Triton Dsouza(Host): Welcome everyone to this episode of our podcast. This is a series of podcasts that Netcore has started as an endeavor, focused on MarTech. My name is Trident D’Souza. I am the country manager for Malaysia for Netcore and I am very pleased to have today as our guest speaker, Mr. Tarun Valecha. So, Tarun is the senior product owner at ‘AirAsia’. Everyone knows about AirAsia as a brand and the phenomenal work that AirAsia has done over the years. Apart from that the Tarun also comes with a very rich experience of working with some of the large big names like ‘Tokopedia’ in Indonesia, ‘Flipkart’ in India. I am very pleased to have Tarun and I’m sure his insights will definitely benefit all our listeners. Welcome, Tarun. It’s a pleasure to have you today.
Tarun: Yeah, thank you Triton for inviting me. It’s a pleasure to be here too.
Triton: So Tarun. We are so glad that AirAsia and you have been a part of this but my first question will be you know today all of us are hit by the covid-19 pandemic. Lots of things around the lockdown have completely hit and impacted a lot of businesses. So I wanted to ask you, “What are your observations around this? How has the market been impacted and what do you see as a user behavior change from this pandemic”?
Tarun: So, I think it is something which is unprecedented. It’s a phenomenon that has caused ripples across the industry. For travel, of course, it has impacted in a huge way. None of the travel companies either offline or online are working right now and therefore they are all gearing up that when this pandemic goes away and people start to travel again, then how they can mostly work towards having a better user Behaviour there and they are gearing up for that. But as a landscape, I think for a lot of companies it has benefited as well. Especially if you consider content companies, a lot of content companies have got a push in their user installs, in their user average user time spent and this is just because people have more time and they are in a lockdown state. Therefore there are a lot of companies that are being used by users who would not otherwise get so much of a push and this is not just valid for specific geographies, which is the main thing right? This has been valid across the world. If we consider places like this, of course in India it’s locked down, in Indonesia, it’s locked down, in Malaysia It’s locked down, in Vietnam it’s lockdown and if you see across all these geographies, the top apps would be the gaming apps. But, then there are certain vernacular content apps, which have also started coming in, which are catering to specific use cases. One of the apps ‘Google classroom’ also got a big push as classroom studies pushed to online learning, which would not have been expected had it not been for covid-19. So I think overall the industry is moving more towards a sort of remote way of life. Apps that were primarily dealing with that have benefited. Other apps will have of course taken a hit and in some time when covid-19 goes away, of course, there would be a recovery period too. It’s not like it would be back to normal straight away. What it means is that simultaneously you have to keep on working on initiatives that retain the users in some way or which have kind of the pull that whenever users come back, they are sticking to your platform for more than they already were. So this is your moment to take advantage of as well.
Triton: Very well put Tarun. I think you have brought down not only the impact but you have put it in a very positive way. Also, you showed what are the positive impacts and very valid points. This is the time to be there with your customers and to retain them so that when things are better, you can reap the benefits. That definitely brings to my next question actually. “So what are the Innovative strategies or any such strategies that you are using right now on user retention, keeping the users engaged and hooked on to your platform”?
