>> Live from Las Vegas, it's theCUBE. Covering Adobe Summit 2019. Brought to you by Adobe. >> Hello everyone, welcome back to theCUBE's live coverage, here in Las Vegas for Adobe Summit 2019. I'm John Furrier, Jeff Frick my co-host this week. Michael Yung is the CIO of Asia Miles. Welcome to theCUBE, thanks for joining us. >> Great to be here. >> So take a minute before we get into the conversation about machine learning, and all the cool tech. What does Asia Miles do, what's your role there, and stuff they do? >> Asia Miles is the loyalty reward program of the Hong Kong, Cathay Pacific Airways. So, typical airline, but we have the reward program to support our members of Cathay pacific airways. We have over, about 11 million members, and over 700 partners around the world. >> How many members? >> 11 million. >> 11 million? >> Wow. >> That seems like a lot to me. (laughs) >> We are the leading loyalty program in the region, in Asia. In fact we started the program about 20 years ago, so in 1999, so this is our 20th anniversary. >> Wow, congratulations. >> So, similar to any Loyalty program, our members can earn miles by flying, traveling, dining, shopping. Even have your mortgage with our banking partners. At the same time, using the miles, you can redeem rewards. Hotel stays, flight tickets, and even for tablet computers or mobile phones. So you can do all of this. >> So, you did the web 1.0, web 2.0, web 3.0. (laughs) You've lived the journey. >> Paper 1.0. >> So my job is actually leading the digital part of the team. As you know, like loyalty program, you don't have protection lines, you don't have branches, everything is digital. So our web, our mobiles, our engines to support the earnings, and engines to support the reductions are all digital. So basically, we are more like a digital marketing company, we links the partners, their products, their offers, to our members. >> So, important is obviously the data, it's super important. And having connections points, APIs, open systems. Is it open APIs? >> Yes, all of these are technologies in our stack. So, basically our membership profiles are databases. And then with APIs we can do all sorts of modeling, or calculation, or segmentation. And then we push through our marketing offers, or campaigns, to our targeted members. >> That sounds like good architecture. Now what, specifically of Adobe product stack, are you using, for Adobe? >> We used almost the whole suite of Adobe products. We started our baby step about three years ago with Adobe Experience Manager. Basically our contact management systems are website or mobile. And then we extended to campaign to automate our marketing campaigns. And then later on audience manager, target and analytics. So it has evolved. So basically a full stack. >> So you're a big customer of all the products. So one of the big things they're talking about is the data, role of data, and machine learning's coming up a lot. How are you applying machine learning, with all those millions of members, and all the different diverse contact you have, and the different connection points to partners. You have to, kind of have this free flowing operating environment, platform yourself. So how are you using machine learning to either automate away things that you're doing manually, or creating new innovation insights. >> As I mentioned, we have to match the offers from our 700 partners to 11 million members, right. And therefore we build certain technologies, like propensity modeling, that we can tell, say from you miles balance, your life stages, your persona, and your lifetime varial, and then we do, what we call the partner recommendation engine. So the recommendation engine will push certain offers to John, or to Jeff already, based on all your profiles. And that requires some machine learning and modeling as well, from our data scientists. >> I'm curious how the expectation has changed over time in terms of, kind of what your members expect to get out of the application. Because I assume they want more, and more, and more, what was special today is common tomorrow. And how you've been able to continue to adapt and change what you often experience. >> Right, great question. First of all, our members really like to go mobile, so our offers have to be location based. So with your mobile apps, then you can see, okay what are the popular restaurants around me, that I can earn miles easily. Or, if it's a Monday, then you can earn, say double miles if you buy something with retails partners as well. So all this, the partners, and the members expect more. And, secondly, members are smart enough to tell that, oh, your offers is generated by a machine. It's not personalized enough. For example, if I just fly to San Francisco last week, why'd you promote San Francisco flight ticket to me? Or hotel again? >> Right. >> I'm not going to San Francisco again. >> The re-targeting thing is brutal. >> Brutal, yeah. So you have to really base it on the transaction history, and the other features or signals, and then define the next offer. And this is really important. >> And do you help the customers figure, because you just said if you eat out on a