#35 - AirAsia Super App Transformation and Lessons from COVID-19 Pandemic - Pablo Sanz

Episode Introduction [00:00:43]

Henry Suryawirawan: [00:00:43] Hello everyone. It’s always great to be back here again with a new episode of the Tech Lead Journal podcast. Thank you for tuning in and spending your time with me today listening to this episode. If you’re new to the podcast, know that Tech Lead Journal is available for you to subscribe on major podcast apps, such as Spotify, Apple Podcasts, Google Podcasts, YouTube, and many others. Also, please check out and follow Tech Lead Journal social media channels on LinkedIn, Twitter, and Instagram. Every day, I post words of wisdom from the recent podcast episode, and I share them on those channels to give us some inspiration and motivation for us to get better each day. And if you’d like to make some contribution to the show and support the creation of this podcast, please consider joining us a patron by visiting techleadjournal.dev/patron. I highly appreciate any kind of support and your contribution would help me towards sustainably producing this show every week.

As we all know, due to the COVID-19 pandemic, the last one year has been challenging for all of us. And in particular for a few lines of businesses, like travel and airline companies. My guest for today’s episode is Pablo Sanz. He is the CTO of AirAsia. In this episode, Pablo shared with me the tough challenges that AirAsia had to go through during the recent COVID 19 pandemic, and how the company has to pivot and transform in order to survive. He explained what the AirAsia leadership team did to realign the company vision and to steer the business and technology roadmaps during the challenging situations while keeping the people’s morale high, which is definitely not an easy thing to do in such a tough situation. If you follow some of the latest AirAsia news and updates, you would have known that AirAsia lately has been doing a lot of business transformations, providing a digital platform, and expanding into multiple new lines of businesses, such as food delivery, shopping, logistics, rewards programme, payment, digital academy, and a few others. Pablo shared with me further about those transformations and AirAsia’s ambition as a company to transform from being a traditional airline company into becoming a digital platform and ASEAN super app. Pablo also shared how he envisioned to build a data-driven product engineering culture within AirAsia, and how it will help AirAsia to continue coming up with innovations based on data driven hypotheses.

I hope you will enjoy this episode. And if you like it, consider helping the show by leaving it a rating, review, or comment on your podcast app or social media channels. Those reviews and comments are one of the best ways to help me get this podcast to reach more listeners. And hopefully they can also benefit from the content in this podcast. Let’s get this episode started right after our short sponsor message.

Introduction [00:04:12]

Henry Suryawirawan: [00:04:12] Hey, everyone. Welcome back to another new episode of the Tech Lead Journal. Today, I’m so excited to have with me, a guest from AirAsia. He’s the CTO of AirAsia. His name is Pablo Sanz. So, as we know that the last one year or so, airline industry has going through a tremendous difficulty. What AirAsia has shown in the last few months, especially from Singapore here, I can see a lot of things from the news, AirAsia has transformed a lot in terms of not just doing airline, I mean like really just doing, cause they have been doing such a number of things. But transforming how they do their digital transformation, their application to become a super app and going into new digital services. I think it’s pretty remarkable. So today I hope to hear a lot from Pablo on their journey in the last one year or so in order to transform their technology team and come up with all these new services and the digital transformation that they went through. So Pablo, welcome to the show. Very excited to have you here.

Pablo Sanz: [00:05:07] Good afternoon, Henry. Thank you for inviting me.

Career Journey [00:05:09]

Henry Suryawirawan: [00:05:10] So Pablo, in the beginning, I would like to ask you to share your maybe career journey or introduction about yourself. Maybe you can mention your highlights and maybe turning points in your career.

Pablo Sanz: [00:05:19] Let me start by giving a small correction to make justice. On AirAsia, we have two CTOs. One takes care of the airline operational stuff. So he’s more responsible of bringing you from A to B and making it in a safe way. And we have another one in charge of the super app. That’s myself, and I’m more responsible on the customer journeys on the app and on the purchasing experiences, and all the client-facing application. So just to be a little bit fair, Vishnu takes care of the airline, so he’s also involved. A little bit of the highlights, I’ve been doing this work almost for 9 -10 months. I was previously the Chief Product Officer only, and now I have product, tech and data responsibilities. In the past I was a consultant for a Spanish company that started first doing kind of horizontal industry agnostic projects. But I joined the airlines division, and it was a pretty cool division within the company, because in a different way than at the telcos or the banks where the way it works, they were only selling meat, basically time and materials. In the airlines, we had the capability of building products. So based on the first projects and experiences, we started building products.

I think it was a very natural path from engineering to product building, and I did sales engineering also at that same time, and I was consultant for multiple airlines before joining AirAsia. That’s the path. By training, I’m an electrical and electronic engineer. But I think I don’t remember much of that. And then I did an MBA with focus on international trade. That was a public MBA with a sponsorship of the Spanish government. So, humble MBA, let’s call it. I think I have both angles. I can connect the business with the technology in a pretty comprehensive ways. So I think that gives me a unique perspective.

Changing Career into Tech [00:07:08]

Henry Suryawirawan: [00:07:08] Thanks for sharing your story. I’m quite interested because so many people these days also thinking of switching career, right? So maybe you can share from your point of view, from your journey from being an electrical engineer into doing a lot of technology, maybe building products, involve in technology software development and things like that. Maybe you can have tips for some people who are thinking of changing their career?