Varun: For AirAsia, we are constantly working towards a lot of technicians too so she is now evolving into a full-fledged tech company and therefore a lot of us. Our Focus has been to build a lot of things which are more than just an airline app. Therefore those initiatives are going on and one of the initiatives which we region recently launched was a chat feature. So we have created groups around chats for a lot of popular destinations and soon we will be launching a lot of great upgrades to that as well. I think the plan with AirAsia app as well as that would make it more engaging, make it more user-friendly, you can access some of the features currently on AirAsia app as well and which is something really good. You are constantly evolving, you are also constantly experimenting just like a start-up. On the landscape as whole other apps are also simultaneously doing a lot of similar things. For a lot of travel companies, the focus has been on customer service right now and therefore even a lot of our initiatives are towards making sure that our customers who are suffering because of flight cancellations are able to resolve their queries as fast as possible. Of course, there have been a lot of concerns which have been raised across social media as well. But I think that considering everything, everybody is trying as hard as they can and they are trying to cater to the spike of customers who are now reaching out to them in customer service. Across AirAsia, we have a lot of other employees also volunteering to be customer service and resolving queries. So as to handle the increased load, so definitely it is something which we are all in it together. We are trying to cater to the requirements of the user. So as of now, it might be customer service, but we are planning for the future as well. How the landscape is changing, doing small experiments around a lot of things. It’s not something that would be limited to a few companies. All companies would be doing something or the other which is possible in their scope to do some small experiments because this is the time for it. Normally, the BAU won’t allow you to do a lot more experiments per se because you would have a lot of BAU activities to cater to as well. But simultaneously considering the market scenario right now, I think there would be a lot of things which would be built in the short period of time and which may become or become a normal for user behaviour in the future. Maybe we are going to be using a lot of these destination chats which we have built. For a lot of user interaction later on as well across Asia, this will create a big group of tourists within the AirAsia app which not only benefits us in the short term but even in the long term.
Triton: I have also come across a lot of posts on social media showing this togetherness of how people have volunteered even in customer service and you know coming together and I think very important point which you have put, ‘all the efforts are going towards making your platforms, especially your app more engaging and more user friendly’. So, I think this is the time and you guys are utilizing so congratulations on that.
Tarun: Thank you.
Triton: The question I want to take forward here is Tarun, while you’re doing and putting all these efforts to make it better for your consumers. “What are the most important KPIs today for you to track from the perspective of this user engagement”?
Tarun: So, for us the most important KPIs, of course, some of the KPIs which we definitely follow are currently mostly related to customer service. How many average queries have been resolved, and that is something which we are following or wide because those are the most relevant user behaviors which are there right now. So this goes from the product. How our NPS is holding up, but simultaneously, I think when we consider what would be the KPIs which we would be normally handling if we were not a travel company and if we were a content app, then it goes back to that form the current scenario. A lot of companies are facing the fact that users are much easier to acquire than they were when covid-19 was not there. And this is because a lot of users now have time to explore the apps to go on Play Store just look for which are the apps which they want and trying out those apps and because of this the most important KPIs then becomes that how is your lifetime value (LTV), another KPI, which becomes very important is how much is the frequency of the user visit. Not really the average user time spent because that might not be the KPIs you may be able to sustain beyond covid, but if you are able to develop a sort of stickiness that has become part of daily user routine, then I think that can sort of extend even when these current covid-19 scenarios over and you will be able to engage that user beyond covid-19 as well. So for us on the chat platform, we are seeing a lot of good activity as well. We are seeing a lot of People who are visiting constantly, average messages sent have gone up for users and I think those are certain KPIs which we are following per product as well. So of course there would be certain KPIs that would be more important right now, which is mostly making sure that our customers are happy but simultaneously small experiments will have their own respective KPIs and we are trying to do well along making sure that users are getting value even when there is no travel. A lot of other companies are making sure that even though the CAC is down, they are benefiting from the LTV. They are making sure that the user is sticking to their platform which can extend beyond covid. So I think those are the right retention parameters to look at.
Triton: Great. I agree with you that in these times I think the LTV and keeping your customers happy become the most important KPI. Well, so now Tarun, the KPIs you mentioned, are there any challenges you face in tracking or keeping it up to the expectations?
