 From Seattle, Washington, it's theCUBE, covering AWS Imagine. Brought to you by Amazon Web Services. Hey, welcome back, everybody. Jeff Frick here with theCUBE. We're in downtown Seattle at the AWS Imagine EDU conference. It's the second year of the conference. Part of the public sector, kind of a carve-out with Andrew Coase Group, really all about education. And education from K through 12 to higher education, community college education, retraining of people coming out of the military. It's a huge segment and we're really excited to have our next guest. He's going to give a keynote later this afternoon on a new paper that he just published. So we welcome Richard Palmer, the practice leader for public sector for Open. Richard, great to see you. Good to be here. So tell us it's called Reaching for the Cloud. Yes. Look, what we've found is that for many universities, moving into the cloud has proved to be difficult. That there are lots of barriers in the way and they get a part of the way along and all of a sudden they hit a wall and it takes time. The big number that we keep looking at is only 30% of application workloads are in the cloud at this point. And after six, seven years of the public cloud being available, it really suggests that there are barriers, significant barriers there. Right, right. So what are some of those barriers? I mean, again, as you said, we're kind of well down this path. So whether it's just legacy stuff that's not worth moving, but I would imagine most of the new workloads are coming in. I mean, they've got to be getting with this program. Purchasing SAS is an obvious ploy. It gets you right out of all of the problems that you had before. Look, the first thing that barrier that people find is that cloud's different. So the skills that you've got in your team, the way that you work finances, your project methodologies, everything is different to engage with cloud properly. And the way that you design and build applications is different in the cloud. So taking a traditional organization, trying to go cloud, has everybody involved from the CFO, with funding cycles, to governance boards, which are the most wonderful thing ever in higher education, all the way through to staff skills and the way that staff think about applications. And if there was one thing in my time in higher ed that I saw time and time again, it's the instant legacy problem. So somebody creates something and does it a special way because they know better than the vendor, and we had this infrastructure anyway. So why not reuse it? And they create an orphan that is neither manageable by the vendor nor manageable by the organization, requires that individual to remain with the organization, well, well past their expiry date. Let's put it that way because they've put things in that are just unique to this single installation. And that's the transformation you see in the cloud. It's software as a service, that's, it's a native thing you don't look at, how it's hosted, you don't care about anything but using it. But the danger point is in moving to infrastructure service or platform as a service that you carry over that customization thinking, which creates, if you like, instant legacy. So that's one of the barriers that we see. I'm just curious, because you brought up two really big things. Just the whole financing and the way you buy it, the way you budget it is completely different than a big capital expenditure that you're depreciating over time. And then as you said, the skill set. So in the enterprise space, everyone's got big piles of money and they hire the big SIs to come help them. They have instant skills, they can bust them in by the minis, the dozens and help them with some of that financial engineering. How is the system integrator or the services kind of industry evolving in education to help them make this transition? So there are two ways. And education being a public sector enterprise has that awful problem that if you provide advice, often you can't provide services. We tend to be getting over that a little bit now, but the obvious way is most higher education institutions who've moved through that process have engaged a strategic partner to help them to plan. So that's the first piece. And there are lots of them around and often they're very good, but moderately expensive. But the thing that they don't tend to do as well is to find the right partners for the actual transition. So often engineers are trying to learn cloud technologies and apply them and get it right the first time around. And of course we all know that in experimentation you want to have learn fast and then relearn when you've come to something that you shouldn't have done. But if an engineer is thrown into a live production style project, there's no time for that relearn, replatform, learn as you go. So having not so much an SI, but an implementation partner is really important. And luckily many of the vendors or their networks are really quite good at doing those mid-level implementation projects now. So it's a matter of finding the right one. But certainly in my home Australian context, for almost all moves to the cloud, there's somebody who's done it several times before in education and has a good reputation. So I suspect in the US that's multiplied, it's a 10 times larger economy. There's probably 10 times as many people who've done it well before. Right, so the other piece that I'm curious if this came out, right? So there's the cloud as a more efficient way to run to your infrastructure and all that that means and cost savings. But much more importantly in some of the things we're hearing today is really to enable innovation. To enable you to develop stuff faster whether it's Alexa or some of these other things we're hearing about. I mean, how does that play in people kind of getting through the pain of getting through this process because if you don't innovate, and we just had somebody on before you said they're worried about competing with online and really having a good experience for the students on campus. Is that the driver? Is it the cost savings? Is it, how do you see that kind of slicing? We're seeing several drivers. The one that's most common is student success and retention. That is ubiquitous in higher education. To bring the cost down and to make sure that every intervention that the college or school does is meaningful and produces a positive outcome. So that's kind of the core business. And so things like analytics play into that and now machine learning more and more. But the motivator, yes, there's competitive motivator but it actually works the same for they're on campus as they're online that if you can help every student to be successful, you'll gain reputation. If you can do it efficiently, you'll drive down