Monday, maybe you get double miles because the restaurants are slow. Is that something that you guys have discovered in your analytics, that you're helping your partners to get more pull on their offers, or is that being driven from them? Because you have a lot, you've got a lot more data than an individual restaurant, or some of your other partners. >> But I mean, even in Hong Kong, Monday's a slow day for business. >> Right, right, right, right. >> So it's good to help out the partners a bit, you earn double miles. Or on certain important days, or holidays, you get triple miles by buying something. So it's natural for our partner's, and our member's expectations. >> You have an economy. (laughs) It's like, you've got to have a fiscal policy. >> Well let me tell you all loyalty programs pretty run like this. >> It's really highly data driven, you have reputation, you have influence. >> Exactly. >> It's very important, I'd almost imagine, contextual understanding about what's happening, and having the right data. You mentioned that re-target thing, about San Francisco. I see this all the time on re-target, they don't have the context. I mean that really makes for a really poor personalized experience. Talk about context, having data in context to something. How hard is that? >> Well it's really from data, turned into information, and then actionable insight, it's really hard. So, even we have so many team members doing all this modeling, there are times that we need powerful tools to do proper segmentation and targeting. And that tool's got to be really flexible, and fast responsive to certain context. And with that Adobe products help us a lot. >> What's the biggest to do for you, going next step as you continue to grow. You're digital, all digital. You have Adobe Suite, cloud computing scale, a lot of data context, a lot of usable data. What's next for your business, what's next for you. >> Well, last year we started to test the water to try out blockchain technology. So we have the marketing campaign rules, and packed that in a blockchain smart contract. And this is one of the things that we invested a lot of time and resources into it. We believe in the future marketing campaign has got to be more real time, and you can earn your bonus miles straight away, instead of waiting two, three months until the end of the campaign. So hopefully with the marketing platform, and also newer technology, and better data, we can do better campaigns. >> In terms of skills to deal with the kind of things that you're doing, with future proofing your business with blockchain, love that. Smart contracts going on, peer to peer, immutable, love that value proposition. You get reputation, move that over into currency. >> One of the options. >> Asia coin. (laughs) >> Optimize is one of the options. >> What else is on your mind? KPIs, how do you look at data sets, how do you guys view? >> Measure success. >> How do you measure success? >> Well, I would say first of all, all the stakeholders have got to be happy with the program. I mean, the stakeholders include our members, partners, our shareholders, and our employees. They're important to make sure that the program is successful. And also including the engagement ratio, and our package ratio, where there are a lot of members that usually don't have chances to redeem things, and then they let the miles expire, for example. So helping them maintain a healthy package ratio is also a KPI that we measure carefully. And then, other than that I think all our employees or staff, they let you know, or they need to understand how technology and business mix together. If you're good in business, but not knowing marketing technology, for example. Or if you only understand technology but not the business, for example, it's just not good enough for the future. So the skillset why you have to understand both. >> How are you using technology, especially Adobe, how is Adobe helping you, and then what other things you might be doing, to help internal processes get better? Because one of the things I'm seeing here at this show is, with the platform, as you start to thread the data together and let the data, kind of naturally flow, with machine learning and the different data points, you can start to get some visibility to insights that might not be there. So that's going to cause some internal disruption. People might lose there job, or new jobs emerge, there's always conflict when you're progressing. How do you use technology, and this technology, to keep getting higher functionality, better economics, what's the internal struggles, and gains look like? >> Well, for example, before the days of marketing platform, or Adobe days, you may need to take weeks to prepare a campaign, if not months. Because you need to prepare all the contents, all the lead assignments, and then you push out through all the different channels. But now you can be always on campaign, different dates. And, for the blockchain example, we can actually eliminate the reconciliation and settlement effort. So the back office operation team, they can move along to do something else. To do more campaigns, or to talk to the partners more, to