Tarun: Well, of course one of the most important things is general user behavior across. People rate when they are unhappy but when they are happy, maybe they will not rate. So it’s very tough to know whether you are catering to a large audience or you are only getting the ratings of the unhappy customers and which would be valid across anyways, but especially in the times like these, maybe there is a spike in one of the two and you will have to go case-by-case try to resolve each of them and make sure and eventually hope that those customers who were giving bad ratings come up to the good rating part. But simultaneously, I think accuracy in many of those KPIs as of now might not matter as much as you are not making sure that it’s exactly accurate as of now, but you’re trying to have a good range. Eventually when you are not in a scenario when there is a spike in ratings, which is right. Now then you would want to be in that good range of things so that even when the sky is the spike is not there you are maintaining a good rating across various user behavior KPIs. You should still so if currently, the spike in average user visits to your app has gone up to five to eight, you still want to win that good range of plus three. You don’t really care if it is eight or seven. Similarly, if the NPS has sort of been in a Range which is good. It’s fair enough. You don’t really care if it’s at the highest level right now because many of those would be catered to as the spike goes down as a lot of happy customers also were so the accuracy might not be very high as of now, but simultaneously you have to see that the scale of things or the people who are doing those ratings right now has gone drastically up. So it is different from your normal scenario as well and therefore it should be treated as such. You could still be focusing on improving those KPIs through your multiple initiatives and you should see the impact of those KPIs from your current levels itself. What will truly come into the picture is that after covid-19 is probably over you will see whether you are able to sustain those levels of LTV, whether you are able to sustain those levels of NPS, whether you are able to sustain those levels of average visits of the user. I think that would be really interesting to see as well. What you can only do is try to remain in that rain so that you are fairly clear that these are the things that are working and then you continue working on them. So I think that is the key thing and we of course use a lot of tools as well to monitor all these metrics across. We definitely use Analytics comprehensively. AirAsia is definitely one of the most analytics first companies. So we have a lot of data to look at and we are constantly following it. But simultaneously we are aware of a lot of other things which are ongoing. I think that is why we are trying to do a lot of our initiatives with a mindset that users should be happy right now and some of those KPIs should constantly show the effect of the good efforts which we are putting in.
Triton: ‘Absolutely’. I think some useful tips. What you also mentioned is, of course, you know the scale of the data and the scale at which AirAsia operates. I think it is tremendous. So the scale also brings me to the next aspect. When we talk about scale, personalization is something that has helped a lot of brands, some of which we are working very closely with, so I wanted to know your thoughts on personalization at scale. “How do you do them? What are your thoughts? What are the benefits and at AirAsia have you guys done anything for personalization”?
Tarun: All right. So personalization is a kind of must-have in the current scenario, especially for the app. I will also write from day one or day two when an app has been built and it has started getting some downloads not day one or day two but maybe day thirty when you have had sufficient app installs, there is sufficient user interaction. Then you start personalizing in some way. Maybe that initial personalization is just based on some collaborative filtering model. You are just trying to show users the content which is relevant for them based on other user behaviour data because you don’t have as much data for them, but eventually that moves towards them a furthermore personalized model as well. So personalization is something which is therefore more stats right from day 1 day 2 day 34 now. It is an industry-wide practice, especially in the current Tech World where it has become much more than having a big team. It is more about figuring out what sort of personalization you need and implementing those basic levels of personalization as well. So I think one of the basic personalizations, especially in the communication part when you are reaching out to the user you are reaching out with the first name, you are reaching out in a channel, which they prefer you are reaching out at a time which they prefer. So this is already a sort of personalization, but of course, for bigger companies including a ratio it has escalated to beyond communication to everything they do on app and website. You can probably see that the widgets which appear on the app and web are personalized based on your user behavior. If you go to certain destinations, then those destinations would be suggested to you. Similarly for Grab, if you are constantly ordering from one restaurant that would be available on top for you. If you are going to a particular destination that would be prefilled. So I think it has been across web apps and communication personalization is well ingrained in our ecosystem now, especially for bigger us because we have a lot of data and when you have a lot of data you try to make use of that data in the best way possible. So you try to plug it into your current user behavior. Try to see if plugging in this data at this point sort of increases the user conversion. Similarly by Saving user history in terms of where they are traveling to, pre-filter searches and see if that leads to shorter time in terms of conversion than in terms of more conversion percentage conversion as well. So I think one personalization at every level is there right from the level of website app to communication. Even to a certain level, we are trying to extend personalization beyond the conventional things and predicting a lot of things now. So it transcends the boundaries of conventional personalization. And then you go to sort of predictive personalization in which you’re trying to predict what they will do next and suggest things based on that. So, that is also personalization, but that is sort of just moving on from a basic personalization, which is based on current user behavior and what currently serves to predict for the future and serve the future as well.