Pablo Sanz: [00:07:29] Yeah. I think we have a lot of examples. Actually, in AirAsia, we are doing cool stuff in that way. We have a partnership with Google actually called RedBeat Academy, where we have our own General Assembly entity, that started to retrain pilots, cabin crew, and our operational staff into digital technologies to bring them to the avant-garde of their jobs, and allow them to change jobs. As a curiosity, I had a chat last week with a friend of mine, has a startup in Spain that tries to help high school people to choose their career paths. He was very interested on this topic also because we have people that managed to change careers, and adapt, and find their place in the tech world, where coming from careers that had nothing to do with it. I think you knew Declan that used to be our CIO. He was an anthropologist, then he became a CIO. We have in our office in Singapore, Wanzhen, was VP of products for banks. And she decided to drop it, and go General Assembly and do engineering, and now she’s working with airline data teams. We have people from psychology, and so on that in my country at least would be mainly unemployed, doing user research on the research team reporting to Nathan.

Yeah, we have a lot of transformation on those areas. We are hiring currently, one intern from RedBeat Academy to join also the design team after 2-3 months course. We are doing a lot of those things. We are transforming data analyst that previously had basic SQL knowledge into product managers and data scientists. Even like I have someone in my team also, Mahisha, that was signing an NDA on Monday to get access to data, to do sentiment analysis on our NPS scores. So there is an evolution. I don’t think what you study defines where you have to go. There are different paths to reach different places.

Henry Suryawirawan: [00:09:19] So it’s very interesting. I read and heard about RedBeat Academy as well. I’ve watched several times on the social media platform, like LinkedIn, where AirAsia staffs has transformed themselves from doing maybe for example, air stewardess or pilots, or used to fly around. Now that there’s no flight available, they transformed themselves and went through this. What will be the thing that they should focus on? Like for people, they want to go through this tech career, but they just started, and went through like General Assembly or RedBeat Academy. What will be the things that they should focus on in order to succeed?

Pablo Sanz: [00:09:51] I think it does not apply only to them, right? Also, when we select interns. I was having a conversation yesterday, and it’s like how tough my guys are being with the interns on the interviews. I just want them to be willing to learn. Probably to have the patience to learn the two, three years in order to assimilate all the concepts, and to find a company that has the patience to groom them and bring them up to speed. Because we are in a world that churns people very fast. I think we need a little bit of patience when training and when mentoring these people, because someone had patience with us in the past also in order to allow us to grow. So I would have two advices. One for the employers that to have a little bit of patience and invest a little bit. And the other one for the people that wants to change career, to have the patience to learn, and not to try to think too much about the compensation and so on. I think as far as you are learning and you’re progressing on your lane during the four, five first years in the career path that you think that you want to go, don’t bother too much about those things. That will pay off along the way. But you need to invest a little bit also, and if someone is giving you the opportunity to learn, appreciate it and take the opportunity because sometimes it’s not easy to get those opportunities. I think the two angles are needed in order to progress in that direction.

Henry Suryawirawan: [00:11:10] Fully agree with you. The only thing is the willingness to learn and be patient, and not to think too much about compensation, or the status, and things like that.

AirAsia Recent COVID Challenges [00:11:18]

Henry Suryawirawan: [00:11:18] So in the last few months, or one year or so, I’m sure internally within AirAsia, you guys also have a lot of challenges to go through. Personally, from your point of view, from the technology side, especially end product, what will be some of the challenges, if you can share?

Pablo Sanz: [00:11:32] Well, you have collaborated with us in the past. When we were an airline, just migrating our tech from A to B, and you helped us set up some CI/CD pipelines. I hope some of them are… I don’t hope they are there still, Henry, just to be clear. I hope they have evolved. We had a lot of systems and we migrated a lot of system, but we were an airline. So, before the pandemic, we started exploring this idea of becoming more than an airline. I mean, it was different business, but they were some way disconnected. So we ended up creating a Frankenstein. You had probably 7 payment gateways, integrated. When someone would approach you to see, I want to do a 10 Ringgit of campaign for Visa. It would be, wow, four months delivery date. Because you need to go there regarding to agreement.

So one of the technological opportunities that we have taken during the pandemic has been transforming the mindset to thinking on products, to thinking on platforms. And I think that’s the foundations of the super app. I’m not a hundred percent sure what being a super app means. But I know that it’s not sustainable if you don’t think on building it as a platform and reusing tech pieces and customer journey pieces across your flows. I think we are on, between alpha and beta of being a super app. I think you guys are going to see a lot of cool things happening in the next 3-4 months. You are going to see more native-like experiences. In our parts, collaborations with external companies. We launched a mini app in our ecosystem last month with Naluri. That is another startup in Malaysia to give capability of tracking food journals. It was a very small use case. But it tallies very well with our health vertical. You are going to see things that are going to allow collaboration and going to make us believe more that we can call ourselves a super app. So we want to progress and increase the ecosystem.

Another lessons learned is that we have a lot of valuable people in the company. I don’t know if I would call it family, but I would call it pride, or common pride at least. We went through very tough salary cuts and hard impact, and we had attrition. But there are people that stayed and valued what we were doing, and was excited about continue learning and the opportunities given. We have suffered quite a bit. We continue suffering a little bit in Malaysia. We are hoping to go back to normal and be able to fly. Everybody pushing together into the same direction, and collaborating, and doing cool stuff, and having more platform thought. I think the main takeaways from this 13 months are it’s very hard to say. It’s painful.

