.
So, in summary, the focus when we look at the M&A is largely focused around the core strategy of consolidation and being number one or a strong number two player in each of the markets and driven a lot from the short to medium term accredited to the group as such. So I just covered this. It's pretty much the same thing which I talked about, but important that this allows us for continuous portfolio rebalance and control over the strategic assets we have. These are some of the examples of the M&A which we've done since 2008. I'll not get into details. But let me just do a couple of examples of why we did some of those M&As. For example, the Smart Hello merger. I think this was really focused around moving from being number three operator to being number one operator in that market.
I think it's been successful in the sense we have by far the number one operator in Cambodia at this point in time. This has also driven a lot of business synergies and cost benefits leading into the EBITDA margins of nearly 50% in Cambodia space. XL Axiata Axis merger was largely focused around spectrum. We did get the benefit out of the spectrum acquisition coming out of the merger. We've seen significant development. I would say a little bit behind on what we had looked at Axis from a performance, but that's largely the industry structure.
Robi-Airtel merger in 2016, again, focusing on moving from being a number two or a number three player to becoming a now strong number two player in this market, which did allow us to acquire spectrum, benefited out of large spectrum, allowed us not to go in for new spectrum auction and convert from a technology neutral at a much cheaper value. I think we are doing far better than what we had set as a business case on this merger. Airtel, which was a loss-making business, has now turned profitable, and we are seeing the effect of that in terms of our ability to take a strong data position in that market. Acquisition of Ncell, I think this was clearly cash accretive for us. We've been able to take out around $400-plus million of dividend from this acquisition.
Continues to be extremely profitable, contributes around 25% of our profit at this point in time. By and large, the focus around M&As has been in line with what our key strategies on M&A have been. From an M&A perspective, key takeaway, it is highly focused around our in-market consolidation and realizing the synergies across the M&A activities we do. It's focused around right investment thesis, which is what I talked about more from a quantitative perspective, that these should be accredited to Axiata. It has to be supplemented with facts and figures. That's basically where we are focused on the M&A aspect.
Yeah, that's it. Thank you. Thank you for me.