Triton: Absolutely. I think you have brought out a very important thing like personalization and some way or the other is being done by a lot of brands but this next phase of predictive personalization like how can you predict or recommend such as certain things I think is changing. Some brands have started with that but it’s still quite a journey far away. So what are your thoughts on this? So I think for this at scale and doing this prediction you will require a lot of Predictive Analytics. You require AI machine learning to be working. So what are your thoughts and some of these aspects and any examples if you have come across?
Tarun: So I will say that of course, you require a lot of data in the first place. So a lot of users need to be using your app well so that you are able to draw out meaningful insights because with less data. If some new start-up tries to do it then maybe this predictive analysis might not be that accurate again and therefore a rudimentary personalization would work better in that way or a rudimentary recommendation might work better that way. But with a lot of apps who have a lot of data, then they can probably experiment with some serving such a certain set of predictions across their behaviour. And I think with respect to that it becomes important that when you start small you don’t focus on a lot of things. I think with a lot of Brands what also happens is that you see a lot of big levels of Predictive Analytics going on across the ecosystem and you try to imbibe those features directly in your ecosystem, which is not really possible or relevant. It really has to come down to that you start small, you start seeing some effect of some predictive recommendations for you. So like for us there are destinations which are suggested we try to do predictive modelling around a lot of things. We have a big data team so we can afford to do that. Therefore a lot of those insights help us in user conversions in drawing out suggested destinations for the user and figuring out what are the destinations we should be selling right now also or giving a focus on our website for that matter. So a lot of those Predictive Analytics is there but for other brands, I think it would come down to also figuring out what is the sort of areas where they want to have this kind of analytics. Maybe they would want to have this kind of analytics first plugged into that predicting what is the next time they will purchase. That’s the easiest low-hanging fruit because it does not really do a lot of changes on your ecosystem, but it is just a reflection of whether you have gotten the right data crunched or not. Whether you have gotten a ride inside or not. And when you sort of sum up, take it there then you extend it to your ecosystem. You try to extend the same thing in your ecosystem. So recommendations for whatever thing works for you try to extend the same thing and then go on to the next experiment or even if you are doing simultaneous experiments try to treat each experiment in parallel. If you will do five experiments in your app per say, and you are testing all of them and conversion happens, you won’t even know that due to which particular change did that conversion happen. So it has to be sorted out in a much better product for you. Probably the product team has to think about how all of these changes are not related to each other and only put in the changes which are not related to each other so that they are able to drop real insights. There should be a real Matrix, which they should be tracking for each of these conversions or success metrics for each of these Predictive Analytics on each of the components. So predictive analytics could be across it could be based on recommendations of products. It could be based on pre-purchase behavior. It could be based on retargeting. It could be based on post-purchase Behavior to induce the next purchase. So there are a lot of funnels that exist. What’s important to select is to select the right funnel to try out or small experiments to see if the insights for that final are working or not. And then to go ahead and Implement that at scale and even at that point be mindful that the experiments are not overlapping with each other in some way.
Triton: ‘Very well put Varun’. I think what you are saying and giving as an input to all our listeners also is to first understand where you are in your journey, understand your customers, and then you can plan because there is immense technology and a lot of things that are available, but if you don’t do the right thing, it may not really give you that impact.
Triton: Wonderful. I think Tarun, this has been a really good podcast says session with some very good and valuable tips. And I think this also comes from a lot of you know experiments that you are doing and also your experience that comes across. We are so thankful to you to be a part of this episode. I think my listeners would have immensely benefited from all your thoughts.
Tarun: Thank you Triton and it was a pleasure to be here and share some of the things which I have been able to gain insights on and therefore sort of thinking back on a lot of those things. I also get some clarity. So it was definitely helpful for me as well.