understand their needs. Instead of just number crunching, we do reconciliation. So I think, it's not about with less resources, but with the same resources, how to do more things. >> Right. >> And it's almost continuous improvement on the campaign. >> Yes, continuous, all the time. >> Versus just, you know, let's plan a campaign, run a campaign, measure the campaign. It's just constantly going. >> Prepare, run it, and then measure. Just never ending. >> As an Adobe customer do you like the direction that they're going? >> Yes, yes. All exciting products are in the road map. And we are ready to explore more in the future. >> Michael, thank you for coming on and visiting us. >> Okay, my pleasure. >> We appreciate it. Here inside theCUBE we're taking all the action, here at Adobe summit. Getting the data, sharing it with you out in the open internet. Thanks for watching, I'm John, with Jeff Frick. Stay with us for more coverage from day one after this short break. (upbeat music)

>> Narrator: Live from New Orleans, it's theCube, covering VeeamON 2017. Brought to you by Veeam. >> We're back at VeeamON 2017 in New Orleans. Andrew Christensen is here. He's a senior systems engineer with the Global Data Centres Study Group, higher ed organization. Andrew, welcome to theCube, thanks for coming on. >> Thank you very much for having me. >> You're welcome. So interesting name, tell us about the organization. >> Study Group, Study Group began life as a small education college in the UK, about 15 years ago. Over the time, we've grown to a global organization. We take on about 85,000 students a year, from close on 160 countries. We have 85 sites around the world, very much a global footprint, very thin corporate terms and in IT terms. >> Okay, is this your first VeeamOn? >> No, it's not. I came to the previous one in Las Vegas, and that experience meant that I had to come back to this one in New Orleans. >> Really, why? Tell us about that. >> It's a great experience. They know how to do their events very well. The information is first class, and as a Veeam product user the information and the experience, in the room available to you is wonderful. >> How long have you been a Veeam customer? How did you... Tell us about your journey. >> Well, our journey, we were very much in the legacy ballpark of backups. We had some unnamed products that we were using, which were very old school in their thinking and how they did things. We realized quickly that the amount of data that we took on was growing and it was going to outpace our capacity for backup, and the simplicity and complexity was growing too fast for us, couldn't manage it. It wasn't going to be feasible, went to market to start looking for a better solution, and Veeam was top of the list. >> So, you mentioned data growth as one of the catalysts for, which created more problems, obviously for your backup, made it harder meet, maybe it was backup windows at the time, or RPO and RTO. Did your decision to change your backup also coincide with an increase in virtualization? And did that have a ripple effect? Can you explain, we've been talking about that all week, but I'd like to validate it with a practitioner. >> No, spot on there. We virtualized quite early on in the grand scheme of things. We went to VMware very quickly, we're now running Hypervisor with a vShpere 5.5 environment. Now that was a well good, I don't think the practices that we took in and a lot of the infrastructure alongside that kept up with that, backups is one of those things. And when we started looking at what we needed, to really work with our environment, get the most out of our virtualization project, we needed to do something very quickly and backups is a key feature. >> Andrew, as a global organization, how does Cloud fit into your architecture, what you're doing, maybe you can kind of sketch out a little bit for... So we know where Cloud fits. >> Our solution, although quite simple in principle, it's never simple, let's face it, anything IT, especially in the engineering scope guide, never simple. >> Dave: This keeps getting more complicated. >> Exactly right, and you know for better or worse, that's how we do these things, especially when it comes to a Cloud scenario. You add a little bit of complexity, but often it pans out to be worthwhile, especially in dollar value. Our solution takes local backups in a hub and spoke concept, our data center as being the hub, and our branch starts being the spoke, consolidate the data from all the sites, hold a decent amount of data as backup onsite, and then everything else will actually ship out to the Cloud, and that being AWS in Glacier storage. Now that came about mostly because our core data center is in Las Vegas, that we have no hands on site. So we didn't really have the option of a manual tape service, a paid for service, very expensive. So we needed to shift away from your old school, typical tape service environment. Having good bandwidth in Las Vegas, and availability to get to the AWS readings, made it a good solution for us. Our tool to do that, already in place with Veeam, made it very simple. >> So what are you... So your target is