Transforming Into Super App Platform [00:14:03]

Henry Suryawirawan: [00:14:03] Thanks for sharing that. You mentioned you guys have to transform from different disparate systems that not talking to each other, different products, and becoming a platform. So in terms of technology leadership, what do you need to do in order to make this happen? Because there are so many things, so many technology legacies, so many people as well at the same time with not so well morale. So in terms of technical leadership, what did you do in order to push forward and make sure that this thing happens?

Pablo Sanz: [00:14:30] Before pandemic started, we had two kind of projects, the urgent and the important ones. Unfortunately, we needed to stop. So that was my division, and that was my success recipe. I wanted to do the ones that were important, perfect, no discussions, needed to go live, be scalable, super fast, 130,000 concurrent users from the day one. Everything, right. We needed all that. No failures allowed. But then we needed to migrate many more projects into the important or the urgent bucket to preservation. An example was Unlimited. That was a successful on facade, but product and technology wise was a disaster. The first weekend that we had it, we had to patch a lot of things. We had to manage a lot of the customer expectations. Because we took a Mickey Mouse product and we use it to do Unlimited. Now, the last ones have been much better, and it works pretty smoothly in Thailand. Indonesia is going on this week, actually. We don’t do it perfect yet, but we are on the path.

I think we took two or three projects and we kept them in the important bucket. One was Order Management System, a digital order. Don’t think on it as an e-commerce platform, but our order was captured in our airline reservation system till then. That was Navitaire. So when you would buy something, Navitaire will be the single source of truth to what you have purchased. So we needed to decouple from it. And it was convenient because it allows you to automate refunds, automate the discussions with the client. It’s pretty convenient. But then you were having all these payment gateways, hosted ones, non-hosted, some of the guys using Stripe, using different various flavors of Adyen, another one of our providers. So we try to unify that. By 15th of April, we will only have two Order Management System, one for flights, and another one for the rest that was key. And then having Single Sign-On capability that would allow us to track across the ecosystem. Anything and keep the customers logged on, and recognize them and award them with points, and keep their saved cards to perform payments in a seamless way was also very important.

Third is where two years and a half ago, we had to take the decision of wrapping our website to the app. We use PWA, Angular. I wouldn’t say that was the wrong decision, but because we were suffering of feature parity in our app. So we started undoing it. We started building flows that were either native or native-like using React Native and reusability of the components across the platform. We did it for SNAP. We did it for hotels. We will go live this month hotel with flights, in that same condition. For other airlines, not for our own flights, that will take a little bit of a while more. We integrated our React Native experience with Naluri, from a third party that I think is what really defines super app which have those ability of all there is to collaborate in your ecosystem. So we give them SSO, we allow them SSO credentials. In the future, we will be building Order Management APIs, and anyone that wants to come to our ecosystem can. And we are using the same principle in order to build e-hailing. We are partnering with someone that has an application in React Native, and they will sync with our Single Sign-On. And in the future, we will integrate our payments so everybody will benefit for all the features in payments in a single manner. All the payment methods, the saved cards, everything will be unified. So I think those are the two, three main characteristics.

I think the one missing, and that we are working on also is the chat capabilities. I think we grow a lot. So I think payments, sign-on, and chat are the characteristics that all the mini apps have. We want to add what WeChat does in a much smaller scale and control. But we want to be able to partner with a startup that have cool ideas to give them also an ecosystem. As part of the community, we want to create a developer community around us. And those probably four angles are the ones that we have worked during the pandemic most to create a super app ecosystem.

Aligning Vision [00:18:31]

Henry Suryawirawan: [00:18:31] So in terms of all these changes that are happening at the same time, I’m sure one of the responsibility of a CTO, the leaders within the company, is to set a vision. Aligning everybody that, okay, you have so many things that you work on before, and now because of this urgency, from now on you need to focus on this, maybe try to unify things. So what are some of the steps that you do in order to align the vision and make sure that everyone went through the process and do exactly what is expected to do in terms of the company?

Pablo Sanz: [00:18:58] I think from product, we have been trying to anticipate more and more of the roadmaps. From engineering, we probably would have taken more of a people approach there on how to reshuffle people. So we have created bigger tribes in order to be more optimized, enforce collaboration, and be able to pivot faster into some of the needs. So before pandemic, you will have flights team, hotels team, a SNAP team, and so on and so forth. Now you continue having them, but they all pertain to the same uber tribe because it has more than one manager, probably. So there is more flexibility of moving resources and sharing technologies across them. For the data analytics, probably we have morphed into more specialized way into the verticals. But for data engineers and data scientists, we are trying to create as many platform level products as we can. Like NUDGE that is our recommendation engine, it’s now being plugged across the ecosystem. The head of that project mainly is the seller of his services. Because now he can recommend you for food, for fresh, for shop, and he can test different algorithms. But the APIs and the contracts have been standardized to be able to plug it in an easier way. So revenue management has become also a platform game, and experimentation also is in that same sense.