costs. So that's beneficial. But then you're asking about innovation. That's a step after you've put all the pieces together to do core business well. And the key elements in doing core business well is shifting from traditional to agile because agile projects have benefits on the business side as well as the technical side. One of the most important things is to be able in the agile space is to be able to iterate quickly. And that is just as important on the academic side as it is on the technical side because usually the academic or the administrative folks don't know what they need until they've actually experienced it. Most times when you're replacing the system you ask the people on the front line what they want and they answer exactly what the last system did, but better. Right, right. So that innovation cycle, do then measure and then cycle through is part of the agile piece. And the second part of it is being able to differentiate between what is actually going to make a difference for your students and what is just pure whim. What we think might be better but is actually going to cost money, create legacy, move us away from standard practice and actually is going to bring no benefit. So really important to attach real KPIs to differentiating practices. Right. And get away from customization where it produces no benefit. The third element is platforms. Once upon a time we used to build our special systems from the code up. We shouldn't do that anymore. We shouldn't be caring about what database is underneath. Application platforms are faster, more effective and require less in-house skills to maintain over lifetime. So that's the third element. So one of the things that the enterprise has been able to benefit from is we just leave the ERP alone, right? Just, there's a lot of stuff that's just not worth lifting and shifting. But it's kind of customer interaction applications and there's a whole kind of class of applications that open up the opportunity to leverage all these kind of platforms and fast development, et cetera. How's that playing out on the public sector side? Because before we turn to the cameras, you talked about just the pain of lift and shift and you run into all kinds of issues. You don't get that good easy win, that good fast win. Are they thinking in terms of setting aside kind of an innovation development team that's working on some of these kind of new age things that aren't kind of the core systems that maybe he don't necessarily want to lift and shift anytime soon? I'm a big fan of innovation teams when you're working directly with research. Not sure that that's the best model for mainstream innovation. It's much more useful to leverage the folks who are actually working directly with the business, people like business analysts and to shift those into thinking beyond the mundane. Because the business analyst usually has a very intimate relationship in the nice way with their business partner and they can engage with what would make life better, what would make things more productive and then to quickly bring resources in behind that idea into a quick proof concept. But you've hit on another whole issue there that the idea of a ubiquitous engagement layer that both delivers a really high quality online or digital student experience but also provides a whole lot of information that can be then analyzed to work out what was the best thing to do with a student is really transformative. And we're seeing the best vendors move into that space even with traditional systems and what they're doing, and I'll use a couple of student management system vendors as an example without naming them, but their traditional systems, they will either host them for you or you can do it on premises. But their new analysis engagement systems are cloud based. So it doesn't matter where your implementation is, you can buy new software as a service that gives you really good analytics and a new communications collaboration engagement layer often with CRM and collaboration tools mixed in in a brand new platform. And that's really transformative so it allows you to keep your transactional system in place but re-skin it with a new engagement layer. And if you can clip your university services into that engagement layer, then you get that 360 degrees view of the student with actually out having to shift major systems. And as you said, that's just money spent to lift and shift a system because there's no strategic benefit except if it allowed you to upgrade because so many universities are stuck with an old version of a system either because they customized it or they haven't got the infrastructure to host the new version or whatever it happens to be. So there is a strategic benefit to be able to stay with the latest version, particularly as most good vendors are providing new features pretty regularly on their most up-to-date and are only doing maintenance releases for previous versions. It's pretty interesting that Andrew pulled out of the three themes for the show. Tomorrow's workforce, role of ML and innovation transformation. That ML got its own bullet point because it is such a kind of an underlying infrastructure that drives so much value across lots of applications. And I just found it interesting that you get kind of the retrain academic institutions in the ways of big data because it's very different than maybe the way that they grew up thinking about data, the quantity and the way that you deal with it and how much you have and sampling of old data versus all the real-time flow. So yes, but the next generation is autonomous. And whether it's self-driving cars or student advisement, we're seeing the leading edge providers provide in the education space pretty much autonomous student advising now, except to the point where you go out of the mainstream. But that rapid, good advice from bots, basically. But when you get to the real world and autonomous systems, we're gonna see a real shift in even the university sector of that interact with people and the environment. If you're doing self-driving cars, you're talking sub-millisecond responses. So that whole world of IoT plus sensing technology plus called smart campus all coming together in the next iteration and being paralleled in the services and maybe even the academic world, that'll probably be a bit slower. But taking the same autonomous kind of thinking and moving beyond just supplementing human transactions. Right, right. All right, well, Richard, thank you for taking a few minutes. Good luck on your keynote this afternoon and we'll look forward to a dig it into the paper. It's been a pleasure, thank you. All right, he's Richard, I'm Jeff. You're watching theCUBE. We're at AWS Imagine EDU in downtown Seattle. Thanks for watching, see you next time.