Glacier? >> Correct, yes. So long term retention, and legal retention especially, we push everything out to Glacier to fill that need for us. >> And okay, so that's the last thing, maybe I missed it, there's something in between obviously, if you need to do a recovery right? >> Correct, so we keep some local storage as well, depending on the environment and the data itself, we'll keep it locally on site in our racks for a certain amount of time, maybe a year, maybe two years depending on some of the data. Everything else has a duplicate and a long term, goes out to AWS. >> Alright, there were a couple of announcements this week about AWS and also about Glacier. What did you hear? What interests you? >> Well, I mean the V10 announcement and its interaction with AWS, hooking in your AWS accounts to s3 Glacier and what not, very promising, very promising, very excited. I'm going to hit up my account manager for a trial on that very soon, cuz that could simplify our process, and I imagine a lot of other people with hybrid cloud scenarios will leverage it as well. For those people that have work loads in AWS, the agentless backup function, very promising. It's a logical step I think, in the partnership that's built that is very logical as well, it's going to help a lot of people. >> What is driving, in your industry, the availability needs and how has that evolved over the last couple of years? >> Well, it's a catch 22, for us data security is paramount, a student come to us, they sign up for a course, in a lot of cases it'll be an international student. Now that's a well and good, but when we look at the data that we take from that student to get them into a course, it's essentially a how to kit for data theft and identity theft. So we need to protect our data very well, you know, we've got a lot of personal information, we've got passport photos, we've got visa replications, the whole shabang. So being able to make sure that A, it's available to the people that need it, so that they can get them into their courses, get them learning stuff, which is what our ideal is, and making sure it's secure, no matter where we put it, backups, availability, all that sort of stuff, needs to be secure. So a solution then at all has to incorporate that as easily as possible. >> Bill Philbin was asking the audience this morning, have you ever had to do a recovery? He said about a third of the audience's hands went up, presuming your hand was up. >> Yes we have. We've both tested, and we actually had incidents where we've had malware come to the business in certain aspects and having a good recovery point on site, and a quick easy interface, the single pane of glass, to coin a pro word of the moment, was very useful. You know, stops the heart a little bit when it does happen, but after you go through the hoops, and you understand what you're doing with the product, it really does give you a sense of security. >> You know, large organizations, big banks and the money business for example, they have very explicit disaster recovery plans, they might have three site data centers, they get gobs of money they can throw at this stuff. Higher education tends to be, a little tighter with the budget, fair to say, but also a lot of smaller and mid size organizations, I think it's fair to put you in that category, oftentimes had very little, if any, sometimes data... Disaster recovery, and what they've done is when they re-architected the backup, they said, you know we can kill two birds with one stone. And so to bring those two worlds together. Is that what you did? And how would you describe it? >> I'd call our solution a bit of a hybrid, in line with the backup scenario that we do have, both offsite and hybrid cloud scenario, we also do a DR solution internally. So we have a data center in Las Vegas, we also have one in Sydney. So we do take some DR concepts down to Sydney to hold on to that, very limited. Your bang for buck with DR, it's very hard to justify when you go to manage it and say, well you know, the cost of failure needs to be calculated here. It's very difficult to make that argument successfully. So having a tool, that we already used, that could also do that, very helpful in the first place. You're right in that we are an SMB in the traditional sense and the feature set that does come with Veeam, is quite good for that I think. We're quite OPIC shy, as a tradition, so being able to put a little bit of infrastructure in place, and sort of pre-purchase these things, get the cost out of the way, with CapEx, helped us a lot. So, no more licensing involved, Veeam took care of it in house already, and a little bit of expenditure took that solution very well for us. >> Enter one of the interesting discussions we've seen in the last few years when we talk in the education realm is the importance of data and how can you leverage that data? Of course, we talked about some of the security aspects. How has the role of data changed in your world? >> It's a bit of a catch 22. It's recognized that we do take on a lot of data. How we use that, it's an ongoing question. I mean people have put a lot of BA type roles in place to try and leverage that