Also, not during the peak of the pandemic, but now we started also a project called Minerva, where we are trying to re-platform the data engineers to become machine learning engineers and data engineers. So they can broaden their scope and we can bridge the gap between data scientists and data engineers, so they can help each other. We did a pilot where we train them, and then the data scientists will monitor it for two weeks on the progress and their doubts. A little bit of homemade way with our current resources, but we did a pilot. And now we are trying to look at our first use case where the 7 data engineers that were part of the pilot should build a data science use case together and put it into production. So we are trying to bridge that gap also and transform them.

AirAsia Data Platform [00:21:02]

Henry Suryawirawan: [00:21:03] I’m interested in the data story that you have here because, as we all know, data is very important. And especially when I read some of the AirAsia news, as a company, you emphasize a lot to actually make use of this data that you have. You have been around for many years. Everyone has transacted mostly on digital. So data is one of the important piece asset in your company. So what do you mean by building platform? Because data has been around there in the AirAsia, and what does it mean to build a platform? And how would you foresee in terms of roadmap, the data being used in AirAsia?

Pablo Sanz: [00:21:36] Yeah, I think the data is there. So the trick is how you efficiently create user experiences around that data. You can do it very fast in silo use cases, and go one by one and not reuse anything, or you can try to think on common use, common algorithm, common pieces that can be reused across the ecosystem. And don’t have to do anything with the intrinsic value of your data, right? The data is there, and the value is probably extracted by the person that is looking at it, or by the machine that is looking at it, or by the engineer that is building machine learning algorithm on top of it. The data itself that has the value that you want to give but you need to extract it. And in that sense, we have, I believe, relatively valuable data. We are trying to increase the frequency of that data. So that’s probably more on the product side of the things that we are doing. You realize e-hailing, food, e-commerce have a higher frequency than our traditional business. Our traditional business has higher shopping cart. So we are having the reverse problem as other super apps in that sense. So we make money. Without pandemic we make money, but we are increasing the frequency.

I think we are in a good path to do it and all, but then in order to collect that data, we needed an automated way of tagging and tracking and so on. It was not sustainable to continue having the typical GA analyst that will go, and will write requirements for every product and every new feature that you write. So we have moved to a self target environment. For instance, we are pushing all that data into the platform. So that’s a platform game. If I can tag product by product, but then I’m being highly inefficient. That was one step that we have taken. It’s very coupled in our case with our design system. We did a design system that we called Phoenix and a publishing platform called COSMOS. So anything that flows through there, it’s self-tagged. Any of the new flows are self-tagged in any technology that we build in. So a Baton is tagged as Baton and throws an event. And you know that if it comes from this to this page, you have a click through. You don’t need to define every time what that click-through is. Then when collecting, we were in an environment where ETL batch processes were, okay, because you fly, you have 14 days lead, you fly three times a year. I don’t really need much real time on that sense. We needed to invest also in a little bit of a streaming capabilities, and move from T+4 hours into something much more real time.

That also impacted probably some of our architectures. Because before we were throwing everything to BigQuery and then consuming it. If you do that, you cannot be real time. There are different solutions, but changing the mindset that “everything needs to flow through a single point of truth” to “it needs to be a single point of truth, but it can flow to other places before reaching the single point of truth”. That was also something that we needed to do. And then connected to that, we needed to start predicting intents of our users. At some point, we realized that we have three different teams create an intent prediction algorithms. One for flights. So if you think it is a product, you need to extract commonalities, and then personalize. So we chop it into predicting what is your intent and predicting what would be your intent for a defined business? We have intent prediction platform that we call Natch, that decides what I should show you next. What I should push you next. What I should try to sell you next. What should I inform you next. But then, if what I should show you next is hotels, the hotels platform decide which is the hotel that they want to sell you.

So we needed to decouple some of those to take advantage of reusability and of verticality. So we needed to tackle the problem in two different pieces. And plugged into that problem, we built also the experimentation platform moderator. In order to be able to stream real time also to get flags real time, we used to have experimentation, and we have been doing experimentation for very long time. Injecting GTMs. We call it Red Optimize. But we needed to create a platform that would allow us to do A/B testing in a sustainable way. So we can move from a 100 a year to 1,500 or 2,000 a year. So that was also an investment in a platform.

And also two more pieces on the data platform. One on the revenue management side, where currently applies for flights, but there is plans to make it platform wise so we can apply for hotels and for packages and so on. And even be a discounts, probably for e-commerce. And last piece where we leverage a third-party tool. That is on the growth side and the ecosystem where we use CleverTap. So our algorithms and our intent predictions can target the users using CleverTap to communicate at different points of time with it. So if you buy a flight, we know that we’ll need to take you check-in, that you have a baggage. You need to know where your belt is. And probably before all of that, I needed to try to upsell you a hotel because you had a flight at a preferential price. We have some of those flows that are manual because the partner has a journey builder. And some of those flows that are triggered by APIs through and connecting to Natch with the predictions. So basically, that’s the data investments on the platform side that we have had to do.

Aligning Technology Roadmap with Business [00:27:07]

Henry Suryawirawan: [00:27:08] So with all these in parallel running tracks, I’m trying to understand a little bit, so during the crisis you have from the business side, business point of view that you need to survive, create new revenue generation. But from the technology side, you also have your own vision, how this could be done. So maybe you can share a little bit. How do you actually align from the business point of view and the technology point of view? And make sure that come up with a consensus that this is the option, or this is the roadmap that we need to do in order to survive together as a company?

Pablo Sanz: [00:27:36] So I think it comes with very frank conversations between the business and technology teams. Like I think cash was king for us, and we move a lot of initiatives in order to preserve. So, Unlimited become key during time in the pandemic in Thailand, in Indonesia, in Malaysia. Besides the international one that is on hold at this time, the rest I think work pretty well. And then refunds and customer service in especially first 4-5 months, we had a lot of work on it because the customers were suffering almost the same as us. So we needed to automate a lot of things on there. But at the same time, the pillars that I was talking to you before were non-negotiable. Creating a unified checkout as a service with Order Management System, it was needed to recover and to build another business lines in a sustainable manner. So it comes from conversations, collaborations, understanding, and ability to explain what are the trade-offs? What we can do?

Also, as I was telling you, we put together bigger teams in order to pivot more. An example is Wi-Fi. At around November, December, we joined the Wi-Fi team. We embedded it into the DevOps and e-commerce tribe. So they could learn new technologies. Probably we were a little bit slower for the first 15 days to one month, but they could learn. They could help each other because on the Wi-Fi area, next six months, we wouldn’t have much work. So we created a bigger team that still have their knowledge and their specialty. But they can collaborate a little bit and help more to create additional bandwidth to tackle this crisis. The same we did with activities and we put it together with Order Management System. Probably that was a little bit less of an evident fit. But yeah, we have been doing these things of moving teams, making them bigger, so they can collaborate better. I’m giving to the engineering managers, a little bit of tools in order to balance their workloads and tackle more dynamically priorities.

Keeping The Morale High [00:29:35]

Henry Suryawirawan: [00:29:35] So I like the way that it’s being tackled. Like you have a frank conversation, and Cash is King, like you said. The urgency and the priority for the business is the first thing that is very important. So I understand as well, like you mentioned earlier in the chat, some of the staffs went through all these tough challenges. And some even had to go through like salary cut. So how do you keep the morale within the team? Especially from the technology team. Make sure that people are aligned, and the things that they do still matter, and went through all this reshuffle. I’m sure, like it’s not easy to keep people motivated at that time.

Pablo Sanz: [00:30:07] Yeah. It was not easy. November, December were couple of months that were very tough for me. I had examples in that this is not my first crisis. When I came out of the university, we have Lehman Brothers in Spain, and it took us 4-5 years to recover. And I was 3-4 years with zero salary increases and whatnot. So, I had an example and I use it multiple times. My cousin was an architect, building architect, not the tech architect during that time. And he worked in a very small company, 12 people in Madrid. Architecture was in a very deep, long tunnel at that time. The guys got their salary 40% for three years, but they didn’t fire anyone. For me, it’s a very inspiring story. I think that it shows a lot of commitment. So, I have to confess that when we got the pay cut, I texted our bosses and I say, thank you. The first time in March. I think that was the responsible thing to do. It was more difficult to explain it then to the rest of the people. And then we took an approach of, I try to be very transparent and give a lot of updates. Probably over transparent in some areas like I have a meeting every two, three weeks that from time to time I cancel, I have to acknowledge what I present the sales to the tech team. So they know how are the sales? How are they doing? They know if investment has come in the last placement, in the last month. I needed to be transparent. I didn’t see another way of doing it.

And then also we try to recover those salaries in a bottom up approach. I think we try to take care first of the people that have lower salaries. I think that a lot of people understood it properly and thought that was the correct thing. Probably others didn’t think the same way. They thought that they deserved to be rebalanced before. But taking into account the cash flows and so on, we started recovering all of us, our salary. Basically, that’s how we dealt with it. I’m proud of how we dealt with it. It was not always good. It was complicated, and it has its downs and its problems. But I’m very thankful for the whole team to have resisted with us because it has been a very tough episode. I think we are in a much better situation and we see the light at the end of the tunnel. I am very thankful for the whole team. I think we could have not done anything without them.

Henry Suryawirawan: [00:32:24] Thanks for sharing that. I think it’s pretty well done in terms of how company is doing that, and making sure that not many people are really suffering, and show that the people who needed the most actually the first one to be taken care of. Rather than having some firing, retrenchments and things like that.

Pablo Sanz: [00:32:39] Yeah. I think you have seen unfortunately in the airline, some of those things have had to happen. But you have seen that we have crew becoming drivers. We have crew becoming influencers for some of the brands. We have a responsibility also from the tech side. We have a responsibility with all the flight attendants, all the ground staff, and so on. I think Tony says it a lot of times. We want to rehire as much people as possible on the airline side. That’s something that I put on myself also from the tech side. We need to help. We need to be efficient and run and build these awesome products so we can bring them back on board. So that’s one of my missions for the next year, probably.

Henry Suryawirawan: [00:33:22] It’s pretty awesome mission. I wish you all the best!

Critical Things of a Super App [00:33:24]

Henry Suryawirawan: [00:33:24] So coming back to the super app. We mentioned a little bit about super app a lot of times. So what do you think are the critical things to consider for the super app to become a success? Because in this region at least, there are a few players that are trying to be a super app. What do you think are the critical things?