a bit further, get some use out of it. We have this data, it should be an asset to us. It's very difficult to do successfully I think. People don't really know the questions to ask, of their data. You know, maybe there's a bit of thought leadership or some extra disruptive technology that should come along and help that out a little bit. I think in the near future, it will be a very big question there to be answered and a lot of demands can be met by that. >> How about your students? You know, there's got to be, from all the devices, what kind of pressures does that put on the IT role? >> Well, it's substantial, I mean in our particular role, especially in the UK and Europe, we actually house our students. So, everything from living aspects to education whatnot, everything is handled by us. So their safekeeping, or their lifestyle, their quality of life and such. Now, in today's modern age, you now have two iPads, you have a Chromebook, you have an iPhone and whatnot, all that needs to be handled by us in a secure fashion. The data that comes from that, the content that gets delivered to the students, both privately and during their education, it needs to be both readily available and useful. Making it available to the students, as well as protecting them I suppose, in a secure fashion, making sure that the data that they hand out over these networks and use is safe. That's a big concern to us. >> So a lot of talk this week about ransomware of course. A guy in your position, talk about make the heart stop a little bit. How do you look at that problem? What solutions do you have? And what would you like to see the industry do? >> It's a difficult question. There's no easy answer to that at all. Recently we've heard a lot about machine learning and predictive analytics and whatnot. We use some products that do real time assessment of file stores, file usage, and whatnot, and predict excess usage I suppose. All of a sudden you can start seeing if there are extra files being encrypted very quickly. You can take action based on that, so these are a clear sign of ransomware. That said, we educate a lot of young people, we educate a lot of young people in IT as well. We have identified that a key threat is often going to be from internal. How we protect against that has really shifted our mindset a fair bit. A lot more legislation is coming in, in the UK especially, starting to come in in the US and Austral-Asia, and the requirement for that is only going to grow. It's a challenge that I can't really say in the future how we're going to predict it and act on it, but it is always going to be in front of mind. >> Do you think you could use your backup data, because essentially you're pushing change data over the network, you know, constantly, or at least multiple times per day I presume, right. Do you have the tooling to monitor that activity and identify anomalous behavior, where maybe you're pushing more data, or you're seeing more encrypted data going at a particular time. Do the tools exist to do that today? >> To an extent yes, getting them all together, to be viewable and visible data for your technician, your engineers and whatnot is a bit of a challenge I think. Anti-virus and security software is out there that can do this for you. Also, the data analytics tools, they're out there at the moment. VeeamONE is actually a useful tool on that front, can help us out a lot there. Making sure that the person responsible or looking for that trend knows where to go and has a good single pane of glass per say to actually identify issues. I think that's the key. >> Could you... Another, I've been thinking about sort of how to solve this problem. Could you put like phantom files out there in the network? Phantom high value files like, student credit card list. >> Andrew: You mean the honey pot scenario? >> Yeah use it as a honeypot, yeah. >> Absolutely, I mean a lot of the more enterprise size corporations are doing this, and you can actually leverage that, take them on as a service if you need to. There are companies out there that will offer this service to you. It is quite expensive for what it is, and yet when we calculate the cost of failure, I think the expense might be justified. >> Well like you said, it's hard. Your CapEx, CapEx phobic, I forget what you said, but I'll say CapEx phobic, challenged. Okay, we're out of time so the last question. Takeaways from VeeamON 2017, things that you're excited about? >> Once again, the AWS announcements in v10 and the partnerships coming from that, very exciting, very exciting. Looking forward to that, to being able to test it a little bit. The feature set that keeps growing. When we started out in 8.5, then 8.5 went on to 9.5, and the growth from 8.5 to 995 and now 10 on the horizon, it's massive. If they continue this growth it's going to be one of the best products out there. I'm very happy about that. >> Alright Andrew, thanks very much. Appreciate it. >> Thanks very much for having me. >> You're welcome. Alright keep it right there everybody, Stu and I will be back with our next guest, right after this short break. This is theCUBE, we're live from VeeamON in New Orleans.