Pablo Sanz: [00:33:41] I think we need to find our way to create frequency, as I was telling you. I think we have a good opportunity on e-hailing, and also we needed to find a way on scaling to all the countries. I think we are on that path. I think you are going to see new experiences. So Food is in Singapore and is in Malaysia. You will see it expanding to the rest of the ASEAN countries. Shop is already on the Philippines and Malaysia, and you will see beauty and shop appear also in Indonesia and Thailand in the next week or so. Food is getting a new platform. So we hope that experience is better. Then we are rolling out e-hailing in Malaysia. Obviously there, per country needs to be a place where we compete because I think we have some key advantages against other players in that.

The expansion and the frequency are the two keys. Because in a normal situation, we have the average basket on the profitability that others don’t have. The increase of the petrol price makes them raise their fares by three times. I think you are experiencing it in Singapore these days. We have some tools to bring stickiness. I think Unlimited is going to be key. Also on that, stickiness, frequency, expansion through ASEAN I think it’s going to be key. Now, I think we already have a technological shape. Like we have our checkout as a service. So when we try to build something additional, like a discount or a voucher of cash scheme or points or other payment methods, we can expand it in only one deployment. Before, we needed seven or eight. So, the same with e-commerces and so on. It takes us now 10 days to spin a new business that is e-commerce related. It took us probably six months before. So, getting there was important, and now it’s a matter of finding ways to compete with others, and finding unique value, unique propositions in some of those areas. The same way that we have in travel that brings us frequency and a stickiness.

Integrating with Partners [00:35:38]

Henry Suryawirawan: [00:35:38] So you also shared a couple of times during the chat, like you partner with a lot of other third party or companies or even startups, in order to build this platform and the super app. What do you think are some of the key things in terms of technology, when you deal with so many different partners suddenly coming and trying to build a unified and stable platform?

Pablo Sanz: [00:35:58] So I think that has ups and downs. Some of them try to change the way we operate, and we need to be clear on that. We are trying to, and I hope that we can publish it at some point, where the people just can come, register and get access to our APIs and our checkout flows, and use us almost as they will use Shopify. Register, 1-2-3, load products and go. But I think we are clear on what we want from them. In this case, it’s a more of a collaboration, and what we offer like Single Sign-On and checkout, in the medium term are not negotiable. I think need to go through a unified platform. So that’s more on the collaboration with startups, or developers. When we select vendors or we collaborate with shops, we try to find a cultural match and commitment. I like small to medium companies that want to partner with you and want to develop their products with you. Because I think there is a win-win situation where you benefit from them, considering you unique on their business and you get that benefit. On the other side, they get a free product roadmap, right? They don’t need to think if it’s going to work, or it’s not going to work. They just need to think if they reuse it or not, because the use cases come from us.

So I think we have had two or three of those that work very well. Basically, that’s what we are trying also on those mini apps. We set it in React Native because it’s the technology. We have this discussion between Flutter, React Native, and so on. We saw in order to enable this for third parties, the CodePush capability of React Native as being a key on the decision, and the availability of JavaScript developers also being key. We have always this discussion on which technology is better, and also everything that we do, I hope that in the future becomes native even. But to make it sustainable in the short term, and to partner with others, we needed bridges and frameworks that will allow to collaborate native and React Native.

Henry Suryawirawan: [00:37:56] So what’s your take in terms of security and the data that you are sharing with these parties?

Pablo Sanz: [00:38:01] Basically, all go through our multi-domain SSO. So all is tokenized and so on. We sign data privacy contracts with them because at the end, they are part of our ecosystem. Don’t think about it as a redirection where I pass the information to someone else. It’s part of our same ecosystem, and they are mini-app on AirAsia, right? So, I don’t think there is many issues in there because we treat their security and the data privacy as any other app on AirAsia. I was talking this morning about how to automate some of those things on the CI/CD pipeline. So, when they push a new package on React Native, all the CI/CD evaluates those different vulnerabilities, regression tests, critical use cases. So, the AirAsia team will only have to do are CodePush validation. And also, we are trying to save the quarterly, full merger of the code without CodePush. That’s part of the plan, so it becomes more automated.

Frequency of Deployments [00:38:59]

Henry Suryawirawan: [00:38:59] So on this topic, CI/CD, maybe during this pandemic time, if you have some data, if you can share as well, how frequent has AirAsia transformed from the past? Before pandemic and up to now. Like how frequent the deployment has been? In terms of quality, how do you ensure the quality in terms of the delivery to the super app?

Pablo Sanz: [00:39:17] I think we found a balance. We deploy and we have web hooks telling that we deploy whenever we want. But having daily deployments also puts a lot of pressure in the teams, and the sanity tests then it’s not very efficient, and also makes the team behave a little bit more Kanban, so makes everything more disorganized. So I think there are teams that work in one week, two weeks or three weeks cycles in general terms. Three weeks cycle was… it’s mobile team, currently, is the exception, but the rest normally work in one or two weeks cycles. And it’s just a matter of being efficient between tech and testing, right? Because if not, you start creating bottlenecks and so on. Obviously, there are some sensitive pieces like payments where that is a little bit more tricky. But basically, that’s how we have evolved. We also, you know it perfectly because you were there, we evaluated like from the Google native tools to GitLab to a full set of them. And now, basically it’s a unified way and we are trying to build it in the new businesses. Because now we have a design system also on the UI. The automation of those test pyramids and so on, on the UI becomes also a little bit more efficient because there are test cases that basically can almost be running in all the LOBs. So, the sanities a little bit this year, and the monitoring a little bit this year. But also the expansion of LOBs means that we are catching up all the time. So that’s how we have evolved.

Data-Driven Product Engineering Culture [00:40:45]

Henry Suryawirawan: [00:40:45] Before this recording, you mentioned a little bit also that your focus is actually to build a data-driven product engineering culture. Maybe you can share a little bit, what do you mean by that?

Pablo Sanz: [00:40:54] So, I’m in the process of doing the OKRs at the end of the Q1 on Q2. I mean that I want goals that are quantifiable. So data is key for all what we do. They are aligned because there is no product without engineering, engineering with a product, neither product without data. Engineering and data might be able to live separated but shouldn’t. The thing is, we still do things by gut, and we are not still purely data driven, and there are market in actions and so on. But my mission is that all those three pieces that are in the same mark, have common goals. So I have all the reviews of the OKRs of the leadership team and then beginning of next week. How I’m trying to change that mindset from writing a super beautiful algorithm in R that does this prediction with machine learning, to verbalize in us, I want to sell 2% more. How? By implementing these super beautiful algorithm in R that does this.

So, it is a change. It is a unification and organization. But we have very successful use cases in flights and homepage and hotels, where the teams are collaborating and have common objectives. It needs to be aligned. That’s within my organization, but all flows from the business objectives. So it is outside of the organization funnel, how it starts flowing. I was toying around, and calling all of them data-driven product engineers, no more data scientist, product engineer associates, product manager, engineering manager, all of them the same title. I don’t think they would like it. But yeah, it’s a fantasy. I’m thinking what will happen, but at least their goals should be aligned.

Henry Suryawirawan: [00:42:34] It’s also a little bit mouthful, data-driven product engineer.

Pablo Sanz: [00:42:37] Yeah. But product engineer is easy, but you need to plug data in, if not, it loses all the sense. But default product engineering should be data driven, but it doesn’t happen always. So probably it was a nice remember.

From Gut-Based to Data-Driven [00:42:50]

Henry Suryawirawan: [00:42:50] So, I’m sure a lot of companies, also startups, for example, they operate by gut, right? They don’t have this kind of a data-driven mindset. They don’t have this sort of OKR maybe from the top that, we have to work based on a common goal and then a set of data as a verifiable metrics. So what do you think are some of the challenges, transforming from gut based decision, maybe from the business, from the technology, into this data-driven? I know this the gap sometimes can be quite a challenge to overcome.

Pablo Sanz: [00:43:16] Yeah. I think I have had this discussion so many times. I mean, we still lack on gut. We cannot be fully data driven. If not, it wouldn’t be any value on hiring a smart people. I think it was Sujit, our ex-Head of Engineering that say that, I went into this interview. The answer from this guy was everything A/B tested, A/B tested. And then he replied, why we need to hire you if I’m going to A/B test everything? He yearn have safety net all the time. Why should I hire you? So like always, there is a ratio on the things, and there are facts. If your website performs faster, you are going to have better conversion. Maybe there is a couple of edge cases, but in general terms, that’s a fact. I mean, that’s a fact that this happened. So there are some universal… less clicks 90% of the times, and that was a little bit more tricky than the performance, but drive you to a better conversion. So, it is a balance on gut and data driven, and the data needs to support your gut. So the fact of having gut, that doesn’t make sense.

So, we are evolving to increase the percentage of time that we act in data. I’m going to put an example with Tony. The other day he came and say, “I want to change all the icons in the homepage to red instead of multicolor”. It’s the second or the third time that we are toying with it. I say, “Sure, but I’m going to A/B test it.” Then we did it, but with one exception of some icon that was already in red, and it was also red in the B version. All of them had better performance and better click through on the homepage. So I went and tell Tony, “See, the data shows that you were right.” And I said, “You didn’t need to A/B test it to tell me that I was right.” And I was like, “You’re not helping.” So, you still need some gut, and I think it’s good to have gut and, that comes with having people. If not, we all will be robots and it wouldn’t work. But we need to progress. And I invite those startups, as we were saying, to little by little move themselves to more A/B tested hypothesis, and more data points and so on. That allows you also to take bigger risks, but allows you also to capture a bigger value in some of the aspects that you didn’t expect. Like the low hanging fruits, they are there. They are easy. By a little bit of observation and data analysis, doesn’t need an automation of data, right? By going to the top five sold hotels, and put them the first, you’re going to sell more hotels. But then that’s not scalable.

I think that my advice would be first, you have guts on the low hanging fruits, then data analysis already helps. And then, once you go on beyond that phase, you need to start investing in A/B testing and data science use cases. If you don’t have those two first pillars evolve in there, probably doesn’t allow you to capture the value that you expect from that transition. So it is a path. And then the last step is increasing the number of times that you do A/B test against the number of times that you act based on data analysis or gut. I think there is a path. It’s not about having MVP and a North Star, it’s about how you go from the MVP to that North Star sometimes. It is not so much about what do you want to become or how you want to start, but the when and how you do the changes. So just make sure that you are ready before investing in those things. Because if not, you can get frustrated.

Recent Interesting A/B Tests [00:46:30]

Henry Suryawirawan: [00:46:31] So on A/B test, maybe if you know some of the A/B tests that you have done lately, which you think was like a surprising things? Maybe during the transformation into super app.

Pablo Sanz: [00:46:41] Yeah, I can do two or three. That’s one on the colors. So basically when we move everything to red. The ones that move from orange to red, and from black to red, they automatically had a higher click through. The ones that went red to red, mostly were flat, but a couple of them suffer because their non-red brothers took over. That was one that we did. Now, also we are toying between a right to left scrollable tiles in the homepage and grid. Grid was what we had, with show all. And we have now on a scroll up wall without show all that gives more real estate to do other things in our home screen. We are A/B testing that at the moment. And we also A/B test the capability of someone that starts searching for flights. We A/B tested this only in Thailand, to give them the option to book a SNAP that has a higher shopping cart. That showed that the people is interested in SNAP. So it was I think 25% of the people click on SNAP. So they understood the value. But then, it brought us a higher drop on the comparison because the shopping cart of a SNAP is higher than the hotels. There are more dominant players on the hotel space. So the people needs comparison before doing the purchase. So, that’s three of the last ones that are on top of my mind. A fourth one and very small. It was to test the SNAP. Allowing baggage or not baggage because baggage adds an API call that inject some latency into all the flow. So eliminated baggage brings lesser shopping cart, but maybe more comparison. So that’s one that we are testing also in the next days.

Embracing Innovation Mindset [00:48:19]

Henry Suryawirawan: [00:48:19] Wow. Some cool A/B test cases that you have there. Thanks for sharing that. So AirAsia I know is well known for a lot of innovation even before the pandemic and all these super app journey. So you guys have gone through online, maybe one of the first few airlines within this region that does so. In terms of budget airline, you’re also quite high-tech. I still remember at some of the airports, I can check in through face recognition and things like that. What do you think are important inside the company and the tech division to embrace this innovation mindset?

Pablo Sanz: [00:48:46] I think we have been all the time innovating. And that’s on my debt list, probably. Because of all the pandemic and the urgency matters, I don’t think we have clicked with our innovation framework. So I don’t think 20/80 works everywhere. I don’t think it’s about resources necessarily. It’s about having people that is motivated to try new things, and a space to innovate. So if you invest 10% of their time, they probably put 10% extra because they are extra excited to do that. We have done two or three pilots. Like universal share just started like a pilot. It was more on the design side of the innovation, and we are working on it in the next quarter. The discount engine was a project that I threw. I have a dream to having shopping cart as a service, not only checkout as a service. Then, COSMOS was something that grew naturally, but it was incremental innovation. That is our publishing platform.

So there is a sparks, and it happens. I tend to don’t do that innovation based on buzzwords. But in use cases that are attractive. I don’t think we have clicked still the framework across the organization to make it systematic. I have ideas, but I have been playing, and sometimes someone drops the ball, and it doesn’t work in formal innovation. Sometimes a recognition is expected, and it’s not still in shape. This year we will have a framework, I’m pretty sure. Probably we cannot afford an 80/20, like you guys. But we will have some based on incentives or recognition, or contribution. Like I have a friend in Amazon. In order to become a Tech Lead, I don’t remember exactly the role, you’d have to have done an innovation that is implemented across X percent of the products of the organization. Those kind of things are going to take shape with us. But I don’t think we have clicked still the one that fits on our culture.

Henry Suryawirawan: [00:50:44] I’m sure AirAsia will keep innovating. I’m very excited to see the changes that are upcoming in terms of products, in terms of business, and also the super app itself.

3 Tech Lead Wisdom [00:50:52]

Henry Suryawirawan: [00:50:52] So Pablo, it’s been a great conversation. But before we end the conversation, normally there is one question I always ask for all my guests. Which is about three technical leadership wisdom that you want to share with the audience?

Pablo Sanz: [00:51:04] You gave me heads-up of the question, but I didn’t want to prepare it. So let’s go and try. The first one, I think it’s about people. I learned very early in my career that you can go faster if you go on your own. But you need a team and manage a team is amazing to go further. So we are here for the long run and we want to go further. So I think that’s very important. And we have a lot of people that is grooming the individuals in a super cool way.

The second one is about engineering, I think. Engineering is about problem solving. I cannot say no to anything that anyone asks me because technology can do anything today. It’s only a mean in order to solve a problem, right? Identify the problems, prioritizing properly. There is some guidelines in our company, but basically this microservice has opened us the possibility of having one technology for every microservice that I think is amazing. It should be very exciting for the developers. And even we have it now in the frontend, we have all this micro frontends, and we have a polyglot ecosystem. Now the app with React Native, native and web views. And we choose what is better in every occasion.

And the third one, I’m going to go with four. To take the data driven decisions as much as possible. Build platforms where you can, not stick only to products. In order to reach scale. So, sorry, I needed four.

Henry Suryawirawan: [00:52:23] No worries. The more, the better. So, thanks again, Pablo, for your time. So if people want to find you online, maybe to connect or ask you questions, where they can find you and connect?

Pablo Sanz: [00:52:33] Yeah. I think my LinkedIn is pablosanz, that’s easy, and if they want to reach me, pablosanz@airasia.com . That’s my email. I’m not very good at emails, so don’t hesitate to pass me two or three times. Eventually I will catch up.

Henry Suryawirawan: [00:52:47] So, yeah. Thanks again, Pablo. I wish you good luck with all the journey with AirAsia.

Pablo Sanz: [00:52:51] Thank you, Henry. And thank you for this. Take care.

– End –